JOURNAL oF tHe ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 59 NUMBER 1 US ISSN 0004-2625 Journal of the Arnold Arboretum Published quarterly in January, April, July, and October by the Arnold Arboretum, Harvard University. 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 1-45, reprinted, and some back numbers of volumes 46-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, U.S.A. EDITORIAL COMMITTEE B. G. Schubert, Chairman S. A. Spongberg P. F. Stevens C. E. Wood, Jr. ASSISTANT EDITOR E. B. Schmidt Printed at the Harvard University Printing Office, Boston, Massachusetts COVER: The stylized representation of fruits of various Leguminosae, form- ing the design r Volume 59, is based on material from plants growing in the Arnold Arboretum af Harvard University. The design was planned and executed by Karen S. ure, who lso drawn several Velmur of the preceding covers as well as various deve used in the Journal of the Arnold Arboretum and on the offprint cov The three ap eaullgs of the Leguminosae are represented by the fruits shown in the desi The genera are, top row (left to right), Albizia (Mimosoideae), Gleditsia (Cuca IpINGienE) Caragana (Faboideae), Gymnocladus (Caesal- pinioideae), and lower row, right, Colutea (Faboideae). The family Leguminosae, the second most perpen family of flowering plants, is to be the subject of an International Legu onference at the Royal Botanic Gardens, Kew, Richmond, Surrey, aan in July, 1978. This cov- er is our salute to the complete success of that Conference. Second-class postage paid at Boston, Massachusetts JOURNAL OF THE ARNOLD ARBORETUM VoL, 39 JANUARY 1978 NUMBER 1 FOSSIL DICOTYLEDONOUS WOODS FROM YELLOWSTONE NATIONAL PARK, II ELISABETH F. WHEELER, R1cHaArD A. SCoTT, AND ELso S. BARGHOORN THIS PAPER is the second in a series dealing with silicified dicotyledon- ous woods from two regions of Yellowstone National Park, Wyoming and Montana. In the earlier paper (Wheeler, Scott, & Barghoorn, 1977), we described eight species of fossil woods in five genera (Alnus, Carpinus, Ul- minium, Magnoliaceoxylon, and Plataninium). Species in nine additional genera are described here; other dicotyledonous woods will be described in a third paper which will include floristic and paleoecological discussions. In Yellowstone National Park, silicified woods occur most prominently in the striking succession of fossil forests at Specimen Ridge and Amethyst Mountain along the Lamar River, and in the Gallatin area some 40 miles to the northwest in Montana. The fossil forests in both areas are in volcaniclastic rocks of the Absaroka volcanic field. The Amethyst Moun- tain and Specimen Ridge fossil forests are in the Lamar River Formation; the Gallatin forests are in the Sepulcher Formation. These two formations intergrade laterally and are of late early to early middle Eocene age (Smedes & Prostka, 1972). Detailed locality information was given in our first paper (Wheeler, Scott, & Barghoorn, 1977) on these fossil woods. ANACARDIACEAE Rhus crystallifera Wheeler, Scott, & Barghoorn, sp. nov. FicuREs 1-3. Growth rings. Present, distinct, 1-3.5 mm. wide. Vessel elements. Semi-ring porous, transition from earlywood to late- wood gradual; extent of earlywood zone variable; pores mostly solitary and round in outline, some multiples of 2 to 3 and clusters present; tan- gential diameter of earlywood pores 105-205 pm., mean 175 pm.; radial diameter of earlywood pores 118-210 ym., mean 178 pm.; tangential di- ameter of latewood pores 40-100 pm., mean 60 um.; radial diameter of latewood pores 38-125 p»m., mean 80 pm.; length 275-460 pm.; perforation plates simple; intervascular pitting crowded alternate, to 10 pm. across; pits to parenchyma horizontally elongate and irregular in shape; tyloses present. © President and Fellows of Harvard College, 1978. 2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Parenchyma. Paratracheal, vasicentric, generally with 4 cells per strand. Rays. Mostly 1 to 2 cells wide, rarely 3 cells wide; heterocellular; al- most all rays with inflated crystalliferous cells, in radial section these ap- pearing twice the size of non-crystalliferous cells; multiseriate rays 6 to 27 cells, 92-600 pm. high. Imperforate tracheary elements, Septate fibers present. MATERIAL. One specimen of mature, silicified secondary xylem. HototyPe. U.S.G.S. Fossil Wood Collection No. D-2054A-14, measur- ing 70 « 95 & 150 mm. vyst Mountain, U.S.G.S. Paleobot. Loc. No. D-2054A. — LocaLity. Amet Both the Anacardiaceae and Burseraceae have species with the combi- nation of characters descriptive of this fossil; that is, pores solitary, rounded, and in radial multiples, simple perforations, vessel—ray pitting simple and variable in outline, paratracheal parenchyma, septate fibers, and narrow crystal-containing rays. The similarity of the wood anatomy of these two families is well known and is used as evidence of their close relationship (Heimsch, 1942; Webber, 1941). The Yellowstone wood is semi-ring porous. No members of the Bursera- ceae are known to show any tendency to ring porosity, while ring porosity is known in the tribes Spondieae and Rhoideae of the Anacardiaceae (Heimsch, 1942). Unilaterally compound vessel-ray pitting, a feature not observed in the fossil, is a distinguishing characteristic of the Burseraceae (Brett, 1966). For these reasons, we consider the fossil to be a member of the Anacardiaceae. Spondias axillaris Roxb. is the only member of the Spondieae in which ring porosity has been observed. We examined sections of this species and S. mombin L. Both have very wide rays (up to 6 to 8 cells wide), some rays have intercellular canals, and there are no more than two vessels per pore multiple. Pistacia, Cotinus, and Schmaltzia are ring porous members of the Rhoideae, but pore distribution and ray structure of these genera appear different than in the fossil. The anatomy of the fossil and extant species of Rhus is similar: the subgenus ToxIcoDENDRON is the only sub- genus of Rhus reported to contain septate fibers, and septate fibers are present in the fossil (Heimsch, 1940); crystals are more common in the fossil wood than in any extant Rhus examined, but in all other features the fossil’s anatomy is consistent with that of extant species of RAus. The numerous reports of fossil woods of the Anacardiaceae, most from Old World localities, are listed in papers by Awasthi (1965) and Kramer (1974). Fossil woods presumed to have affinities with Rius have been as- signed to Rhoidium Unger (1850). Three species, Rhoidium juglandifolia Unger, R. ungeri Mercklin, and R. philippense Orie, are known. All were described in the middle of the 19th century, a time when descriptions of fossil woods were very brief and generally unaccompanied by illustrations. Consequently, detailed comparison of the Yellowstone fossil wood with jean 1978] WHEELER ET AL., FOSSIL WOODS, II 3 these species is not possible. Without examination of the type material, it is also impossible to determine whether the woods called Rhoidium have any relationship to Rhus or the Anacardiaceae. Among the oldest known woods of the Anacardiaceae are those reported from the Eocene Eden Valley, Wyoming, locality (Kruse, 1954). The two species described from this locality are quite different from the Yel- lowstone wood. Schinoxylon actinoporosum Kruse is diffuse porous and has radial chains of up to 25 pores, no axial parenchyma, and intercellular canals in the rays. This wood was said to resemble Schinus in those fea- tures seen in longitudinal sections: however, it does not have the distinc- tive ulmiform pore distribution of extant species of Schinus. It is unclear why this wood was assigned a name implying affinities with Schinus, as the Eden Valley specimen lacks the pore distinction diagnostic of this genus. The second species from Eden Valley, Edenoxylon parviareolatum Kruse, does not resemble any one extant genus of the Anacardiaceae. It is diffuse porous and has numerous small vessels; the rays commonly have intercellular canals. This species was subsequently reported from the low- er Eocene of Herne Bay, Kent, England (Brett, 1966). These two species, a wood assigned to Tapirira from the Clarno beds of Oregon (Manchester, 1976), and Anacardioxylon magniporosum Platen (1908) from California are the only fossil woods of the Anacardiaceae known from the United States. No Anacardiaceae are reported from the Kisinger Lakes—Tipperary flora (MacGinitie, 1974). A member of the tribe Spondieae is reported from the Golden Valley Formation, latest Paleocene age (list prepared by Hickey in Leopold & MacGinitie, 1972). Rus is on a preliminary list of megafossils from the Wind River Formation, late early Eocene age (Mac- Ginitie, 1969), and it is also reported from the Wagon Bed Formation of middle Eocene age (Leopold & MacGinitie, 1972). Two species of Rhus and one of Toxicodendron occur in the Green River flora (MacGinitie, 1969). There are two species of RAws in the Florissant flora (MacGinitie, 1953). There are 15 species in the subgenus TOXICODENDRON (Brizicky, 1962). They are mostly temperate North American and Asian in distribution, al- though Rhus striata Ruiz & Pavon grows in tropical South America. CYRILLACEAE Cyrilloxylon eocenicum Wheeler, Scott, & Barghoorn, sp. nov. FIGURES 4—7. Growth rings. Present, distinct, 2-5 mm. wide. Vessel elements. Predominantly solitary, some pairs, angular in outline; gradually decreasing in size through the growth ring, with a tendency to form a band of pores at the beginning of the growth increment; tangential diameter of earlywood vessels 56-85 pm., mean 70 pm.; radial diameter 67-102 pm., mean 85 pm.; tangential diameter of latewood vessels 33-66 4 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 pm., Mean 47 pm.; 53 to 78 per square mm.; length 670-875 ,m.; per- foration plates scalariform, generally with more than 40 bars; inter- vascular pitting usually opposite but sometimes transitional and scalari- form; vessel-ray pits similar to intervascular pitting. Parenchyma. Not observed. Rays. 1 to 7 cells wide, multiseriates with a single marginal row of square and upright cells, a few homocellular rays; multiseriates 9 to 32 cells, 164-800 pm. high; uniseriates rare, composed of upright cells, 2 to 19 cells high. Imperforate tracheary elements. Viber tracheids with small bordered pits. MATERIAL. Two specimens of mature, silicified secondary xylem. HoLotryPe. U.S.G.S. Fossil Wood Collection No. D-2054B-28, measur- ing 52 & 68 &* 45 mm. Locality. Specimen Ridge, U.S.G.S. Paleobot. Loc. No. D-2054B. There are many plants which have wood with small angular vessels, scalariform perforation plates, and opposite and scalariform pitting. How- ever, the occurrence of these structural features in combination with multi- seriate rays more than three cells wide that often have a single marginal row of upright cells and rare uniseriate rays is not common. Such struc- tural features are characteristic of the Cyrillaceae. The wood described here also resembles Cornus, a genus that has been reported from the Yel- lowstone compression flora (Dorf, 1960) and from other Tertiary floras of the Rocky Mountains (Leopold & MacGinitie, 1972). Cornus differs from the wood described here and from Cyrilla as it has numerous uni- seriate rays and markedly heterocellular rays. Slides of six specimens of Cyrilla racemiflora were examined. As might be expected for such a wide-ranging species (southeastern United States to the Amazon), there is considerable structural variability. Maximum ray width ranges from four to eight cells, and parenchyma abundance and dis- tinctiveness of the growth rings varies. One specimen (Hw 19068) from the Amazon has ill-defined growth increments, while some specimens from the southeastern United States tend toward being semi-ring porous. Thomas (1960) earlier noted the absence of distinct growth rings in Cyrilla wood from the more tropical parts of the range of the genus. The fossils resemble those woods from the northern part of Cyrilla’s range as they tend toward semi-ring porosity and have distinct growth rings. There are two other genera in the Cyrillaceae, Purdiaea and Cliftonia. Purdiaea (two specimens examined) differs from the Yellowstone wood as it has relatively numerous uniseriate rays. Van der Burgh (1964) and we observed that rays in Cliftonia have more marginal rows of square and upright cells than do rays in Cyrilla. The rays of the Yellowstone wood resemble Cyrilla more than Cliftonia as they generally have but a single marginal row of square and upright cells. Cliftonia has less parenchyma 1978] WHEELER ET AL., FOSSIL WOODS, II 5 than Cyrilla. The amount and distribution of parenchyma in the fossil was difficult to assess because of the relatively poor preservation. It is possible that the Yellowstone wood may resemble Cliftonia more than Cyrilla by having scanty parenchyma. Consequently, as the fossils may be intermediate in structure between these two genera, we assign them to Cyrilloxylon. The two specimens assigned to this taxon, D-2054B-28 and -35, differ slightly. D-2054B-35 has wider growth rings, vessel diameter does not change as much throughout the growth ring, and its vessel ele- ments are slightly longer (up to 920 pm.), than D-2054B-28. Fossil wood of the Cyrillaceae has been reported from the Brandon Lignite of Vermont (Spackman, 1949) and the brown coals of the Netherlands (van der Burgh 1964, 1973). The Brandon Cyrilla has 20 to 25 bars per perforation plate, vessel elements 550-650 pm. long, and rays up to 5 cells wide with pronounced margins of upright cells. Cyril- loxylon europaeum van der Burgh lacks growth rings, its perforation plates have 15 to 30 bars, and its rays are up to 5 cells wide with 1 to 7 marginal rows of upright cells. This species appears intermediate between Cyrilla and Cliftonia as the abundant parenchyma is Cyrilla-like, but the rays are Cliftonia-like. The Cyrillaceae, a New World family of subtropical and warm-temperate areas, has three genera: Cyrilla, Cliftonia, and Purdiaea. Cyrilla, a mono- typic genus, grows in acid soils along the margins of swamps and streams and in wet pinelands. In this country, it occurs in the coastal pine belt from Florida to southeast Virginia to southeast Texas; it is also present in Mexico, British Honduras, the West Indies, and northern South America. Cliftonia has a more restricted distribution and occurs in the Coastal Plain of Georgia, Florida, Alabama, and Mississippi. Purdiaea is found in Venezuela, Colombia, Peru, Cuba, and the Isle of Pines (Thomas, 1960, 1961). There is no previous record of Cyrillaceae in the Rocky Mountains, as far as we know. FAGACEAE Quercinium amethystianum Wheeler, Scott, & Barghoorn, sp. nov. FicuREs 8, 9. Growth rings. Present, 2.5—-4.5 mm. wide, not always distinct. Vessel elements. Semi-ring porous; solitary, circular to oval in outline; in broad radial bands; tangential diameter of earlywood pores 110-200 pm., mean 140 pm.; radial diameter 165-265 um., mean 225 »m.; tangential diameter of latewood 33-94 »m., mean 64 pm.; length to 494 pm.; per- foration plates simple; intervascular pitting alternate; vessel—ray pits variable, typically large, and often vertically elongate; tyloses present. Parenchyma. Abundant, in 1- to 4-cell-wide bands, appearing wavy in cross section, particularly in the latewood; as isolated cells; and in the earlywood intermingled with tracheids forming conjunctive tissue between the vessels and rays, not swollen crystalliferous. 6 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 Rays. Rays of two distinct sizes, uniseriate (rarely biseriate), aggregate and compound; homocellular and heterocellular with marginal rows of square and upright cells, uniseriates 2 to 20 cells, 47-360 pm. high; large rays to 31 cells broad and up to 8 mm. high. Imperforate tracheary elements. Vasicentric tracheids with pitting on radial and tangential walls, and fiber tracheids with bordered pits. MATERIAL. One specimen of mature, silicified secondary xylem. Hootype. U.S.G.S. Fossil Wood Collection No. D-2054A-11, measuring 41 « 50 &* 120 mm. Locatity. Amethyst Mountain, U.S.G.S. Paleobot. Loc. No. D-2054A. Wood of certain evergreen oaks may be distinguished from deciduous oaks, as the wood of the former group is diffuse or semi-ring porous and has aggregate, rather than composite, rays in the mature wood (Williams, 1942; Brett, 1960). The wood of Lithocarpus is indistinguishable from the wood of such evergreen oaks. Brett (1960) emended the organ genus Quercinium Unger so that it would encompass species of fossil wood with evergreen oak-Lithocarpus structure. The specimen described here, be- cause it is semi-ring porous, is assigned to Quercinium. The name Quer- coxylon Krausel (1939) emend. Miiller-Stoll & Madel — is also used for fossil oak woods; however, Quercinium has pri Three species of fossil oak wood have been described on the Yellow- stone fossil forests: Quercus rubida Beyer (1954), Ouercinium knowltonii Felix (1896), and Q. lamarense Knowlton (1899). Quercus rubida Beyer is distinct as it is ring porous with an abrupt transition from a well-defined earlywood zone to the latewood zone, while the two species of Ouercinium ave evergreen oak-Lithocarpus structure. The description of the earliest-named species, Quercinium knowltonii Felix, is quite brief. It was described as having oval earlywood vessels with a tangential diameter of up to 210 »m., latewood vessels with a tan- gential diameter of 80-120 um., and parenchyma in short uniseriate bands associated with the pores. The only illustration accompanying the diagnosis of QO. knowltonii is a schematic drawing of a transverse section. A detailed comparison of Q. amethystianum with Q. knowltonii is not possible, but the parenchyma distribution in the two seems distinct and warrants the use of a different name for the specimen described above. An emended diagnosis of Q. lamarense follows, and differences between it and QO. ame- thystianum will be discussed following that diagnosis. jer Quercinium lamarense Knowlton, emend. Wheeler, Scott, & Barghoorn Ficures 10-12. Growth rings. Present, 2-10 mm. wide, not always distinct. Vessel elements. Semi-ring porous; solitary, oval in outline in earlywood, circular in outline in latewood; arranged in flamelike, radial pattern; tan- 1978] WHEELER ET AL., FOSSIL WOODS, I 7 gential diameter of earlywood pores 80-250 pm., mean 200 pm., radial di- ameter of earlywood pores 195-320 um., mean 240 pm.; tangential di- ameter of latewood pores 50-120 um., mean 90 pm.; length to 645 pm.; perforation plates simple; intervascular pitting alternate, pit-pairs circu- lar in outline; vessel-ray pits large, and often vertically elongate; tyloses present, but infrequent. Parenchyma. Predominantly apotracheal, in distinct bands 1 cell wide. at times 3 cells wide (within some growth rings these parenchyma bands more closely spaced in the earlywood), as isolated cells, and as broken uniseriate lines; also paratracheal; not swollen crystalliferous. Rays. Rays of two distinct sizes, uniseriate (occasionally partially bi- seriate) and aggregate; homocellular and less frequently heterocellular with marginal rows containing some square and upright cells; uniseriates 2 to 20 cells, 64-460 pm. high, numerous, 13-17 per mm.; aggregate rays to 1 mm. broad and 8 mm. high; units of aggregate rays, mostly 8-cell- wide rays. Imperforate tracheary elements. Vasicentric tracheids with pitting on radial and tangential walls, and fiber tracheids with bordered pits smaller and less frequent than on the vascular tracheids. MATERIAL. 12 specimens of mature, silicified secondary xylem. Hotortver. 2 slides: USNM 245721 a and USNM 245721 b. LocaLity. Specimen Ridge. One of the most common dicotyledonous woods in the collections from Specimen Ridge is this species of Quercinium. We have examined sections of seven specimens collected by Scott and an additional five specimens col- lected by Knowlton and Read that belong to the National Museum and the U.S. Geological Survey. The original diagnosis of Q. damarense Knowl- ton (1899) is quite brief and is accompanied by four line drawings. All the drawings are at a fairly high magnification showing in detail a few pores, but not showing overall pore or parenchyma distribution. Quanti- tative data were given only for the vessels. In his monograph on the Yellowstone fossil forests, Knowlton wrote that the material he described as Quercinium lamarense came from an upright stump measuring four feet in diameter and was collected on August 22, 1887. Among Knowlton’s collections housed at the National Museum, there are two slides (numbers 149 and 150) described in his notebooks as sec- tions from an upright stump measuring four feet in diameter collected on that date. As collection data are identical and these two slides are the only ones of oaks in Knowlton’s slide collection, it seems safe to assume these are the sections which Knowlton described and figured, and so we designate them the type of Quercinium lamarense Knowlton, emend. Wheeler, Scott, & Barghoorn. Knowlton acknowledged that Quercinium lamarense is similar to the 8 JOURNAL OF THE ARNOLD ARBORETUM [VoL, 59 earlier-described Q. knowltonii Felix (1896), but he suggested that the shape and size of the pores were different. It is possible that O. knowltonii and Q. damarense are different names applied to the same type of wood. However, as the original material Felix described was not available for study, we are using the name Q. /amarense for the woods we have ex- amined. Quercinium amethystianum has more parenchyma than Q. lamarense, and aggregate and compound rays rather than just aggregate rays as does QO. lamarense. Ring porous oak woods are among the most commonly described fossil woods, perhaps due in part to the ease of recognizing their distinctive wood structure. Reports of fossil woods resembling evergreen oaks and Lithocarpus are not as numerous. Lists of the occurrence of fossil oak woods have been prepared by Prakash and Barghoorn (1961a), Miiller- Stoll and Madel (1957), and Privé (1975). There are five species of fos- sil wood of the evergreen oak type known from the United States, all from the Tertiary of California. Four species, Ouercinium anomalum, 0; solerederi, Q. wardii, and Q. lesquereuxii, were described by Platen (1908). The fifth species, Quercus ricardensis Webber (1933), is from the Ricardo Pliocene. All five differ from Quercinium amethystianum and QO. lamarense in pore size and/or parenchyma abundance. Williams (1942) found in North American oaks a loose correlation be- tween crystal abundance in wood and dryness of habitat: generally the dryer the habitat, the more common the crystals. Neither Quercinium amethystianum nor Q. lamarense has crystalliferous parenchyma. Oak leaves are abundant in the Yellowstone compression flora (Dorf, 1960, 1964). No species of Quercus are reported from the Kisinger Lakes- Tipperary flora (MacGinitie, 1974), but there are oaks in the older Lost Cabinian Wind River flora and the younger Green River flora. JUGLANDACEAE Pterocaryoxylon knowltonii Wheeler, Scott, & Barghoorn, sp. nov. Figures 13-15. Growth rings. Distinct, 5-6.5 mm. wide, visible to unaided eye, demar- cated by 3 to 6 rows of flattened, thick-walled, imperforate tracheary ele- ents. Vessel elements. Semi-ring porous; mostly solitary or in radial multiples of 2 to 4; earlywood pores and pore multiples in diagonal arrangement: solitary pores oval in outline; latewood pores tending to be squared off, but not highly angular; tangential diameter of earlywood pores 160-310 um., mean 215 yum.; radial diameter of earlywood pores 225-400 p»m., mean 325 wm.; tangential diameter of latewood pores 70-140 pm., mean 118 uM. 3 radial diameter 110-190 pm., mean 165 pm.; length 460-850 pm.; per- foration plates simple; intervascular pitting alternate, pits 10-12 pm. across; vessel-ray pits not observed; tyloses present. 1978] WHEELER ET AL., FOSSIL WOODS, II 9 Parenchyma. Apotracheal, as short uniseriate lines in cross section, scanty in earlywood, more frequent in latewood with lines of greater tan- gential extent more closely spaced. Rays. Multiseriates to 3 cells wide, 5 to 26 cells, 155-515 um. high; with uniseriate margins of 1 to 8 (mostly 1 to 3) rows of square and up- right cells; uniseriates 3 to 14 cells, 105-385 wm. high, cells of uniseriate rays similar in shape to uniseriate margins of multiseriate rays; 6 to 9 per mm. Imperforate tracheary elements. Pitting not observed. MATERIAL. One specimen of mature, silicified secondary xylem. Hotrotyre. U.S.G.S. Fossil Wood Collection No. D-2054B-37, mea- suring 35 & 24 & 41 mm. Locarity. Specimen Ridge, U.S.G.S. Paleobot. Loc. No. D-2054B. The fossil resembles the wood of Pterocarya and Juglans sections CAR- DIOCARYON (oriental butternuts) and TRACHYCARYON (J. cinerea, the American butternut). Wood of the black walnuts, both tropical and tem- perate, is distinguished by ray cell shape, the presence of crystalliferous parenchyma strands, and the reticulate thickenings in the vessel elements (Miller, 1976a, 1976b). The rays in the fossil are three cells wide. Rays only two cells wide were observed in the five specimens of Pterocarya rhoifolia; three-cell- wide rays occur in Juglans. Kribs (1927) noted rays are only one to two cells wide in Pterocarya. We examined four species of Péerocarya in the Harvard Wood Collection and observed some occasional three-cell-wide rays in one preparation of P. paliurus. Miiller-Stoll and Madel (1960) re- ported these rays in Pterocarya caucasia C. A. Meyer. The three-cell-wide rays of the Yellowstone fossil are more suggestive of Juglans than Ptero- carya, as such rays are rare in the former. However, there are ten species of Pterocarya (Willis, 1973), and examination of all species is necessary to determine if ray width is a useful feature for separating Juglans and Pterocarya. One of the distinctive features of the fossil is the pronounced diagonal arrangement of the earlywood pores. Drs. Regis Miller and B. F. Ku- kachka of U.S.D.A. Forest Products examined thin-sections of the fossil and compared then to sections of Juglans and five specimens of Ptero- carya rhoifolia. Drs. Miller and Kukachka did not observe diagonal pore arrangement in any extant species of Juglans, but did in the five speci- mens of P. rhoifolia. Diagonal pore arrangement also occurs in P. paliurus. Specimens of the other eight species of Pterocarya were not available for study. Drs. Miller and Kukachka suggest that the fossil resembles Ptero- carya more than Juglans, a suggestion with which we concur. Miiller-Stoll and Miidel (1960) established the genus Pterocaryoxylon for juglandaceous woods with the following characteristics: diffuse or semi- ring porous; vessels thin-walled, solitary, and in radial multiples of 2 to 4; 10 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 simple perforation plates; alternate intervascular pitting; parenchyma in one-cell-wide, short, irregularly spaced bands, and 1 to 3 seriate rays. As the Yellowstone wood exhibits these characteristics common to both Jug- lans and Pterocarya, we are naming it as a new species of this genus. No fossil wood resembling Pterocarya has been previously described from the Tertiary of North America, although Pterocarya is among the genera of woods listed as occurring in the Miocene Vantage, Washington, fossil forests (Beck, 1945; Prakash, 1968). Only one species of fossil wood of Juglans has been described from North America. Juglans fryxellii Prakash Barghoorn (1961b) from the Miocene Vantage forests resembles the extant black walnuts and is distinct from Pterocaryoxylon as it has crystal- liferous parenchyma, smaller vessels, and no diagonal pore arrangement. To date, five species of Pterocaryvoxylon have been described. These are P. pannonicum from the Pliocene of Hungary, P. honshouense (originally called Pterocarya rhoifolia by Watari in 1952), P. chinense from the Ter- tiary of China, P. pilinyense from the Tertiary (Helvetian) of Hungary, and P. subpannonicum from the Pliocene of France (Privé, 1974). Ptero- caryoxylon honshouense and P. chinense both have rays only 1 to 2 cells wide. Pterocaryoxylon honshouense also is characterized by pores in di- agonals and so appears structurally similar to Pterocarya. Pterocaryoxylon pilinyense Greguss also has 1- to 2-cell-wide rays, but it has crystalliferous parenchyma, a feature not associated with Pterocarya or the butternuts. Pterocaryoxylon pannonicum and P. subpannonicum have rays to 3 cells wide, but do not show any evidence of diagonal pore arrangement. Woods assigned other generic names, but which might be transferred to Ptero- caryoxylon include Juglandinium caucasicum Gaivoronsky from the Oligo- cene of the U.S.S.R., J. fasseewii Naschokin from the Tertiary of the USS.R., and Juglandoxylon sp. Petrescu & Nuta from the Miocene of Ru- mania (Privé, 1974 Both Juglans and Pterocarya are members of the Kisinger Lakes flora (\acGinitie, 1974), and both occur at many Tertiary localities in the Rocky Mountains. The history of Pterocarya in the Rocky Mountain re- gion was summarized by Leopold and MacGinitie (1972) pe — mH MyRICACEAE Myrica absarokensis Wheeler, Scott, & Barghoorn, sp. nov. FIGuRES 16-19, Growth rings. Present, distinct. Vessel elements. Diffuse porous: Dea aaa solitary, some pairs tangential diameter 30-85 ym. (mean 50 pm.) ; radial diameter 47-118 ee mean 75 wm.; 65-110 per square mm.; length 505-735 «m.: perforation plates all scalariform with 7 to 16 bars: Intewascul: ar pitting predominant- ly opposite, some transitional, vessel-ray pits similar to intervascular pit- ting 1978 | WHEELER ET AL., FOSSIL WOODS, II 11 Parenchyma. Apotracheal diffuse, as isolated strands, often inflated and crystalliferous. Rays. Multiseriates to 4 cells wide; 4 to 25 cells, 80-600 pm. high; with uniseriate margins of 1 to 9 rows of square and upright cells; uniseriates 2 to 8 cells, 92-350 pm. high, apparently composed of only square and up- right cells; 8 to 12 per mm. Imperforate tracheary elements. Fibers with bordered pits on both radial and tangential walls. MATERIAL. Five specimens of mature secondary xylem. Ho otyer. U.S.G.S. Fossil Wood Collection No. D-2054B-11, measuring 75 < 59 & 50 mm. Locatity. Specimen Ridge, U.S.G.S. Paleobot. Loc. No. D-2054B. There are five specimens assigned to this taxon (D2054B-11, -21, -23, -25, and -34). There is some variation in quantitative characters, particularly vessel density, but this sort of variation is not considered taxonomically significant. The minimum tangential diameter of a vessel element is 32 pm. in all specimens, the maximum tangential diameter varies from 60 to 87 pm., the mean varies from 46 to 57 pm.; vessel density ranges from 50 to 200 per square mm.; and the number of bars per perforation plate ranges from 6 to 20, although in all specimens it is most frequently 12. Vessel elements up to 1200 pm. long occur in some specimens. The combination of characters of these woods (small solitary vessels, scalariform perforation plates, opposite intervascular pitting, and apotra- cheal diffuse parenchyma) occurs in many unrelated families, of which only four have an average vessel element length of less than 800 pm. These are the Hamamelidaceae, Cornaceae, Cyrillaceae, and Myricaceae. In the Hamamelidaceae, vessel-ray pitting is typically large and elon- gate, and intervascular pitting is predominantly scalariform. The wood of the Cyrillaceae is characterized by perforation plates of 30 to 50 bars and few uniseriate rays; that of the Cornaceae, by perforation plates with many (more than 20) bars, multiseriate rays frequently higher than 1 mm., and no crystalliferous parenchyma. Metcalfe and Chalk (1950) report that Kaliphora and Torricellia have simple perforations, but Willis (1973) places Torricellia in its own family and notes that Kaliphora, which is in- digenous to Madagascar, should probably be excluded from the Cornaceae. As the Yellowstone woods have small vessel-ray pitting, opposite inter- vascular pitting, less than 20 bars per perforation plate, low rays, an inflated crystalliferous parenchyma strands, they differ from members of these families. The structural features of these woods match those of the Myricaceae, and they are assigned to that family. Slides of the wood of eight species of Myrica were examined. Of these species, the Yellowstone woods most closely resemble M. cerifera and M. rubra, as both have inflated crystalliferous axial parenchyma and pre- dominantly scalariform perforation plates. Some species of M-yrica do have ig JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 mostly or exclusively simple perforation plates and lack crystalliferous parenchyma. Myricoxylon hungaricum Miiller-Stoll & Midel (1962) from the Miocene of Hungary differs from the Yellowstone wood as this Hungarian fossil wood lacks growth rings and has predominantly simple perforation plates and a higher ray density. Miuller-Stoll and Middel also noted the structural similarity of A/yrica wood to wood of the Cornaceae and Hamamelidaceae and used number of bars per perforation plate and nature of vessel to ray parenchyma pitting to distinguish between these families. Kruse (1954) described an Eocene fossil wood and referred it to the genus Myrica. This wood from Eden Valley, Wyoming, was described as having perforations with up to 35 to 40 bars, a larger number of bars than is characteristic of the extant Myricaceae. We prepared new sections of the type of Myrica scalariformis Kruse and observed the number of bars per perforation plate to range from 15 to 32, most commonly numbering 20 to 22 bars. This is more consistent with the anatomy of extant Myrica species. A/yrica scalariformis differs from the Yellowstone wood as there are more bars per perforation plate, no crystalliferous axial parenchyma, a lower vessel density, and enlarged cells in the rays and the axial paren- chyma. We did not observe enlarged parenchyma cells in any of the ex- tant species of A/yrica we examined, and such cells are not mentioned in descriptions of the anatomy of the family (Metcalfe & Chalk, 1950: Rec- ord & Hess, 1943). It is possible that the Eden Valley wood has been in- correctly assigned to Myrica. There are numerous reports of putative Myrica leaves from Cretaceous and early Tertiary localities in North America. Chourey (1974) examined these reports in detail and found that most of these fossils cannot be con- sidered to be Myrica. Identification of a leaf as Mvyrica requires use of venation pattern, leaf morphology, and cuticular features. As most of the putative J/yrica species are represented by specimens that are either frag- mentary, poorly preserved, or without cuticle, it is not possible to deter- mine their affinities. Chourey suggests that, to date, there have been only eight occurrences of megafossils which are probably Myricaceae. Berry described many Mvyrica leaves from the Eocene of southeastern United States. His collections and new collections from this area were examined with especial care, but no M/yrica leaves were found in either the original collections or the new ones. Pollen of the Myricaceae is reported from a few North American Ter- tiary localities. According to Chourey, “. . . not until Eocene times is the family Myricaceae well established . . . The limited record of Myrica ollen, however, might also show a more authentic record of the evolution and distribution of M/yrica throughout the Cretaceous and Tertiary than is indicated by common overuse of the generic name for fossil leaves.” Wolfe (1973) noted that the history of the family is uncertain as its pol- len is similar to other amentiferous families. The Yellowstone wood represents one of the oldest occurrences of the family and the genus. It has primitive wood structure for the genus as it jean 1978 | WHEELER ET AL., FOSSIL WOODS, II 13 has exclusively scalariform perforation plates. Today Myrica has a nearly cosmopolitan distribution, but does not oc- cur naturally in northern Africa, central and southeastern Europe, Aus- tralia, and southwest Asia (Willis, 1973). Mvyrica cerifera, a species which occupies a wide variety of habitats, has a remarkably wide range and grows in the southeastern United States, the larger islands of the West Indies, and parts of Mexico and Central America. Myrica rubra is a mem- ber of the mixed mesophytic and evergreen sclerophyllous broad-leaved forest associations of China (Wang, 1961). NyYSSACEAE Nyssa saximontana Wheeler, Scott, & Barghoorn, sp. nov. FIGuRES 20-23. Growth rings. Present, distinct, 2-5.5 mm. wide. Vessel elements. Diffuse porous; solitary and in radial multiples of 2 to 4, and in radial chains with a single fiber separating the vessels; tangen- tial diameter 55-110 p»pm., mean 83 pm.; radial diameter 60-115 pm., mean 97 pm.; 35-57 per square mm.; length 700-1400 pm., perforation plates scalariform with more than 20 bars; intervascular pitting opposite to tran- sitional, pits often square or rectangular in outline; vessel-ray pitting similar to intervascular pitting. Parenchyma. Seen only in longitudinal sections as isolated strands, ap- parently apotracheal diffuse. Rays. Multiseriates to 2 (rarely 4) cells wide; 8 to 32 cells, 440-1500 ym. high, markedly heterocellular with 1 to 18 marginal rows of square and upright cells; some rays composite with uniseriate bridges between multiseriate portions of ray; uniseriates 1 to 10 cells, 155-679 pm. high; 8 to 11 per mm. Imperforate tracheary elements. Pitting not observed. MATERIAL. One specimen of mature secondary xylem. Hototyre. U.S.G.S. Fossil Wood Collection No. D-2054B-41 measur- ing 34 & 22 & 35 mm. Locatity. Specimen Ridge, U.S.G.S. Paleobot. Loc. No. D-2054B. According to van der Burgh (1964, p. 285), crowded opposite inter- vascular pits which are somewhat squared in outline are diagnostic of Nyssa. This type of pitting occurs in the fossil. The combination of char- acters of the fossil is diagnostic of the Nyssaceae as determined by a sur- vey of the descriptions of the wood anatomy of different families (Met- calfe & Chalk, 1950). There is also some resemblance to /lex. A few species of Jlex do have crowded opposite pits which are square in outline. In most species of /lex there are fewer rows of marginal upright cells, and 14 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 the rays are generally wider. One of the characteristics of the fossil is vertical fusion of rays; this feature is found in Nyssa (Record & Hess, 1943; and personal observation) and is rare in /lex. Slides of the wood of both Nyssa and /lex and the descriptions of //ex species in Baas’s (1973) publication were compared with the Yellowstone wood. Wood of Nyssa javanica (Hw 16849) and Nyssa sinensis (Hw 21459) agrees in every detail with wood of the fossil. Vessel element distribution, density, and size; pitting type; ray structure; and parenchyma distribution are iden- tica The genus Vyssoxylon was established for those woods resembling .V yssa and Davidia. Camptotheca can be distinguished from these two genera as it has perforation plates with less than 20 bars and very few pore multiples. .Vvssa and Davidia closely resemble one another structurally. Comparing the two genera, Metcalfe and Chalk (1950) found that Da- vidia tends to have a larger number of bars per perforation plate, more scalariform pitting, less parenchyma, a higher ray density, and fibers with more conspicuously bordered pits than Nyssa. We examined two slides of . involucrata and found that the vessels were more angular and there were fewer pore multiples than were found in the fossil. Five species of fossil woods whose affinities are suggested to be with Nyssa have been described. These are: Nyssoxvlon japonicum Méadel (1959) from the Tertiary of Japan; Nyssa eydei Prakash & Barghoorn (1961b) from the Miocene of Washington; Nyssoxylon haanradense van der Burgh (1964) from the Netherlands brown coal; N. romanicum Petrescu (1970) from the Oligocene of Rumania; and JN. ishikariense Suzuki (1975) from the Eocene of Japan. Nyssoxylon japonicum differs from the Yellowstone wood as the former has a higher vessel density, a greater number of bars per perforation plate, and rays only up to two cells wide. Nyssa evdei is distinguished from the Yellowstone wood as the former has a lower vessel density, more parenchyma, longer vessel ele- ments, taller rays, a higher ray density, and crystalliferous parenchyma. Nyssoxylon haanradense is represented by one poorly preserved specimen, and the features seen in cross section are not described; unlike the Yel- lowstone wood, however, it does have numerous crystalliferous parenchyma strands and its rays are rarely three cells wide. The most recently described species, V. ishikariense Suzuki, has smaller vessels, a higher vessel density, and narrower and more frequent rays than the Yellowstone wood. Fruits, leaves, and/or pollen of Nyssa have been reported from localities in Europe, Asia, and North America ranging in age from Paleocene to Pliocene (Eyde & Barghoorn, 1963). Most of the leaves from North American sites have been compared to the extant species of eastern North America, particularly the Nyssa sylvatica complex. Some of the fruits, notably Nyssa brandoniana Eyde & Barghoorn (1963) and Paleonyssa spatulata Scott (1954), seem more closely to resemble Nyssa javanica (Eyde & Barghoorn, 1963). Brown (1962) listed two species of Nyssa leaves at a number of Paleocene localities in the Rocky Mountains and 1978] WHEELER ET AL., FOSSIL WOODS, II 15 High Plains. A Nyssa fruit resembling the extant V. ogeche was reported from the Metzel Ranch flora of Montana (Becker, 1972). At present, Vyssa is native to eastern North America and eastern Asia, with VV. sylvatica occurring in Mexico at altitudes of 3000 to 5000 feet (Eyde, 1963). Nyssa sinensis grows in the mixed mesophytic forest and evergreen oak forests of China (Wang, 1961). ROSACEAE Prunus gummosa (Platen) Wheeler, Scott, & Barghoorn, comb. nov. IGURES 24-27. Pruninium gummosum Platen, Naturf. Gesell. Leipzig Sitzungsber. 34: 122- Growth rings. Present, distinct, 1-7 mm. wide. Vessel elements. Diffuse porous; solitary, in clusters, or in radial multi- ples of 2 to 4; tangential diameter 32-74 pm., mean 47 pm.,; radial di- ameter of solitary pores 78-98 pm., mean 90 pm.; length to 370 pm.; spiral thickenings commonly present; perforation plates simple; inter- vascular pitting alternate, minute to small, pit-pairs round, rarely touching; vessel-ray pits similar to intervascular pitting; tyloses occasionally present. Parenchyma. Very rare, as isolated strands, generally of 4 cells, seen only in longitudinal sections, apparently apotracheal diffuse. Rays. Multiseriates to 4 cells wide; 5 to 44 cells, 105-763 pm. high; with uniseriate margins of 1 to 6 (most frequently 1 to 2) rows of square and upright cells; uniseriates 1 to 10 cells, 46-285 pm. high, some com- posed entirely of square and upright cells, others with some procumbent cells, in tangential section cells of uniseriate rays oval in outline; 8-1 per mm. Imperforate tracheary elements. Fibers with circular bordered pits on both radial and tangential walls. pits approximately 5 jm. across. Gum ducts. Traumatic vertical gum ducts present, often in tangential groups at beginning of growth increment. MATERIAL. One specimen, D-2054A-21, of mature secondary xylem, measuring 63 & 71 * 41 mm. Locatiry. Amethyst Mountain, U.S.G.S. Paleobot. Loc. No. D-2054A. The combination of characters of this fossil (small vessels, small to minute alternate intervascular pitting, prominent spiral thickenings, and traumatic vertical canals) is diagnostic of the Rosaceae, exclusive of the Chrysobalanoideae. Pore multiples are characteristic of the Prunoideae (Metcalfe & Chalk, 1950), and Prunus is particularly prone to the forma- tion of traumatic vertical canals (Jane, 1970; Panshin & DeZeeuw, 1970; Record & Hess, 1943). As pore multiples and vertical canals are present. 16 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 the specimen is assigned to the Prunoideae. Vertical canals also occur in Laurocerasus of the Prunoideae, and this genus (a segregate of Prunus) is similar anatomically to Prunus. Laurocerasus may be distinguished from Prunus as, in the former, rays frequently have four or more marginal rows and parenchyma is vasicentric (Metcalfe & Chalk, 1950). The Yel- lowstone wood resembles Prunus as it has sparse diffuse parenchyma and generally less than four marginal rows of ray cells. Pruninium gummosum from Amethyst Mountain was described by Platen in 1908. The differences between his diagnosis and the specimen described here are slight. The quantitative data are nearly identical: ac- cording to Platen, maximum tangential diameter is 75 pm., we found it to be 74 um.; he described rays as up to 50 cells high, we found a maximum height of 44 cells. We did not observe any scalariform perforation plates, although he noted their rare occurrence. The diagnosis of Pruninium gum- mosum is so close to the specimen we studied that it almost seems as if we had fragments from the same stump. There is less variation than may be observed in wood taken from different portions of the same tree. We consider Pruninium gummosum Platen and the specimen described here to be conspecific, and their anatomies are indistinguishable from species of the extant genus Prunus. As discussed in the first paper on Yellow- stone fossil dicotyledonous woods, it is not required that species of fossil wood be assigned to organ genera when their anatomical features are those of a single modern genus and such an assignment obscures the contribu- tions of paleobotany to the history of a genus. Consequently, we hereby transfer the species Pruninium gummosum Platen to Prunus and make the new combination Prunus gummosa (Platen) Wheeler, Scott, & Barghoorn. Reports of fossil woods of the Rosaceae exclusive of the Chrysobalanoi- deae are few. Prunus (?) (Beck 1673) is on a list of woods of the Miocene Vantage fossil forests (Beck, 1945). Thin-sections of this wood (Beck 1673, HPC 56667) were examined. This specimen appears to be Ulmus pacifica Prakash & Barghoorn (1961a) rather than Prunus. There are no other descriptions of fossil wood of the Prunoideae. Fossil woods assigned to the Rosaceae include Lyonothamnoxylon Page (1964) from the Pliocene of Nevada, Rosaceoxylon Shilkina (1958) from the late Tertiary of Rus- sia, Maloidoxylon from the Tertiary of France (Grambast-Fessard, 1966) and the Miocene of Colorado (Wheeler & Matten, 1977), and Pomoxy- fon Hofmann (1944). The last is not a valid name as no species name was assigned nor was a species described when the genus was first named. Prunus is a cosmopolitan genus of some 430 species (Willis, 1973); Baas (1973) has found that spiral thickenings are characteristic of the temperate members of the genus. Spiral thickenings are prominent in the specimen described here. Both megafossils and microfossils of Prunus occur in the Kisinger Lakes- Tipperary flora. MacGinitie (1974) considered the leaves similar to those of the extant Prunus serotina Ehrlich. Species of Prunus have also been described from Florissant (MacGinitie, 1953) and the Green River flora 1978] WHEELER ET AL., FOSSIL WOODS, II 1 (MacGinitie, 1969), and reported from the older Wind River flora (Mac- Ginitie, 1969). STAPHYLEACEAE Turpinia lamarense Wheeler, Scott, & Barghoorn, sp. nov. FIGURES 28-32. Growth rings. Indistinct. Vessel elements. Solitary and in radial multiples of 2 to 8, mostly 2 to 4: square to rectangular in outline; tangential diameter 55-115 »m., mean 85 um.; radial diameter 78-155 pm., mean 124 wm.; 14 to 22 per square mm.; length .76-1.47 mm.; perforation plates exclusively scalariform with 26 to 51 bars; intervascular pitting subopposite, opposite, and transitional, often scalariform at ends of vessel elements; vessel—ray pits similar to inter- vascular pitting. Parenchyma. Exact distribution not determinable, isolated strands seen in longitudinal section, some adjacent to vessel elements. Rays. Multiseriates to 4 cells wide, occasionally 5 cells wide; 11 to ey cells, 415-2440 um. high; markedly heterocellular with uniseriate margins of up to 16 rows of upright and square cells; vertical ray fusions; uni- seriates 2 to 25 cells, 127-1840 um. high, apparently composed of only square and upright cells; occasionally crystals in slightly enlarged ray cells; 7 to 10 per mm. Imperforate elements. Some fibers with bordered pits. MATERIAL. One specimen of mature, silicified secondary xylem. HototyPe. U.S.G.S. Fossil Wood Collection No. D-2054B-24, measur- ing 77 & 27 * 93 mm. Locatiry. Specimen Ridge, U.S.G.S. Paleobot. Loc. No. D-2054B. Small vessels that are both solitary and in radial multiples, opposite intervascular pitting, vessel—ray pitting that is round to elongate, exclusive- ly scalariform perforation plates with more than 20 bars, average vessel element length greater than 800 p»m., markedly heterocellular rays, and sparse parenchyma are all diagnostic of the Staphyleaceae. The fossil somewhat resembles members of the Styracaceae, but in this family per- foration plates generally have fewer than 20 bars, and the vessel—ray pit- ting is very fine. Turpinia is the only genus in the Staphyleaceae with pores in radial multiples and more than ten rows of marginal upright cells, so the fossil is assigned to this genus. According to Willis (1973), Turpinia is a genus comprised of 30 to 40 species occurring from Ceylon to Japan, Malaysia, and Central and trop- ical South America. We examined only six of the species. The Old World species examined have predominantly solitary vessels, a fairly high vessel 18 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 density, thick-walled fibers, and elongate vessel-ray pitting. The New World species examined have numerous radial multiples, a relatively low vessel density, thin-walled fibers, and vessel-ray pits that have a round out- line and are not as coarse as those of the Old World species. The structural features of Zurpinia absarokensis are more consistent with those of the New World species of the genus, at least those available to us for com- parison. e Kisinger Lakes flora (MacGinitie, 1974) does not contain any members of the Staphyleaceae. Turpinia appears on a preliminary generic list of the Lost Cabinian Wind River flora of late early Eocene age (MacGinitie, 1969), but we know of no other reports of the genus in the Tertiary of the Rocky Mountains nor of any reports of fossil wood of Tur pinia. Fossil seeds of Turpinia have recently been recognized in the Oligocene Brandon Lignite of Vermont (Tiffney, 1977). The Brandon Turpinia uliginosa most closely resembles the New World representatives of the genus. Tiffney based this comparison on the examination of 20 ex- tant species. pu ULMACEAE Zelkovoxylon occidentale Wheeler, Scott, & Barghoorn, sp. nov. FIGURES 33-37. Growth rings. Distinct. Vessel elements. Semi-ring porous; transition from earlywood to late- wood gradual; tangential diameter of earlywood vessels 83-124 pm., mean 112 »m.; radial diameter 125-190 ym., mean 165 um.; solitary and in radial multiples of 2 to 4, circular to oval in outline; latewood vessels tangential diameter 36-70 pm., mean 50 um., in clusters of 3 to 20 cells and more rarely in radial multiples of 2 to 6, often tending to form tan- gential or oblique bands; perforation plates simple; intervascular pitting crowded alternate, hexagonal in outline due to crowding; spiral thickenings in the smaller elements. Parenchyma. Paratracheal, vasicentric, or not completely ensheathing vessels or vessel clusters, most commonly 4 cells per strand. Rays. Multiseriates to 4 (rarely 5) cells wide, 9 to 52 cells, 170-860 pm. high; with 1 to 3 marginal rows of square and upright cells, frequent- ly inflated crystalliferous cells in marginal rows and along side of rays, some homocellular rays also; uniseriate rays few, low, 1 to 4 cells high. Imperforate tracheary elements. Libriform fibers and vascular tracheids, vascular tracheids with spiral thickenings and crowded alternate pits, gen- erally associated with small vessel elements. MATERIAL. One specimen of mature, silicified secondary xylem. HoLotyre. U.S.G.S. Fossil Wood Collection No. D-2054B-26, measur- ing 44 & 34 & 41 mm. Locality. Specimen Ridge, U.S.G.S. Paleobot. Loc. No. D-2054B. 1978] WHEELER ET AL., FOSSIL WOODS, I 19 The structural features of this fossil, particularly latewood pore distri- bution and spiral thickenings in the smaller tracheary elements, clearly indi- cate its affinities with the Ulmaceae. Genera in the Ulmaceae with some degree of ring porosity and heterocellular rays are Celtis, Zelkova, and Hemiptelea. Hemiptelea is readily distinguishable as it has scalariform perforation plates, few radial multiples, and storied vessels. The Yellow- stone wood shares features with both Zelkova and Celtis. Inflated crystal- liferous ray cells are characteristic of Zelkova. Ray cells of Celtis often contain crystals, but the cells are not enlarged as they are in Zelkova and the fossil. Rays in Celtis are predominantly heterocellular; those in Zel- kova are predominantly homocellular. Wood of extant Zelkova species is strictly ring porous, the transition from earlywood to latewood is very abrupt, and there is a distinct band of large earlywood pores. Radial mul- tiples are very rare in Zelkova, occurring in only 0 to 8 percent of the pores (Sweitzer, 1971). Celtis is an anatomically diverse genus with both diffuse and ring porous species. The abruptness of the transition from earlywood to latewood may vary within a species, probably reflecting differences in habitat. Five of six specimens of C. occidentalis had an abrupt transition from earlywood to latewood, while one was semi-ring porous with a gradual transition in size and distribution of the pores. Sweitzer (1971) studied three specimens of Planera: two were diffuse porous, the third ring porous. The arrange- ment of the pores, in addition to the size, differs between ring and diffuse porous species: radial multiples are common in the diffuse porous species of Celtis, rare in ring porous species. Sweitzer (1971) suggested that ring porosity and spirals should be con- sidered as specializations in Ulmaceae which are correlated with a sea- sonal temperate environment. Cox (1941) studied species of Celtis from temperate and tropical regions and concluded that species of tropical zones are diffuse porous and lack spiral thickenings, while those of temperate regions are ring porous and have spiral thickenings. Spiral thickenings are present in the small tracheary elements of the Yellowstone fossil and it is intermediate in the development of the ring porous condition. The earliest known ring porous woods are Eocene (Chowdhury, 1964). The known early Eocene floras of North America indicate that rainfall was evenly distributed throughout the year. MacGinitie (1974) suggested that by the early middle Eocene, the climate of the Rocky Mountains was sea- sonal. The development of ring porosity may be correlated with the ad- vent of a seasonal climate The fossil wood described here resembles Celtis as both have similar pore distribution, but the fossil wood has a character (inflated ray cells) that occurs only in Zelkova, a genus with ring porous wood. Celtis is not common in the fossil record of the Rocky Mountains, while Zelkova is common. Celtis mccoshii Lesquereux is one of the least com- mon fossils in the Eocene Green River flora and is also rare in the Oli- gocene Florissant flora (MacGinitie, 1969, 1953). Zelkova is a member of the Kisinger Lakes flora (MacGinitie, 1974). One may speculate that this 20 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 wood might represent wood of a tree which had foliage recognizable as Zelkova, but whose wood had not yet developed the presumably derived ring porous condition. The wood does share more features with Celtis than with Zelkova, and it is difficult to assess how much weight should be at- tached to a single distinctive feature, such as these enlarged ray cells A fossil wood that is very similar to the Yellowstone wood is Zelkovoxy- lon dacicum Petrescu (1971) from the Oligocene of Rumania. It has very inflated crystalliferous ray cells, like extant Zelkova and the Yellowstone specimen. It also has vessels in radial multiples, heterocellular rays, and is semi-ring porous, all features present in the Yellowstone wood, but which neither Sweitzer nor we found to be characteristic of Zelkove. Petrescu seems to have weighted the one character, inflated crystalliferous cells, in determining the affinities of the Rumanian woo We do not consider the Yellowstone wood to be structurally equivalent to extant Zelkova, but for the three reasons listed below we name it a species of Zelkov ioe: implying some, but not complete, oo to extant Zelkova. First, ring porosity in the Ulmaceae is, at times, under environmental control, and pore distribution is affected by the nee to which the ring porous condition is expressed. Pore distribution is variable within a genus because of this. Second, the Yellowstone wood has inflated crystalliferous ray cells, a feature unique to Zelkova. Third, there is a precedent for assigning the name Zelkovoexylon to woods with the com- bination of features found in the fossil. There are three species of fossil wood which are very similar to Zelkova, as they are ring porous with a narrow, well-defined earlywood pore zone. These are Zelkova zelkoviformis (Watari) Watari (1941, 1952) and Z. wakimizui (Watari) Watari (1948, 1952) from the Miocene of Japan, and Zelkovoxylon yatsenkokhmelvskyi Greguss (1969) from the Miocene of Hungary. All three are younger than the two semi-ring porous species of Zelkovoxylon. At present, the genus Ze/kova is comprised of six or seven species which are distributed in the eastern Mediterranean, the Caucasus, and eastern Asia (Willis, 1973). — id Wu DISCUSSION The ten species of fossil wood described in this paper represent a combi- nation of so-called temperate and paratropical forms, as do the eight species described in our first paper (Wheeler, Scott, & Barghoorn, 1977). Two of the woods, Cyrilloxylon eocenicum and Prunus gummosa, have structural features (semi-ring porosity and spiral thickenings, respectively) which are generally found in the extant representatives that grow at high- er latitudes, In contrast, Turpinia lamarense has indistinct growth rings, a character considered to be indicative of more tropical conditions. Zelkovoxylon occidentale, a wood that is intermediate in structure be- tween two well-defined extant genera (Zelkova and Celtis), differs from the present-day Zelkova in degree of ring porosity and pore distribution, the 1978 | WHEELER ET AL., FOSSIL WOODS, II rat latter a feature affected by the degree of ring porosity. The ring porous condition in the Ulmaceae is believed to be a specialization developed in response to a seasonal temperate environment. The absence of ring po- rosity in this specimen may reflect the absence of such an environment. This is consistent with the subtropical climatic conditions already inferred for the Yellowstone fossil forests (Dorf, 1960, 1964) and the near-tropical climate of the nearby Kisinger Lakes flora of similar age to the Yellow- stone flora (MacGinitie, 1974) MacGinitie (1974) has commented on the lack of similarity of the Yel- lowstone leaf and Kisinger Lakes floras. The Kisinger Lakes flora is ap- proximately 80 miles southeast of the Yellowstone flora, and the two are of similar age. The Kisinger Lakes flora has a “distinct tropical American aspect” and is apparently unlike the older, early Eocene floras of the Rocky Mountains which appear to have relationships to the evergreen broad-leaved and mixed mesophytic forest types. In contrast, the Yellow- stone fossil flora, as presently known, does not appear to have a strong relationship to the Neotropics. None of the 18 species of wood we have described to date indicates such affinities, and many (Quercinium, Ptero- caryoxylon, Myrica, Nyssa, and Prunus) are related to extant genera which grow in the evergreen sclerophyllous broad-leaved and/or the mixed meso- phytic forest types. More detailed comparison of the Yellowstone flora with other Tertiary floras of the Rocky Mountains will be reserved for our third and final paper. ACKNOWLEDGMENTS We thank the National Park Service and the Chief Ranger Naturalist, Yellowstone National Park, for permission to collect fossil woods in the Park: Dr. R. A. Davis for the loan of materials from the University of Cincinnati Paleobotanical Collections; Dr. Francis Hueber of the Smith- sonian Institution for assistance with Knowlton’s collections; Dr. Francis Kukachka of the U.S.D.A. Forest Products Laboratory for examining the material of Pterocaryoxylon and Turpinia; Dr. Regis Miller of the U.S.D.A. Forest Products Laboratory for reading the manuscript and examining slides of Pterocaryoxylon and Turpinia; and Dr. Virginia Page of Stanford University for reading the manuscript. Much of this study was made while the senior author was engaged on a part-time basis reorganizing the Harvard Wood Collection. Support for that project came from National Science Foundation Grant DEB72-02053 A04 (Reed C. Rollins, Principal Investigator). BIBLIOGRAPHY AwastuHt, N. 1965. Fossil woods of Anacardiaceae from the Tertiary of south India. Palaeobotanist 14: 131-143. Baas, P. 1973. The wood anatomical range in /lex (Aquifoliaceae) and its eco- logical and phylogenetic significance. Blumea 21: 193-258. a7 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 os i = 1945. Ancient forest trees of the sagebrush area in central Wash- Jour. Forestry 43: 334-338, pores H F. 1972. The Metzel Ranch flora of the eg Ruby River Basin, southwestern Montana. Palaeontographica 141B: 1-61. Beyer, A. F. 1954. Some petrified woods from the Specimen Ridge area of Yellowstone National Park. Am. Midl. Nat. 51: —567. Brett, D. W. 1960. Fossil oak wood from the British an Palaeontology 3: 86-92 1966. Fossil wood of Anacardiaceae from the British Eocene. Jbid. 9: 360-364 Brizicky, G. K. 1962. The genera of Anacardiaceae in the southeastern United States. Jour. Arnold Arb. 43: 359-375. Brown, R. W. 1962. Paleocene flora | oy ae Mountains and Great Plains. . 5. Geol. Surv. Prof. Pap. 375: Burcu, J. VAN DER. 1964. Holzer ee Braunkohlenformation. Acta Bot. Neerland. 13: 250- 1973. Holzer der rederreniscen Braunkohlenformation, 2. Rev. Palacobot, Palynol. 15: 73-27 Cuourey, M. S. 1974. A study of the Myricaceae from Eocene sediments of southeastern North America. Palaeontographica 146B: 88-153. CHOWbHUuRY, Kk. A. 19 OF, ates rings in tropical trees and taxonomy. Jour. Indian Bot. Soc. 43: —342. Cox, M. J. 1941. The ak anatomy of the ie xylem of five American species of Celtis. Am. Midl. Nat. 25: 34 Dorr, E. 1960. Tertiary fossil forests of Yellowstone ae Park, Wyoming. Pp. 253-260 in Billings Geol. Soc. Guidebook, 11th Ann. Field Conference. 1964. The petrified forests of Yellowstone Park. Sci. Am. 210(4): 105- EYDE, R. H. 1963. Morphological and paleobotanical studies of the Nyssaceae, I. A survey of the modern species and their fruits. Jour. Arnold Arb. 44: 1-59. Ne E. S. BaRGHOORN. 1963. Morphological and paleobotanical studies of e Nyssaceae, II. The fossil record. /bi es t 1896. Untersuchungen tiber fossile Hélzer. V. Zeit. Deutsch. Geol. Ges. 48: 249-260 GRAMBAST-FEssarp, N. 1966. Contribution a l'étude des flores tertiares des ré- gions provencales et alpines: deux boix nouveaux ss ee ae du Pon- tien de Castellane. Mém. Soc. Géol. France 105: 46. Grecuss, P. 1969, oe oe woods in eet 151 pp. Akadémiai kiado, Budapest, Hung HeEimscu, C. 1940. W a aenee and morphology of Rhus and allied ge Jour. Arnold Arb. 21: 279- . 1942. Comparative anatomy of me secondary xylem in the “Gruinales” and oo of Wettstein with reference to taxonomic groupings. Lilloa 8: 83- HorMann, E. i) “Pisacenieete aus dem sgn ares von Pram- bachkirchen in Oberdonau. Palaeontographica 88B: 1-86. JANE, F. W. 1970. The structure of wood. ed. 2. 478 pp. A. & C. Black Ltd., London. Know ton, F. H. 1899. Fossil flora of the Yellowstone National Park. U. S. Geol. Surv, Monogr. 32: 651-791 1978 | WHEELER ET AL., FOSSIL WOODS, II 25 Kramer, K. 1974. Die tertiaren Holzer Sidost-Asiens (unter Ausschluss der Dipterocarpaceae). 2. Teil. Palaeontographica 145B: 1-150. KravuseL, R. 1939. Ergebnisse der Forschungsreisen Prof. E. Stromers in den Wiisten Agyptens. IV. Die fossilen Floren Agyptens, Teil 3, E-L. Bay. Akad. Wiss., Math.-Naturwiss. Abt., n.s. 47: 1-140 Krigs, D. A. 1927. Comparative anatomy of the woods of the Juglandaceae. Trop. Woods 12: 16-21 Kruse, H. O. 1954. Some Eocene dicotyledonous woods from Eden Valley, Wyoming. Ohio Jour. Sci. 54: 243-268. Leopotp, E. B., & H. D. MacGInitig. 1972. eee and affinities of Tertiary floras in the Rocky Mountains. Pp. 147-200 im A. GRAHAM, Ed., Floristics and Paleofloristics of Asia and eastern North cae Elsevier Publ. Co., Amsterdam, Holland. eee H. D. 1953. Fossil plants of the Florissant Beds, Colorado. Car- gie Inst. Publ. 599: 1-188 1969. The Eocene Green River flora of pew a orado and northeastern Utah. Univ. California Publ. Geol. Sci. 83: 1974. An early middle Eocene flora from the ie Absaroka Volcanic Province. northwestern Wind River Basin, Wyoming. Jbid. 108: 4 a E. 1959. Ein fossiles Nyssa-Holz aus Japan, Nyssoxylon japonicum, Ps ty Sp. oenckend,.. Leth. 400 2hi-227. Mascon SR. 1976. Wood of Tapirira (Anacardiaceae) from the Paleo- ne Clarno Formation of Oregon. Rev. Palaeobot. Palynol. 23: 119-127. ee C. R., & L. CHark. 1950, Anatomy of the dicotyledons. Vols. 1 2. 1500 pp. Clarendon Press, Oxford, England. Mixter, R. B. 1976a. Reticulate thickenings in some species of Juglans. Am. four Bot. 63: 898-901. _ 1976b. Wood anatomy and identification of species of Juglans. Bot. Gaz. 137: 368-377 Miuier-StTott, W. R., & E. Maver. 1957. Uber tertiare Eichenhdlzer aus dem nnonischen Becken. Senckenb. Leth. 38: 121-168. 1960. te 2 ee Holzer aus dem Tertiaér des pannonischen Beckens. ibid Ale 255= x LOG2: a Myrcacen Holz aus dem ungarischen Tertiaér, Myricoxylon hungaricum n. g., Ds LOMAS 373733: Pace, V. M. 1964. Lyonottannasion from the lower Pliocene of western Nevada. ee ies —266. PANSHIN, A. J., & C. DE ate w. 1970. Textbook of wood rechnoleey Vol. 1. ed. 3. 705 pp. McGraw-Hill Book Company, New York Petrescu, I. 1970. ee cae POTEET n. Sp. in Olicscenul din NV Ro- maniei. Contr. Bot. Cluj Dp: “1 1971. Zelkovoxylon ae n. in Oligocenul din nord vestul Tran- silvaniei. 7bid. 1971, pp. 121-130. PLaTEN, P. 1908, Untersuchungen fossiler Holzer aus dem Westen der Vereinig- ten Staaten von Nordamerika. Naturf. Gesell. Leipzig Sitzungsber. 34: 1- 55, 161-164 PRAKASH, U. 1968. Miocene fossil wood from the Columbia Basalts of central Washington, III. Palaeontographica 122B: 183-200. & ARGHOORN. 1961la. Miocene fossil woods from the Colurnbia Basalts of ‘central Washington. Jour. Arnold Arb. 42: 165-203. 24 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 . 1961b. Miocene fossil woods from the Columbia Basalts of central Washington, II. /bid. 42: 347-358. Prive, C. 1974. Pterocaryoxylon pee ela gt n. sp., bois fossile de Juglanda- ceae provenant du Cantal. Rev. Gén « Sir 243~25 . 1975. Etude de quelques bois de ae tertiares du Massif Central, France, oe 153B: 119-140. ReEcorp, S. J., . W. Hess. 1943. Timbers of the New World. 640 pp. Yale Univ. ae Hav scott, R. A. 1954. Fossil fruits and seeds from the Eocene Clarno formation of Oreg on. eos 96B: 66-97. SHILKINA, J. 1958. The fossil woods of the Goderdzy Pass. Izdat. Akad. Nauk — Bot. Inst. V. L. Komarov. Trud. ser. 8, Paleobot. 3: 125-178. [In Russian. | SMEDES, H. W., & H. J. Prostka. 1972. Stratigraphic framework of the Absaro- ka Volcanic Supergroup in the Yellowstone National Park region. U. S. Geol. Surv. Prof. Pap. 729-C: 1-33. SPACKMAN, W. 1949. The flora of the Brandon Lignite: geological aspects and a comparison of the flora = its modern equivalents, 175 pp. Unpublished Ph.D. Thesis, Harvard Uni SUZUKI, M. 1975. Two new species of _ fossil woods from the Paleo- gene of Japan. Jour. Jap. Bot. 50: 238. eee M. 1971. Comparative prs of the Ulmaceae. Jour. Arnold Arb. ay 5 Tuomas, J. L. 1960. A monographic study of the Cyrillaceae. Contr. Gray Herb. 186: 1-114 . 1961. The genera of the Cyrillaceae and ee of the southeastern United States. Jour. Arnold Arb, 42: 96-10 TIFFNEY, B. H. 1977. Contributions to a kaa of the fruit and seed flora of the Brandon Lignite. 303 pp. Unpublished Ph.D. Thesis, Harvard Univ. UNGER, F, 1850. Genera et species plantarum fossilium. 627 pp. Braumiiller, Vienna. Wanc, C. W. 1961. The forests of China. Maria Moors Cabot Foundation Publication 5. 313 pp. Cambridge, Mass Watari, S. 1941. Studies on the fossil woods from the Tertiary of Japan. I. Fossil woods from the River aes Anatal village, Ninohe district, Iwate prefecture. Jap. Jour. Bot. 11: 416. 1948. Studies on ae fossil ai from the Tertiary of Japan. V. Fos- sil woods from the lower Miocene of Hanenisi Simane prefecture. /bid. 13: 5 1952, ee ng woods from the Miocene oe aa a Sea side of Honsyu. Jour. Fac. Sci. Univ. Tokyo Bot. 6(1-3): WesBER, I. E. 1933. W He from the Ricardo oo of Last a Gulch, California. Carnegie Inst. Publ. 412: 113-13 1941. Systematic anatomy of the woods the Burseraceae. Lilloa 6: 441-465. WHEELER, E. F., & L. C. Matren. 1977. Fossil wood from an upper Miocene locality in Seema Colorado. Bot. Gaz. 138: 112-11 . A. SCOTT, & BARGHOORN. 1977. Fossil dicotyledonous woods from Yellowstone National Park. Jour. Arnold Arb. 58: 280-306. WILLIAMS, S. 1942. Secondary vascular tissues of the oaks indigenous to the United States, III. A comparative anatomical study of the wood of Leuco- balanus and Erythrobalanus. Bull. Torrey Bot. Club 69: 115-129 1978] WHEELER ET AL., FOSSIL WOODS, II 20 asanre a C. 1973. A dictionary of the flowering plants and ferns. ed. 8 (te- y H. K. Atry SHAW). 1214 pp. University Press, Cambridge, En- ae WOLFE, i 1973. Fossil forms of Amentiferae. Brittonia 25: 334-355. 4 ‘DIVISION OF UNIVERSITY STUDIES DEPARTMENT OF BIOLOGY AND DEPARTMENT OF WOOD AND BOTANICAL MUSEUM AND PAPER SCIENCE HARVARD UNIVERSITY NortH CAROLINA STATE UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 RALEIGH, NORTH CAROLINA 27607 R.A a: U. S. GEOLOGICAL SURVEY DENVER, COLORADO 80225 26 JOURNAL OF THE ARNOLD ARBORETUM LVoL. 59 EXPLANATION OF PLATES PLATE I GURES 1-3. Rhus crystallifera, sp. no 1, transverse section, K 4 intervascular pitting, « 200; 3, ae section showing septate bers nd rays with ne fa oes erer cells, * 50. Ficures 4 Cyrilloxylon eocenicum, Sp. nov.: 4, transverse section, x 45; 5, salasform perforation plate, x 14 ; opposite a eee intervascular pitting, X 140; 7, tangential section, x eu FIGURE 8. Quercinium amethystianum, ae nov.: transverse section, note abundant parenchyma, G; PLATE II FIGURE 9. Quercinium amethystianum, sp. nov.: transverse section, & 50. FIGURES 10-12. Quercinium lamarense Knowlton, eel Wheeler, Scott, & Barg- hoorn: 10, transverse section, growth ring boundary, D-2054B-42. « OF di transverse aaee type specimen, <- 16% 12. transverse eee D-2054B-42, FIGURES 13-15 _Pterocaryoxylon knowltonii, sp. nov. , transverse sec- tion donne eae pore arrangement, * 25; 14, oe a pitting, * 200; 15, tangential section, note three- ety rays, asl PLATE III Ficures 16-19. Mvyrica absarokensis, sp. nov.: 16, transverse section, & 80 17, tangential section, note as parenchyma strands, 120; 18, on iform eras plate, X 300; 19, vessel to ray parenchyma pitting, 300. Sees 20-2 Nyssoxylon A eae hire sp. nov.: 20, transverse section, 5 epee transitional intervascular pitting, X 220; 22, scalariform per- fration ane x 375; 23, tangential section, note composite rays, X 100. Frc- E 24. Prunus gummosa, comb, novy.: tangential section, & 80. PLATE IV Figures 25-27. Prunus gummosa, comb. : 25, transverse section show- ing traumatic gum canals, X 40: 26, spiral eras in small vessel element, 100; 27, simple perforation plate and - to ray parenchyma pitting, x 20. FIGURES 28-32. Turpinia lamarense, s sp. nov.: 28, transverse section, < 60; 29, tangential section, & 50; 30, opposite Praneeen re pitting, & 300; 31, scalariform perforation plate, X 220; 32, vessel to ray parenchyma pitting, 220. PLATE V Ficures 33-37. Zelkovoxylon eee sp. nov.: 33, transverse section, note semi-ring porous condition, & 40; 34, Eee perforation plate, X 150; 35, spiral thickening in small tracheary ee ye 36, tangential section, 4 1 Os 37, tangential section showing crystalliferous ee ray cell, X 180. PiatTE I Jour. ARNOLD ARB. VOL. 59 WHEELER ET AL., Fosstt Woops, II PLATE II Jour. ARNOLD ARB. VoL. 59 —— Pe goede a S~ ie as Ee a See es 22 i 2 > te = ae iT ) OODS WHEELER ET AL., Fossit W PLaTE III Jour. ARNOLD ArB. VOL. 59 cont = fgg: ea nate Stns oes Pee aes <4 = - “ete en y—~ LY im Re Se 88s Se ae ait mes Sadan Sn ae. or eee bs oS; ee won 9 lg tae ee SS ¥ 3° z a a ae by as tae en ssh as page, soem 5 . ee ae II ? WHEELER ET AL., FossiL Woops PLaTE IV Jour. ARNOLD Ars. VOL. 59 S gt 6 08 ee _. ets wl é fs ‘ ie. ak 2 AT’ cam — x H.S. MacKee 3818 Lam é& Meeuse 6068 Carlquist 1017 Carlquist 6085 Stebbins & Keighery 4-25 Kraehenbuehl 2153 Carlquist 6056 DJ. Whibley 1081 E. Pritzel 572 Stebbins & Keighery 1-9 A, Morrison sn. Carlquist 5940 A.C. Smith 3546 A.C. Smith 3045 P) COLLECTION HERBARIUM LOCALITY VOUCHER * XYLARIUM New Caledonia 1G; oa S. Australia L — Papua L — New Caledonia ? BRIw 13033 New Caledonia L — Madagascar L — W. Australia ? BRIw 11684 W. Australia RSA RSAwW W. Australia RSA RSAwW W. Australia UCB a S. Australia NCU — W. Australia RSA RSAW S. Australia NCU — Australia L — Australia UCB —- W. Australia L ae Australia RSA RSAW Fiji MAD, NY SJRw 28402 Fiji SJRw 27903 New Caledonia BRIw 12037 [SL61 VILYAIdIH “TV LA NOSIMDIAG TABLE 1. Wood specimens of Hibbertia examined (continued). SPECIES montana Steudel ngoyensis Schltr. potentillifiora Bentham procumbens (Labill.) DC. scandens (Willd.) Gilg scandens (Willd.) Gilg stricta (DC.) F. Mueller trach yphylla Schltr. uncinata (Bentham) F. Mueller virgata R. Br. ex DC. COLLECTOR Stebbins & Keighery H. S. MacKee 32303 G. L. Stebbins A-53 Hoogland 3067 Stebbins & Keighery N.T. Burbidge 3326 E. F. Constable 4434 R. D. Hoogland 3115 E. F. Constable 3850 M.C.R. Sharrad 342 C.L. Wilson 855 DJ. ia 855 Stebbins & Keighery Bd. “Blaylock 575 COLLECTION LOCALITY Australia New Caledonia N.S.W., Australia N.S.W., Australia (Queensland, Australia Cult. BRI s.n. N.S.W., Australia S. Australia Australia New Caledonia Australia S. Australia * Abbreviations follow Holmgren and Keuken (1974) in Index Herbariorum. HERBARIUM VOUCHER “* NYLARIUM BRIw 13034 $e WOLAXYOdaVY GIONUV AHL AO TVNUNO! 6S “I0A| 1978 | DICKISON ET AL., HIBBERTIA 35 OBSERVATIONS GrowTH RINGS. Well-defined growth rings are recognizable by the pres- ence of thicker-walled imperforate tracheary elements in the late wood of the Australian species Hibbertia drummondii, H. huegelii, H. montana, H. exutiacies, H. virgata, H. potentilliflora, H. saligna, H. stricta, H. un- cinata, H. hypericoides, and H. obtusifolia, Very weak growth rings are also evident in 7. cuneiformis and H. sericea. Growth rings are irregularly spaced and conspicuously wavy in outline as viewed in transection. Woods of H. exutiacies and H. virgata are, in addition, distinctly ring porous, as shown by the restriction of larger vessel elements to the early wood of a growth ring. VESSEL ELEMENTS. Selected characteristics of the wood of species stud- ied are summarized in TaBLe 2. Four species, Hibbertia lucens, H. salig- na, IH. cunciformis, and H. coriacea, formed the basis of the description of Hibbertia wood by Dickison (1967). In spite of the fact that they be- long to different sections of the genus (which in floral morphology differ from each other so much that they might best be regarded as subgenera), these species are very similar to each other with respect to xylem struc- ture. Wood of the remaining species is here recorded for the first time and reveals an interesting diversity of anatomical structure previously un- reported for the genus Hibbertia. n most plants the shape of the pores varies from slightly angular to circular. In Hibbertia saligna, H. lucens, H. cuneiformis, H. trachyphylla, and H. obtusifolia the pores are decidedly angular, whereas in H. scandens, H. huegelii, and H. exutiacies there is a strong tendency toward a circular outline. As reported by Dickison (loc. cit.), vessel elements are solitary in distribution; the paired condition is present only in the region of vessel element overlap. Walls are typically of medium thickness, but decidedly thin-walled vessels occur in the Australian H. scandens and in H. ngoyensis of New Caledonia. Vessel walls of the other New Caledonian plants ex- amined also show a tendency toward the thin-walled condition. Mean pore diameters range from extremely small in the low, shrubby species to mostly moderately small in the larger tree forms. In the single collection of H. hypericoides studied, pores have extremely small diameters, making it difficult to distinguish vessels from imperforate tracheary elements in transection. Vessel diameter in H. scandens has clearly increased in re- sponse to the liana habit. The number of pores per unit area varies con- siderably, being highest in H. stricta and H. exutiacies, and lowest in H. scandens, Within the genus, mean vessel element length ranges between medium sized and moderately long. Intervessel pitting is sparse, but where present is opposite to transitional to scalariform in distribution. Pits are circular to scalariform in outline and range in size from minute to very large. Ray—vessel pitting is half- bordered with a distribution pattern similar to that of intervessel pitting. Hibbertia procumbens, a prostrate semi-xeric species from southeastern TABLE 2. Summary of selected wood anatomical features of Hibbertia in relation to growth habit and habitat. from immature samples are in parentheses.) The numbers corresponding to — states are: 1, number of ae studied ; Z, eae aa distribution; 3, ecole gical preferences; 4, growth es , leaf size; See of stems studied; 7, growth rings; range in v element lengt th, um. mean vessel element length, un 10, mean es diam a Be ea plates scalari- form only; 12, aia plates lees and simple; 13, bars per cre aee perforation, pens ‘14, mean fiber length, um.; 15 fiber wall thickness Section and species 1 2 3a 4 5b 6c 74 8 9 10 ll 12 13 14 15® Polystiche baudouinii 2 New Caledonia M rosette tree, L. S. 4 M - 660-1408 1078 57 ay 17 1441 m, t 10 m. Spicatae ngoyensis 1 New Caledonia M rosette tree, Let 35-3 M = 759-1243 968 95 + 22 1804 t 6 m. lucens 3. New Caledonia, SX pee or Ly S83 M - 625-2037 1074 61 F 42 1551 My: «iG Fiji etbhe tree trachyphylla l New Caledonia SX shrub or Lie 4iS%,.3 6.0 mm. - (418-1034) (726) (40) + (24) (1056) m, t rosette tree altigena ] New Caledonia SX shrub, 1.0 m. Tyee Se. 2. 4.0 mm. as (341-1012) (682) (35) a (32) (1001) m Cyclandra saligna 2 E. Australia, M erect shrub, lie. Siz 3 M + 770-1694 1199 38 + 46 1351 m, ¢ N.S.W. 60-90 cm. serrata 1 S.W. Australia M shrub, 2-3 m. be Se3 of 7 ig 979 2 + 32 £139 ? obtusifolia 2 S.E. Australia SX small, erect L. &. 3 4.0 mm. + 308-957 627 25 + 17 825 m shrub, 0.1] m. scandens 2 E. Australia M vine L. S. 3 8 mm. - 413-1430 731 45 + 9 943 t 9¢ WOLHYYOPTUVY GIONYV AHL AO TWNYNOL 6§ “10s | montana drummondii. glaberrima potentilliflora virgata procumbens Candollea cuneiformis huegelii uncinata Hemipleurandra lineata furfuracea hypericoides Pleurandra sericea stricta rR = i W. Australia Interior of W. Australia Interior of S. & N. Australia W. Australia E. Australia Tasmania, E. Australia S.W. Australia W. Australia Interior of W. Australia W. Australia S.W. Australia S.W. Australia E. Australia E. & S. Australia SX-X SX-M SX-M SX SX shrub, LO peu fete Nae shrub, 0.5 m. shrub semi-prostrate shrub shrub, 0.5 m. prostrate shrub shrub, 1-2 nm. shrub, 0.5 m. shrub, 0.2-0.5 m. semi-prostrate shrub shrub, 0.5-1.5 m. shrub shrub, 0.5-1 m. shrub, 0.5 m. 3 2.5 mm. 440-990 220-880 220-825 270-622 248-550 165-671 572-1232 330-1265 330-853 473-869 528-979 363-1056 363-1265 193=7:70 tN oa 1402 m, [SL6I VILMAAdIH “TY LA NOSIMDIG exutiacies Pea 901 used: TABLE 2. Summary of selected wood anatomical features of Hibbertia in relation to growth habit and habitat (continued). ] S.E. Australia SX data unavailable Ge ):3 prostrate shrub Le. tou 3.8 mn. + 231-660 430 2] 2 » 613 aspera 1 S.E. & W. SX shrub Liv: 2S, 2 3 mm. Jo2=7 92 561 z + 7 770 Australia Hemistemma coriacea i Madagascar SX shrub, 0.6-2 m. Ls Oe “2S M - 693-1287 957 40 + 19 1474 Le Se. 3 banksii 1 N. Queensland, M shrub, 1-2 m i oe 3: - 671-1430 968 ? + 12 1144 Australia; S. New Guinea “symbols used: mesic (M), semi-xeric (SX), or xeric (X). b 2 P 2 : 2 Leaf size classes after Raunkiaer (1934): (L. S. 1) Leptophylls 0-25 mm.*; (L. 2) Nanophylls 25-225 mm.~; (L. S. 3) Microphylls 225-2025 mm.” ; (L. S. 4) Mesophylls 2,025-18,222 m2. “symbol used: mature wood (M). d. Symbols used: growth rings present (+), or absent (-). “symbols used: walls thin - lumen greater than thickness of walls (t); medium - lumen equal to thickness of walls (m); or thick - lumen less than thickness of walls (th). se WOLAYOdUY GIONYV AHL AO TWNYAOL 6S “T0A] 1978] DICKISON ET AL., HIBBERTIA 39 Australia, is unusual in possessing extensive scalariform lateral wall pit- ting. Perforation plates are exclusively scalariform in the vast majority of species, although there is wide variation in the number of bars per per- foration. In our material mean bar number ranges between a high of 42 and a low of 2 (Frcures 1-10). Bars also vary considerably in width and degree of separation. Branched bars are frequent. It is of much interest cv NUD AUC ANAT CAMA eer) ——— ieee ] CUES UCCUUCN ALM NNtUnHCHE Ls ees, fects Soll MN CLL oe vf —. ™ t aii a Girma & ete AN Hn My Ty] \ PRU Ua iF HAE Mpsssvn over ng posit © 100, 1O00u eet _—— ——— eA TEUELL UE AMANDA Stn cet ——— ane mn a 1OOp Gi Nyy oe eee \ ro) °o = Frcures 1-11. Vessel member structure in Hibbertia and Pachynema: 1, Hib- bertia saligna (BRIw 13034); 2, H. baudouinit (BRIw 13033); 3, H. sericea (Sharrad 342); 4, 5, H. uncinata (Stebbins & Keighery A-3); 6, 7, H. exutiacies (Kraehenbuehl 2153); 8, H. obtusifolia (Stebbins A-53); 9, H (BRI . cuneiformis w 11684); 10, H. banksii (Brass 8431); 11, Pachynema junceum (Dun- 40 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 that simple perforation plates were observed in a low percentage of the vessel elements in H. drummondii, H. glaberrima, and H. uncinata (F1c- URE 4). In these species vessel elements are occasionally present with simple perforations on one end and few-barred scalariform perforation plates on the other. IMPERFORATE TRACHEARY ELEMENTS. Imperforate tracheary elements are fiber-tracheids, although in many of the small, semi-xeric species the pit borders are very reduced, and the elements approach the libriform fiber type. The mean length of the fiber tracheids generally ranges within the very short and short categories; only Hibbertia lucens and H. ngoyensis have imperforate tracheary elements that can be classified as long. Bor- dered pits are usually present on both radial and tangential walls and have minute to large diameters. In the New Caledonian species the pitting is often distributed in multiseriate rows. The walls of fiber tracheids vary considerably in thickness. Thin-walled elements (cell lumen greater than thickness of walls) occur in H. scandens, H. baudouinii, H. lucens, H. trachy phylla, H. saligna, H. virgata, H. cuneiformis, H. hypericoides, H. coriacea, and particularly H. ngoyensis. Thick-walled elements (cell lu- men almost completely closed) characterize H. glaberrima, H. huegelii, H. uncinata, H. sericea, and particularly H. drummondii. Rays. Rays are heterogeneous, and in plants with abundant xylem both multiseriate and uniseriate rays are present. In Australian plants with limited secondary xylem production, generally only uniseriate rays occur, although these may occasionally become biseriate. Multiseriate rays are present in New Caledonian species in the first-formed secondary xylem. Multiseriate rays are 1 to 5, mostly 1 to 3, cells wide, are composed pri- marily of upright cells, and have long uniseriate wing extensions of the upright cells. Rays tend to be numerous to abundant per square mm. AXIAL PARENCHYMA. Axial parenchyma is diffuse and scanty in dis- tribution, although it becomes increasingly scarce in more xeric species. Axial parenchyma is rare in plants like Hibbertia drummondii. Numerous raphides are present in enlarged axial parenchyma cells of H. cuneiformis and #7. virgata. DISCUSSION In the genus Hibbertia, trends of specialization with respect to growth habit have been chiefly toward small, thin-stemmed, prostrate or semi- prostrate shrubs; and with respect to leaves primarily toward leptophylls and nanophylls (sensu Raunkiaer, 1934) having various kinds of special- izations associated with drought resistance. Nodal anatomy has frequent- ly been reduced from the trilacunar to the unilacunar pattern. The result- ing diversity in plant habit, and, therefore, the amount of secondary xylem produced, makes an evaluation of the wood anatomical data obtained from this genus difficult. The dangers in comparing data from woods of different 1978 | DICKISON ET AL., HIBBERTIA 4] ages and growth habits are well known. Nevertheless, it is clear that H76- bertia has undergone an extensive adaptive radiation that has often resulted in reduced plant size in response to increasingly arid conditions. The com- bination of decreased plant stature and decreased water availability has produced significant alterations in xylem structure. Hibbertia species from mesic habitats show an absence of growth rings. Seasonality of climate is reflected in xylem with growth rings and even ring porosity. In Hibbertia, vessel elements have frequently become shorter, a trend noted for other xeric taxa by Carlquist (1975), and their perfora- tion plates have acquired a reduced number of bars (less than 10) with a few species even possessing occasional simple perforation plates. Small differences in vessel member length are not here regarded as important, although variation representing different size classes of elements is signifi- cant. In dicotyledons as a whole, these trends have usually been accom- panied by an increase rather than a decrease in pore diameter. In Hib- bertia a reduction in vessel element length is associated with a reduction in pore width. These tendencies are also combined with an increase in pore number per unit area. It must be kept in mind, however, that there has also occurred a dramatic decrease in the total amount of xylem pro- duced in the xeric plants. The only exception in the trend toward smaller pore diameters in small-stemmed species is /. scandens, a liana, in which vessels are very wide in comparison with stem diameter, yet have retained scalariform perforation plates with a moderate bar number. Vith regard to correlations between perforation plate type and eco- logical factors, van der Graaff and Baas (1974) and Carlquist (1975) pre- sented evidence to show that no general trend can be established in di- cotyledons as a whole, an opinion further enunciated and documented by Baas (1976). Individual genera, however, occasionally do show such cor- relations. Hibbertia, for example, demonstrates a high degree of correla- tion between mesic habitats and longer vessel elements with many-barred scalariform perforation plates, and xeric habitats and vessel elements with reduced length, diameter, and bar number in scalariform perforations. The only notable exception is the xeric species 7. procumbens, which has short, narrow vessel elements, yet has retained scalariform perforation plates with a high number of bars. The scalariform lateral wall pitting on the vessel elements of this species is also unusual. It is very clear that in Hibbertia, habit-related anatomical modification is superimposed upon evolutionary, habit-related, and/or physiological specialization, and the separation of these facets is difficult. Four other families of angiospermous shrubs occur regularly in associa- tion with the specialized Australian species of Hibbertia: Proteaceae, Le- guminosae (especially Acacia), Myrtaceae, and Epacridaceae. In the first three families, vessel elements are short, comparatively wide, and have simple perforations. The Epacridaceae, however, resemble Hibbertia in that several species have vessel elements with small or very small diameters and scalariform perforation plates. The Proteaceae, Leguminosae, and Myrtaceae contain large shrubs and trees, and most of the recorded data 42 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 are from these larger species. The Epacridaceae, on the other hand, con- tain chiefly small, slender-stemmed shrubs similar in growth habit to Hibbertia, Baas (1976) has observed that in the Epacridaceae, the genera with scalariform plates show a mesic preference, and the genera with ex- clusively or predominantly simple plates prefer xeric habitats. RELATION BETWEEN WOOD ANATOMY AND TAXONOMY. As reported by Stebbins and Hoogland (1976), the genus Hibbertia is remarkable for its great range of variation with respect to leaf size, shape, and structure, as well as to floral morphology. The present study reveals an additional large range of variation with respect to secondary xylem structure, particu- larly in vessel element morphology. In many of the principal sectional groups (sensu Gilg & Werdermann, 1925), which differ from each other with respect to floral morphology, the trend toward medium-sized to small, often acicular leaves can be found as an independent course of evolution (Rury & Dickison, 19 Within each section or sectional group (considering HEmMiPLEURANDRA, HeEMISTEMMA, and PLEURANDRA as a single group) there a appear to be parallel trends of specialization. In leaf morphology they consist of re- duction in leaf size and venation, and increase in the proportion of scler- enchymatous tissue; in floral morphology there has been reduction in size and number of parts; and in vessel elements, reduction in length, diameter, and number of scalariform perforation plate bars, all correlated with an overall reduction in plant size. A different combination is seen in Hibbertia scandens (section CYCLANDRA, subsection SUBSESSILES ), in which larger leaves and flowers are associated with vessel elements showing no reduction in diameter. This species has also become specialized differently in growth habit, since it is a vine rather than a small shrub. The three New Caledonian sections exhibit considerable homogeneity in their wood anatomy. In addition, they have many similarities to section HEMISTEM- MA, a relatively mesophytic species group from northern Australia, New Guinea, and Madagascar. A very interesting parallelism in leaf and nodal structure also occurs within these sections separated by floral characters. The New Caledonian subsection TrRimorPHANDRA, placed by Gilg and Werdermann (1925) within the highly heterogeneous Australian section CYCLANDRA on the basis of floral morphology, exhibits a leaf morphology identical to that of the New Caledonian hibbertias (Rury & Dickison, 1977). Although wood specimens of subsection TRrMORPHANDRA were not available for study, we strongly suspect that the wood anatomy of these species would serve to support the suggested transferral of this sub- section to one of the New Caledonian sections of Hibbertia. A lack of correlation is evident when Hidbbertia is compared with two of its nearest relatives, Didesmandra and Schumacheria. These genera have longer vessel elements with larger numbers of scalariform perfora- tion plate bars than any species of Hibbertia (Dickison, 1967). Neverthe- less, their flowers are bilaterally symmetrical, and in the reduction of the ovules to one per carpel, as well as in floral anatomy, they are more special- 1978 | DICKISON ET AL., HIBBERTIA 43 ized than any species of Hibbertia (Dickison, 1968; Wilson, 1973). Pachynema, the third genus related to Hibbertia, is essentially leafless and has a much reduced floral structure (Wilson, 1973). Examination of the secondary xylem of P. junceum (van Steenis 17664 and C. Dunlop s.n.) reveals that vessel elements are medium sized in length (242-627 pm., mean 440 pm.) with scalariform perforation plates. The number of bars per scalariform perforation plate ranges from 2 to 15 with an average num- ber of 7 (Ficure 11). Fiber-tracheids are very short in length (363-902 um., mean 660 pm.). RELATION BETWEEN VESSEL ELEMENT STRUCTURE AND GEOGRAPHIC DIS- TRIBUTION. The species having the least specialized vessel elements include endemics to each of the major geographic regions where Hibbertia occurs (eastern Australia, southwestern Australia, New Caledonia, Madagascar ) with the exception of northern Australia, from which no specimens have yet been examined. However, since the species of northern Australia re- semble H. coriacea of Madagascar and H. lucens of New Caledonia in most characteristics, and occur in similar habitats, there is good reason to believe that most of them have relatively unspecialized vessel members. With the exception of H. scandens, which occurs from southeastern Aus- tralia northward to New Guinea, all of the species known to have more specialized vessel elements occur in southeastern, southwestern, or interior Australia, and inhabit temperate or subtropical rather than tropical re- gions. Trends toward leptophylly with accompanying reduction in length and diameter of vessel elements occur in species endemic to southeastern Australia (H. virgata), to the dry interior (H. glaberrima), and to south- western Australia (H. Auegelii), as well as in species groups that are dis- tributed all across the continent (H. stricta). The three species observed to possess occasional simple perforation plates are all from the interior region of Australia. The evidence from geography, therefore, supports that from floral morphology to indicate that in various parts of temperate Australia, independent trends toward xylem specialization have occurred contemporaneously. Although in Hibbertia reduction in vessel member length and diameter is apparently the result of a multitude of biotic and abiotic factors, these trends do follow the general trends for other dicoty- ledons of decreased vessel element length and diameter with increasing latitude (van der Graaff & Baas, 1974). RELATION BETWEEN WOOD STRUCTURE AND HABITAT. The trends of specialization of vessel elements toward shortening, reduction in diameter, reduction in the number of bars on scalariform perforation plates, and even the formation of simple plates, that are reported for the genus Hzbbertia are clearly associated with occupation of drier habitats, including those like the heathlands of eastern Australia. Similar trends are reviewed by Carlquist (1975) for desert shrubs and high montane or alpine cushion plants belonging to various families. Associated with these changes in Hibbertia vessel structure, is a concomitant decrease in length and in- 44 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 crease in wall thickness of imperforate tracheary elements and an overall decrease in the amount of axial parenchyma. The species with the thickest fiber-tracheid walls is H. drummondii, a species also characterized by the occasional occurrence of simple perforation plates, which grows in the dry Australian interior. Nevertheless, a different picture emerges when one compares the least specialized species belonging to each of the sections with each other, and also when Hibdbertia is compared with other genera of Dilleniaceae. These species, although by no means xerophytic, do not always occupy the wetter habitats of the regions where they are found. Hibbertia saligna, a species with comparatively primitive vessel members, is generally distributed on ridge crests and mountainsides in a forest of Eucalyptus spp. This species of Hibbertia, because of the tendency of Eucalyptus leaves to hang ver- tically, receives abundant light on the forest floor. Hibbertia serrata, H. cu- neiformis, and H. lucens are three additional species which possess vessel elements having scalariform perforation plates with 24 to 42 bars. The first two of these species grow in southwestern Australia where annual precipi- tation is ample, from 40 to 60 inches, but where the summer drought lasts from four to five months. In this region, they occur in relatively dry as well as more mesic sites. Hibbertia lucens has a tropical habitat, but ac- cording to herbarium labels it grows in dry, open forests rather than in wet rain forests. All of the New Caledonian species that were investigated have comparatively primitive vessel members. Hibbertia baudouinii occurs : more mesic situations than does H/. /ucens, and the specimen of H. ngo sis studied grew in a humid forest. The remaining species having — elements with 20 or more bars per scalariform perforation plate are 7. coriacea of Madagascar, which grows in the climatically moist eastern side of the island (but according to herbarium labels in dry, sunny places or on sandy soil that is subject to periodic drought), and H. obtusifolia, the specimen of which was collected in a dry Eucalyptus forest near Can- berra, Australia. Fiber-tracheid walls of these plants tend to be thinner than in other species. Also, all of the above species of Madagascar and New Caledonia have a more regular, better-developed system of both low and high order venation. Present evidence suggests, therefore, that the correlation between vessel element specialization and adaptation to drought holds well for trends within the Australian species of Hibbertia, but that the situation within the genus as a whole, particularly among the species of northern Australia, New Caledonia, and Madagascar, is less clear. Any aaa about these species must await the study of additional materi e same ambiguity arises when other genera a Dilleniaceae are ex- amined. Long vessel members having exclusively scalariform perforation plates with many bars occur in Dillenia, Schumacheria, and Didesmandra. Whereas the latter two genera are rain forest plants, Dillenia grows in a variety of habitats including lowland rainforests, savannas, and monsoon forests with pronounced dry periods (Baas, 1976). Another dilleniaceous genus with scalariform perforation plates is Acrotrema, a woody herb that jean 1978] DICKISON ET AL., HIBBERTIA 45 grows in wet areas. These genera, however, are distinctly more specialized with respect to floral morphology than is Hibbertia (Dickison, 1968; Wil- son, 1973), so they cannot be regarded as being more (or even as) closely related to the common ancestor of the Dilleniaceae than is Hibbertia. The presence in Pachynema complanatum of elongate vessel elements of which the scalariform perforation plates have 16 to 48 bars (Dickison, 1967) is truly remarkable, and a re-examination of this species with new material would be highly desirable. This species grows in the tropical monsoon climate of northern Australia where it is subjected to five months of winter drought. It is a shrublet bearing leafless, flattened stems and much-reduced flowers. The habit of the plant, and even more that of other species of Pachynema, such as P. junceum, suggests a long history of adaptation to aridity. The occurrence of scalariform elements in Pachynema suggests the possibility of an early derivation of this genus from the ancestral hib- bertias and a subsequent radiation into xeric habitats, rather than a more recent derivation from the extreme xeromorphic forms of Hibbertia. Con- sequently, the anatomy and the ecological preferences of the least special- ized species of Dilleniaceae do not support the hypothesis that their an- cestor evolved in a tropical rain forest climate, and that specialization has been continuously and irreversibly associated with adaptation to greater aridity. Woop ANATOMY AND LEAF MORPHOLOGY IN RELATION TO HABITAT. Vege- tative character syndromes observed in the growth habit, nodes, and leaves of numerous species of Hibbertia represent the evolutionary products of morphological adaptation within diverse ecological situations (Rury & Dickison, 1977). The present study reveals ecologically adaptive charac- ter syndromes in the wood anatomy of Hibbertia which are clearly cor- related with both habit and habitat of the species studied (TABLE 3). It appears that the vegetative and wood anatomical character syndromes of these hibbertias have evolved, in general, as a single vegetative unit in response to environmental selection (TABLES 3, 4). Hence the wood anat- omy, nodal structure, and leaf morphology and venation are best regarded as a “morphological continuum” which has evolved as a unit during the ancient adaptive radiation of this genus (sensu Stebbins, 1974). The Australian hibbertias exhibit a heterogeneous species assemblage with respect to leaf size, leaf morphology, and floral structure. The most mesic species of Australia occur as small (Hibbertia saligna) to large (H. serrata) shrubs, or rarely vines (H. scandens, H. dentata). In these plants, features such as long, wide, scalariform vessel elements with numerous bars, and thin-walled fibers are correlated with mesomorphic leaves with primitive (sensu Hickey & Wolfe, 1975) leaf venation patterns. All mesophytic spe- cies examined, with the exception of H. Aypericoides, possess three-trace, trilacunar nodes, a feature considered as primitive for Hibbertia and the entire family Dilleniaceae (Dickison, 1969). The presence of such features as short vessels with few bars (mean 7) per perforation plate, unilacunar nodes, reduced leaves (nanophylls), and reduced leaf venation in H. hy- TABLE 3. Wood anatomical features and leaf sizes of 27 species of Hibbertia in relation to habitat type and geography FIBER WALL LEAF SIZE VESSEL ELEMENT PORE No. BARS PER FIBER SPECIES GROUPING CLASS ° LENGTH, um. DIAMETER, “mM. PLATE LENGTH, “em. THICKNESS ° Australian mesic : ) mean Le Se 851 30 21 1059 m-t range $2 308-1694 14-45 6-80 792-1402 New Caledonian mesic | Per 1023 76 20 1623 m-t ange L.S. 3,4 968-1078 57-95 17-22 1441-1804 Total mesic species (10 spp.) mean ast 885 42 21 1172 m-t range L. S. 2-4 308-1694 14-95 6-80 792-1804 Australian semi-xeric (11 spp.) mean L.S. 2 585 22 10 801 m range de Ss 165-1265 14-40 0-37 517-1402 t-th Island semi-xeric ) mean LS: 863 45 29 1188 m-t range L.S. 2,3 341-2037 35-61 13-50 1001-1551 Total semi-xeric species (15 spp.) mean Le 655 28 5 898 m range L. S. 1-3 165-2037 13-61 0-50 517-1551 t-th Xeric species (Australian) 2 spp.) mean L.S. 2 522 18 3.5 764 m-th L.S. 1-3 497-547 13-23 3-4 760-767 " Leaf size classes: see TAB " Fiji, New Caledonia, and Monacates ° Fiber wall iclinesses: t, thin; m, = ae th, thick. Ob WOLAYOPdUY GIONYV AHL AO TWNYAOL 6S “10A] TABLE 4. Wood anatomical features of 27 species of Hibbertia in relation to their leaf size classes. VESSEL ELEMENT PORE No. BARS FIBER FIBER WALL LEAF SIZE CLASSES" LENGTH, #m. DIAMETER, um. PER PLATE LENGTH, um. THICKNESS ° LBP Sy ys (12 spp.) mean 559 19 11 is m b (4 spp.) mean 726 28 12 1017 m L. S. 3, 4 (11 spp.) mean 897 45 26 1206 m-t * Leaf size classes: see TABLE 2 ” Excluding species with exclusively L. S. 2 or L. S. 3 leaves, i.e., leaves of intermediate sizes. “Fiber wall thicknesses: t, thin; m, medium; th, thick. [8261 VILYAGdIH “TV LA NOSIMOIG 48 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 pericoides, a species of semi-xeric and mesic habitats, suggests that this species may be a secondary radiant into mesic areas All of the ile island species of Hibbertia exhibit a vegetative morphological strategy which differs from that of the mesophytic Aus- tralian hibbertias. The large shrubs and rosette trees of these island habi- tats possess trilacunar nodes and large, entire leaves with strong intra- marginal veins and a very rigid, regular, and well-developed system of low (brochidodromous) and high order venation. These mesophytes, how- ever, possess long, scalariform vessel elements which, with the exception of a slightly larger pore diameter, are very similar to those of their shrubby Australian counterparts. The non-expansible, tear-resistant leaves of these island hibbertias thus appear to represent the result of evolution within a relatively non-seasonal, growth-promoting habitat where the potential for sudden leaf expansion in response to ephemerally favorable microcli- matic conditions would not be advantageous. The occurrence of such mesomorphically advanced leaf venation patterns in plants with relatively primitive wood anatomy represents a type of mosaic evolution which is also evident in other genera of Dilleniaceae. LITERATURE CITED Baas, P. 1976. Some functional and adaptive aspects of vessel member mor- phology. Leiden Bot. Ser. 3: 157-181. a I. W. 1966. The significance the reduction of vessels in the Cacta- eae. Jour. Arnold Arb, 47: 288-29 ee Ist, S. 1975. Ecological een of xylem evolution. xi + 259 pp. University of California Press, keley. COMMITTEE ON NOMENCLATURE, International Association of Wood Anatomists. 1964. Multilingual glossary of terms used in wood anatomy. Verlagsanstalt Buchdruckerei, Winterthur, Switzerland. COMMITTEE ON STANDARDIZATION OF TERMS OF CELL SiZE. International Associa- tion of Wood Anatomists. 1937, ees terms of length of vessel mem- bers and wood fibers. Trop. Woods ae Standard terms of size for ee diameter and ray width. /did. 59: 51, DICKISON, W. ‘< 1967. Comparative eae studies in Dilleniaceae. I. W vod anatomy. Jour. Arnold Arb. —29. 1968. Comparative Be cee ee in Dilleniaceae. III. The carpels Ibid. 49: 317-329. ———. 1969. Comparative morphological studies in Dilleniaceae. IV. Anatomy of . node and vascularization of the leaf. /bid. 50: 384-400. GILG, E., & E. WERDERMANN. 1925, aie In: A, eae & K. PRANTL, eds., Nat. Pflanzenfam. 21: 21- GRAAF, N. A. VAN DER, & P. Baas. on Wood anatomical variation in rela- tion to latitude and altitude. Blumea 22: 101-121. Hickey, L. J., & J. A. Wotre. 1975. The bases of angiosperm phylogeny: vegetative morphology. Ann. Missouri Bot. Gard. 62: 538-590, Hormcren, P. k., & W. KEUKEN. 1974. Index herbariorum. ed. 6. Reg. Veg. 92: = 97 1978] DICKISON ET AL., HIBBERTIA 49 RAUNKIAER, C. 1934. The life forms of plants and statistical plant geography. xvi1 + 632 pp. Clarendon Press, Oxfor Rury, P. M., & W. C. Dickison, 1977. Leaf venation ea of the genus Hibbertia (Dilleniaceae). Jour. Arnold Arb. 58: 209- Sreppins, G. L. 1974. Flowering plants: evolution ae the species level. xviii + 399 pp. Belknap-Harvard Press, Cambridge —— .D. HoocLtanp. 1976. Species diversity, ecology and evolution in a primitive angiosperm genus: Hibbertia (Dilleniaceae). Plant Syst. Evol. 125: 139-154. Witson, C. L. 1973. The floral anatomy of Dilleniaceae. II. Genera other than Hibbertia. Phytomorphology 23: 25-42. W.C.D. AnD P. M.R. G.L.S. DEPARTMENT OF BOTANY DEPARTMENT OF GENETICS THE UNIVERSITY OF NoRTH CAROLINA UNIVERSITY OF CALIFORNIA CHAPEL HILL, NorTH CAROLINA 27514 Davis, CALIFORNIA 95616 50 JOURNAL OF THE ARNOLD ARBORETUM VOL. 59 AN ETHNOFLORA OF CHOKOLOSKEE ISLAND, COLLIER COUNTY, FLORIDA DanteL F. AUSTIN AND Davin M. McJUNKIN CHOKOLOSKEE ISLAND, one of the largest of Florida’s Ten Thousand Island group in Collier County, has a long history of habitation by man. Because of this extended occupation, its flora has shifted in species com- position with each change of people. The native flora was mostly tropical hardwood hammock and mangroves; settlement by the Calusa Indians, the Spanish, the Seminoles, the English, and various groups from the United States has greatly altered the original plant associations. At pres- ent the island is experiencing what may prove to be the most traumatic shift. The newest change, increasing development, threatens to eliminate all traces of past eras. This study records what remains of the past, as well as present changes. HISTORY Primeval Chokoloskee differed from other islands in the area in being higher and protected in an isolated end of the back bay, now known as the Chokoloskee Bay. Here at the mouth of the Turner River, people from the north first settled about 2200 B.C. (Sears, 1956). The earliest re- corded name for the descendants of these people, the Calusa, comes from the Spanish. The proto-Calusa economy was fishing, hunting, and gathering, and was based on the locally abundant oyster (Cushing, 1896; Willey, 1949; Sears, 1956). Quahog clam, horse conch, and the lightning whelk were other important marine products. These were valued not only for their flesh, but also for their shells which were used for tools since suitable rocks were absent except as occasional trade items from the north. Terrestrial food in the form of deer, opossum, raccoon, and rabbits was also available. Perhaps the most important were birds, particularly turkeys and white ibis, and plant products. These filled the need for a good source of thiamin to counteract the problem of thiaminase. Thiamin (vitamin B,) becomes a critical limiting factor in a diet consisting largely of raw shellfish since uncooked fish or shellfish is very high in this particular enzyme (Greig & Gnaedinger, 1971; Bethke et al., 1973). The archeological record (Hrdlicka, 1922) and historical accounts (Barrientos, 1965: Solis de Meras, 1964) demonstrate that these Indians had no gross dietary inade- uacies. To supplement their animal diet, there was an ample vegetable larder. As a starch staple, the Calusa had both red coontie (Smilax spp.) and white coontie (Zamia sp.). Some zamia was gathered on the west coast, but the bulk was tribute from the subservient tribes of the east coast, where the plant was much more plentiful (Fontaneda, 1944). jon 1978 | AUSTIN & McJUNKIN, ETHNOFLORA 51 The Calusa, and probably their predecessors, subjugated the Indians of the eastern Florida coast and extracted tribute from them (Swanton, 1922). Calusas thwarted all Spanish attempts to Christianize them and to en- slave them for work in the gold mines of Hispaniola. This was in contrast to the Lucayans of the Bahamas, and a great proportion of the Matacumbe, Tequesta, and Jaega of southern Florida. Juan Ponce de Leon, a veteran of genocidal campaigns on both Puerto Rico and Santo Domingo, was repulsed and later died as a testament to Calusa ferocity. European diseases and slaving drastically reduced the total aboriginal population of Florida. Spanish slave-hunting from the south declined by the early 1600’s. In the mid-1600’s English and Creek slave raids started, but this time the primary pressure was on the tribes of northern and cen- tral Florida. These factors caused the Calusa to shrink from a once-power- ful tribe to a small group sequestered in the recesses of the lower western coast. From the early seventeenth century on, there is little or no record regarding the Calusa as a people or a culture. The next group to inhabit this region, and without a doubt Chokoloskee itself, was a mixture of Cuban fishermen and Indians. The Indians were the so-called “Spanish Indians” who have been the subject of much in- vestigation (Dodd, 1947; Goggin, 1950; Sturtevant, 1953; Neill, 1955). The most probable explanation of the Spanish Indians is that they were Calusa remnants: other accounts attempt to involve vanguard Seminoles and even Choctaws. The Indian-Cuban relationship was a peaceful, com- mercial arrangement advantageous to both. There were numerous fishing “ranchos” located along the west coast from Tampa Bay southward, with a peak population of around 400 (Ham- mond, 1973). The first English language report concerning the “ranchos’’ was from Roberts (1976). Bernard Romans, surveyor for the British Crown, observed a commerce in 1769 of over 100 tons per year in dried and salted fish and corncob-smoked fish roe with the Havana market (Romans, 1962). Turtles, shark liver oil, songbirds, and “manufactured goods” were also important export items. Cardinals, which brought six to ten dollars each, were caught with a bird lime made from gum elemi, an extract of the gumbo limbo tree (Bursera simaruba) (Williams, 1962; Covington, 1959). Transported live to Havana in willow cages, these birds were es- pecially favored among cigar makers, seamen, and the wealthier classes. Manufactured goods probably included surplus cordage termed “cuba.”’ This cord was made by women from one of the various native or introduced plants, e.g., Vucca aloifolia, Agave decipiens, Sansevieria metallica, Gos- sypium hirsutum, Hibiscus tiliaceus. Since this system of fishing ‘“‘ranchos” was thriving at the beginning of the English administration of Florida (1763-1783), it had been an on- going concern for quite some time, probably active to some degree by the late 1600’s. The largest of these “‘ranchos” comprised eighteen or twenty palm-thatched huts with a population of fifty to sixty people including women and children (Williams, 1962). About half of the male population was Indian. The Cuban men often took Indian wives, but there was at 52 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 least one case in which a Spanish woman became the wife of a Spanish Indian warrior (Sprague, 1964). This Indian later participated in the raid on Indian Key in which Dr. Henry Perrine was killed in 1840. Per- rine was at one time Consul to Yucatan and was particularly interested in the introduction of tropical crop plants into the United States (Robinson, 1942). In typical Spanish tradition (cf. Kimber, 1973), these ‘‘ranchos’”’ had small kitchen gardens in which corn, melons, several types of beans, limes, coconuts, and various kinds of aromatic herbs were grown. However, these dooryard gardens did not supply the bulk of their vegetables. The majority came from dealings with the many “plantations” in the area (Williams, 1962). In addition, coontie was obtained farther inland from less accul- turated Indians. These plantations were rarely over seven to eight acres and were located on the higher keys and on old kitchen middens and mounds along the rivers. Middens have since been favorite agricultural sites owing to the rich, black humus left when the tropical hardwood hammock is cleared. Elevation of the middens above sea level (8 to 20 feet) and their ease of access were added advantages. Corn, peas, melons, pumpkins, several kinds of beans, sweet potatoes, peppers, tomatoes, sugar cane, limes, oranges, and aroma- tic herbs were grown and sold at a high price to the relatively affluent fishermen. Most of the farmers were Cubans who seasonally employed several Indian families, but at least one Anglo-Saxon was recorded. John Durant, of Savannah, Georgia, farmed on Cape Romano in 1828, as did a mulatto whose name was not recorded (Williams, 1962). Mullet fishing was such a lucrative enterprise that it induced Captain William Bunce, the customs inspector of Key West, to establish several “ranchos” of his own in Tampa Bay in 1834 (Dodd, 1947). After the onset of the Second Seminole War (1835-1842), Washington officials, particularly Joel Poinsett, Secretary of War, were highly suspicious of the “ranchos” and their Spanish Indians, believing that they were supplying the beleaguered Seminoles. Bunce, along with other prominent citizens, asserted that the Spanish Indians rarely visited the mainland, and spoke Spanish, and that the Seminoles never recognized them as fellow tribes- men. Notwithstanding, the “ranchos’ were systematically closed down, and the Spanish Indians were shipped to Oklahoma (Foreman, 1953; Cov- ington, 1954). Bunce’s “ranchos’ were burned by the U.S. Army as a suspected “. . . hiding place for a party of renegade Spaniards, who had previously and at this time [sic] intercourse with the savage band . . .” (Phelps, 1847). In 1841 Lt. John T. McLaughlin, U.S.N., submitted to the Secretary of the Navy a map of southern Florida on which “Choko Lithka” appeared for the first time (Buker, 1975). The field reconnaissance for this map was made during the riverine campaign which decimated the Seminoles through every part of their former refugium, the Everglades. We do not know the particulars of the role of Chokoloskee during the Second Seminole War. From 1842 to about 1870, few people frequented the Chokoloskee area. 1978 | AUSTIN & McJUNKIN, ETHNOFLORA 53 Visitors were the occasional plume and hide hunters, the U.S. Army, and the Seminoles. There were also one or two Yankee families that intermit- tently farmed on the present site of Everglades City on the Barron River (then known as Potato Creek) in the 1860’s (Tebeau, 1955). In 1856 an Army scout reported that on Chokoloskee, Seminoles had harvested about an acre of corn, and that about eight miles to the south of Pavionyhatchee (presently Chatham River), a twenty-acre crop of sweet potatoes had been dug. This field was located on an extensive mound system and was set about with clumps of sugar cane (Tebeau, 1968). This is the same agricultural site which was reconnoitered in 1838 by a similar expedition of the Second Seminole War (Covington, 1958). The last major expedition of the Third Seminole War (1855-1858) was launched into the Big Cy- press Swamp from “‘Chokoliska Key.” Chokoloskee appears on the famous 1856 “Ives Military Map” of South Florida. This reliable map has the distinction of being the first vegetation map of southern Florida that is not strictly the product of an active imagination. The name Chokoloskee is from the Muskogean Seminole, ‘“chuko’’ — house, and “leske’ — old. The reason for this name is unknown (Simpson, 1956), but one may spec- ulate that it made reference to an ancient place of habitation. With the advent of modern settlement of Chokoloskee in the early 1870’s, signs of previous occupation were found in the form of several large lime trees (Tebeau, 1957). Latins from Key West formed about half of the population in these early years, with an array of “Anglos” from diverse parts constituting the other half. The prevalent house type of this period was the “trash house,” which had not only a palm thatch roof, but thatch walls as well. Some of these were quite elaborate, having a frame door with hinges. Usually a person had one in which to live and several that served as sheds. Trash houses were notorious fire hazards, so each family would have several sets se- quentially (cf. Simpson, 1920). There are no wild palms to be found on the island today, nor were there any when the Smallwood family moved there in 1897. The primary thatching palms were the cabbage palm (Sabal palmetto), the saw palmetto (Serenoa repens), and the Florida thatch palm (Thrinax radiata). There was also heavy pressure on the thrinax for its trunks for use as poles to construct turtle crawls for which they are high- ly suited (Small, 1925, 1929). Farming, fishing, hunting, and charcoal-making were the available liveli- hoods on Chokoloskee. Farming was primarily for the large population (20,000 in 1910) of Key West. Cigar-making was a flourishing industry near the turn of the century, and the southernmost city proved to be a reliable truck market (Cutler, 1923). Winter produce not sold in Key West was shipped via Mallory Line Steamers to New York. Both tem- perate and tropical crops were produced in great abundance. Tomatoes, cabbage, sugar cane, avocados, potatoes, peppers, cucumbers, eggplants, melons, onions, and bananas (the apple variety was preferred, although plantains or horse bananas tolerate more salt) were the main crops. Guavas, sugar apples, Jamaica apples, both sweet and sour oranges, cauliflowers, and ) 54 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 pumpkins were also sold. It seems strange that certain Cuban mainstays such as boniato, yuca, malanga, annatto, and beans were not produced, even though there were Cubans on Chokoloskee and Key West. Over three-fourths of the island was cleared and planted, with a preference for the high shell mounds for tomatoes, and the lower, heavier areas for sugar cane (Tebeau, 1955). In 1912 the Overseas Railroad was completed from the southeastern coast to Key West, and vegetables which could be pro- duced more cheaply on the mainland were shipped in by rail. With the railroad came the end of Chokoloskee’s farming era, although the large seedling avocado grove which spanned the island was still in heavy pro- duction until it was ruined by the 1926 hurricane. These ‘alligator pears” were packed in barrels and shipped north to Fort Myers and Punta Gorda, where as luxury items they brought a nickel each (Tebeau, 1955). All that presently remains of the grove is about 25 erratic seedling trees, from which all the fruit is either consumed or sold locally. Making charcoal from buttonwood (Conocarpus erectus) and cutting these trees for sale were also important occupations of the early pioneers on the island. Buttonwood charcoal, even-burning and smokeless, was pre duced for the Key West market. It was made as follows: wood was cut into 3- to 4-foot lengths, neatly stacked into great semiconical piles reminis- cent of wigwams, 6 to 9 feet high and 15 to 25 feet in diameter; each pile was then covered with sand and grass, leaving a vent in the top and sev- eral openings around the bottom to fire the wood and allow the gases to escape. If the wood cutters were not going to watch the fire all night after it was started, they would simply close up the vents. By this method, a cord of wood yielded ten bags of marketable charcoal. While activity ceased about 1930, as late as 1954 one could still see areas on the southwest Florida coast where the practice took place. Occasional work was also found at the tanbark factory at Shark River, where the red mangrove (Rhizophora mangle) yielded an exceptionally high-quality tannic acid. This factory was intermittently operated from 1908 to 1923. The gathering of mature royal palms (Rovystonea elata), used in development and landscape schemes in such places as Fort Myers and Tampa, also offered temporary work (Small, 1929). Since the 1930’s fishing has remained the chief occupation. Today com- mercial fishing for mullet and stone crabs is the mainstay of the economy, along with increasing tourism based on sport fishing and charter boats made famous by the Everglades Rod and Gun Club of Everglades City. A causeway was built from Everglades City to Chokoloskee in 1955; this helped the island’s depauperate economy, but ended its insular mystique. jean ETHNOFLORA As is widely known, man is one of the most effective vectors for plant dispersal. The present project began as an attempt to determine whether the plants now growing on the island could be used to trace occupation by the various people who have lived on Chokoloskee. We soon discovered 1978] AUSTIN & McJUNKIN, ETHNOFLORA ob) that the post-Columbian plants could be recognized and assigned, in part, to a definite period. Those plants native to Florida, on the other hand, presented greater difficulty. Many of the plants included in this list were commonly used by prehistoric and historic inhabitants of Chokoloskee, yet their wide distribution does not necessarily implicate man as a vector. Bursera simaruba, for example, may be found in almost any hammock along the coasts south of Tampa; it is distributed by birds which eat the seeds. Conversely, Sapindus saponaria has been cited by Craighead (1974) as an indicator of Calusa habitation. About 42 percent of the plants present, but not now cultivated, on Chokoloskee can be attributed directly to movement by man and to his activities. Many of these plants are exotics whose introduction into Florida can be traced to the 1800’s and 1900’s. Some plants have been brought in from the north (e.g., Lepidium, Portulaca), but the majority came from the West Indies and from Old World floras. There is good documentation for many of the species listed having been used by man in Florida. For others there is no record here, but usage has been recorded elsewhere. In such cases we have usually assumed that sim- ilar utilization occurred on Chokoloskee. No attempt has been made to cite original sources for the use of plants or their products; instead, we have used recent, readily available publications for documenting most of them. FLORISTIC COMPOSITION In southern Florida the high ground usually associated with hammock vegetation has long been favored for habitation. Anthropologists do not believe that the mounds on the Turner River north of Chokoloskee were built by the Calusa solely for the advantage of elevation (Sears, 1956). They were partially the by-product of living and were part of a religious complex. One cannot help believing, however, that these elevated sites were refugia during storms. Particularly convincing evidence for this conclusion was given by Jonathan Dickinson, who weathered a hurricane in an In- dian hut atop a shell mound on the Florida east coast in the 1600’s (Dick- inson, 1945). Accounts of early pioneer life of Chokoloskee (Tebeau, 1955) and other areas with hammocks (Tebeau, 1968; Pierce, 1970) indicate that the European and American settlers sought out hammocks for home- sites. Similarly, the Seminoles used hammocks as temporary and permanent campsites (Tebeau, 1968). A comparison of plant exploitation on Chokoloskee and in some other areas may be profitably made. Yarnell (1964) has reported that in the Great Lakes region about 20 percent of the vascular plants were used b Indians in pre-Columbian times. We know of no similar studies for other areas, but Griffin (1967) was of the opinion that this 20 percent usage might be taken as a standard of comparison. The data for Florida hammocks must be interpreted with caution (TABLE 1). Of the hammocks listed, the Boynton Hammock apparently has 56 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 TABLE 1. Ethnofloristic composition of certain Florida hammocks.* HAMMOCK NAME ETHNIC GROUPS USAGE ETHNOFLORA Chokoloskee Island Calusas, Spanish, Farming, 83% Seminoles, Americans settlement Lignum Vitae Key Matacumbes, Spanish, Settlement, 60% Americans farming Butts hammock Americans Settlement, 31% plant nursery Matheson hammock Tequestas, Seminoles, Settlement. 31% Spanish, Americans public park Indian Key Matacumbes, Spanish, Settlement 30% Boynton hammock Jaegas, Americans Campsite 27% Note: these figures are based on the assumption that all species with ; a aoe of use were indeed used. been least disturbed by man (Austin & Weise, 1972). Indian Key was occupied by Matacumbes and American pioneers, but has not been settled recently. Butts and Matheson hammocks (Avery, 1968; Austin, 1974) apparently have been subjected to little intrusion and disturbance. Only on Lignum Vitae Key (Avery, 1969; Popenoe & Avery, 1971) and on Chokoloskee are there records of both long and recent disturbance, and of the introduction of exotic plants. Thus the data seem to indicate that the ethnofloristic composition is not only a function of plant usage, but also of the recency of arrival and of the length of time used. Chokoloskee and Lignum Vitae have both the longest and the most recent histories of intrusion, and they correspondingly have the largest percentage of plants used. The 20 percent figure for aboriginal utilization may be more realistic than total ethnofloristic composition appears to indicate. For Boynton hammock the 27 percent status indicates only slight European/American influence. Of the 29 assumedly utilized plants, 11 are documented to have been introduced from Old World sources. Thus, aboriginal use may have been only 17 or 18 percent of the present flora. For Chokoloskee, 29 species can be clearly documented as post-Columbian. These account for 21 per- cent of the plants not presently cultivated. Some of the remaining plants cannot be categorized with confidence, but the number used by the Calusa is probably near 20 percent. For the ethnoflora of Chokoloskee, we have determined five categories of usage: 1) food and household plants, 50 species; 2) escaped or per- sistent ornamentals, 20 species; 3) medicinal and poisonous plants, 21 species; 4) homovectant weeds, 28 species; and 5) plants presently culti- vated, 90 species. Those species used for more than one purpose have been counted twice. The remaining 40 species (about 17 percent of the flora) 1978 | AUSTIN & McJUNKIN, ETHNOFLORA a7 consist of native plants having no apparent association with man. The ethnoflora in its broadest sense includes almost 83 percent of the island ora. ACKNOWLEDGMENTS This study spanned the period between January, 1972, and January, 1975, during which time many classes taught by the first author visited the island. The interest and questions of those classes aided considerably in the formation of the end product. Ms. Thelma Smallwood patiently answered many of our questions, as did Mrs. William Deem, Everglades National Park office at Everglades City. We particularly wish to thank G. N. Avery, J. Popenoe, and C. E. Wood, Jr., for many useful suggestions concerning the manuscript. LITERATURE CITED Apvams, C. D. 1972. Flowering plants of Jamaica. 848 pp. Univ. West Indies Press, Mona. Austin, D. F. 1974. Botanical survey of the Boca West tracts. [ Unpublished ——- & i G. WEISE. — weer ao of the Boynton Beach ham- mock. Quart. Jour. Fla. Acad. 35: aoe G. N. 1968. Checklist of the ae hammock, Miami. [ Unpub- ; 1969. Checklist of Lignum Vitae Key. ee eraae | BARRIENTOS, B. 1965. Pedro Menéndez de Avilés, founder of Florida. (English translation by A. KERRIGAN.) Floridiana Facsimile & Reprint Ser. [Original ed. printed in 1567.] 161 + 149 pp. Univ. Florida Press, Gainesville. BETHKE, G., S. CHANGBUMRUNG, & W. FELpHEIM. 1973. Eine Mikromethode zur Bestimmung der erythrocytaren Transketolase-Aktivitat. Int. Jour. Vi- tamin Res. 43: 426-437. Buxer, G. E. 1975. Swamp sailors: riverine warfare in the Everglades, 1835- 1842. 152 pp. Univ. Florida Press, Gainesville. Burkitt, I. H. 1966. A dictionary of the economic products of the Malay oo 2 vols. 2444 pp. Ministry of Agriculture, Kuala Lumpur, Ma- aysl acon. J. W., ed. 1954. A petition from some Latin-American fishermen: 1838. Tequesta 14: 61-65. 958. Exploring the Ten Thousand Islands: 1838. bid. 18: 7-13. 1959. Trade relations between southwestern Florida and Cuba: 1600- 1840. Fla. Hist. Quart. 38: 114-128. CRAIGHEAD, F. C. 1974. Hammocks of South Florida. 7: P. J. GLEASON, ed., Environments of South Florida: present and past. Mem. Miami Geol. Soc. 2: 53-60. Cusuinc, F. H. 1896. Exploration of ancient Key dwellers’ ee on the Gulf Coast of Florida. Proc. Am. Philos. Soc, 35: 329-43 Cutter, H. G. 1923. History of Florida. Vol. 1. 613 pp. aa Publ. Co., Chicago. D’Arcy, W. G., & H. EsupaucH. 1974. New World sae ea lanaceae ) north of Colombia: a resumé. Baileya 19: 93- 58 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Dickinson, J. D. 1945, Jonathan encanta journal or God’s protecting providence. (E. W. & C. M. — vs, eds. of 1699 edition.) x + 252 pp. Yale Univ. Press, New Have Dopp, D. 1947, Captain Sie. Tampa Bay fisheries, 1835-1840. Fla. Hist. Quart. 25: 246-256. FONTANEDA, H. D’'ESCALENTE. 1944. Memoirs of Do. d’Escalente Fontaneda re- specting Florida. (English translation by B. SmirH, 1854.) (Reprinted with revisions, D, O. True, ed.) 77 pp. Hist. Assoc. So. Fla. Misc. Publ. No Miami. FOREMAN, G. 1953. Indian removal. 415 pp. Univ. Oklahoma Press, Norman. GILLIs, W. T. ae Name changes for seed plants in the Bahama flora. Rho- dora 76: 67-1: Goccin, J. M. ce The Indians and history of the Matacumbe region. Te- 4 GREIG, R. A.. & R. H. GNaAEDINGER. 1971. Occurrence of thiaminase in some common aquatic animals of the United States and Canada. U. S. Natl. Marine Fisheries Ser. Spec. Sci. Rep. Fish. 631: 1-7. GRIFFIN, J. B. 1967, Eastern North American archeology: a summary. Science 156: 175-191. HaMMonp, E. A, 1973. Spanish fisheries of Charlotte Harbor, Fla. Hist. Quart. al 355 ae) Harvin, J. W., & J. M. Arena. 1974, Human poisoning a native and culti- vated plants. ed. 2, 194 pp. Duke Univ. Press, Durhan scans U.P., ed. 1919. Sturtevant’s notes on edible slants, N. Y. Dept. Agr. 7th Ann. Rep. 2: 1-686. ena C. B. 1969. Nightshades: the paradoxical plants. 200 pp. Freeman. San Francisco. Hrpiicka, A. 1922. The anthropology of Florida. Publ. Fla. State Hist. Soc.. No. 1. 68 pp. Ives, J. C. 1856. Military map of the pe ie ceas . Florida south of Tampa Bay. Scale 1:400,000. J. H. Colton & Co., New York. Kimber, C. T. 1973. Spatial patterning in : dooryard gardens of Puerto Rico. Geogr. Rev, 63: 6-26. Kiukas, R. W. 1969, Exotic terrestrial plants in South Florida with emphasis on the Australian pine ai equisetifolia). 7 pp. National Park Ser- vice. [Mimeographed report. | Lone, R. W., & O. one on A flora of tropical Florida. 962 pp. Univ. Miami Press, Coral Gables. MacCau_ey, C. 1884. The oo of Florida. Fifth Ann. Rep. Bur. Ethnol. Smithson. Inst. Pp. 469-531. Morton, J. F. 1968. Wild plants for survival in South Florida. 80 pp. Hur- ricane House, Miami. ——— & B. R. Lepin. 1952. 400 plants of South Florida. 134 pp. Text House, MUENSCHER, W. C. 1951. Poisonous plants of the United States. 277 pp. Mac- millan Co., New York. NEILL, W. T. 1955. The identity of Florida’s “Spanish Indians.” Fla. An- throp. 8: 43-57. 1956. Preparation of rubber by the Florida Seminole. Zbid. 9: 25-28. Parsons, J. J. 1972. Spread of African pasture grasses to the American tropics. Jour. Range Managem. 25: 12-17. 1978 | AUSTIN & McJUNKIN, ETHNOFLORA 59 Puetes, S. S. 1847. Senate document no. 52. 29th Congress, 2nd Session, Jan. 11, 1847. Pierce, C. W. 1970. Pioneer life in southeast Florida. (D. W. Curt, ed.) 264 pp. Univ. Miami Press, Coral Ga les PoreNoE, J., & G. N. Avery. 1971. Survival of ornamental plants on Lignum Vitae Key. Proc. Fla. State Hort. Soc. 84: 368-370. Reap, R. W. 1975. a genus Thrinax (Palmae: Coryphoideae). Smithson. Roperts, W. 1976. - account of the first discovery and natural history of Florida, with a particular detail of the several expeditions and descents made on that coast. (Introduction and index by R.-L. Gorp:) xxix 4- xP 102 + 15 pp. Floridiana Facsimile & Reprint Ser. Pe ed. printed in 1763.] Univ. Florida Press, Gainesville. Rogprnson, T. R. 1942. Henry Perrine, pioneer horticulturist of Florida. Te- questa 1: 16-24. Romans, B. 1962. A concise meee history of east and west Florida. 342 pp. Floridiana Facsimile & Reprint Ser. [Original ed. printed in t775,\° Univ, Florida Press, Gainesville. Sears, W. H. 1956. The Turner River site, Collier County, Florida. Fla. An- t Smupson, C. T. 1920. In lower Florida wilds. 404 pp. G. P. Putnam’s Sons, New York Simpson, J. C. 1956. Florida place names of Indian derivation. 158 pp. Fla. Geol. Surv. Spec. Publ. No. 1. Tallahassee. SMALL, J. K. 1910. Additions to the flora of peninsular Florida. Bull. Torrey Bot. Club 37: 513-518. 1919. Cape Sable region of Florida. 27 pp. New Era Printing rays Lancaster, Penn. 1925. Silk-top thatch. Jour. N. Y. Bot. Gard. 26: 49-54. 1929. From Eden to Sahara. 123 pp. Science Press, Lancaster, Penn. Manual of the southeastern flora. 1554 pp. Univ. North Caro- lina Press, Chapel Hill. Sotis pE Meras, G. 1964. Pedro Menéndez de Avilés. (English translation by _T. ConNoR.) 286 pp. Floridiana Facsimile & Reprint Ser. [Original ed. printed in 1567.] Univ. Florida Press, Gainesville. — J. T. 1964. The origin, progress and conclusion of the Florida War. 57 pp. Floridiana Facsimile & Reprint Ser. [Original ed. printed in 1848. | Uae Florida Press, Gainesville. SrurTEVANT, W. C. 1953. Chakaika and the Spanish Indians. Tequesta 13: 35-73. on Swanton, J. R. 1922. Early history of the Creek Indians and their neighbors. 492 pp. Smithson. Inst. Bur. Am. Ethnol. Bull. TeBeau, C. W. 1955. The ae of the Chokoloskee Bay country. 88 pp. Univ. Miami Press, Coral Gab . 1957. Florida’s last oe the history of Collier County. 260 pp. Univ. Miami Press, Coral Gables 1968. Man in the Biverelades: 192 pp. Univ. Miami Press, Coral Gables. WILLEy, G. 1949. Archeology ys - Florida Gulf Coast. 559 pp. Smithson. Misc. Coll. Vol. 113, Publ. 60 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Wi.iiams, J. L. 1962. The territory of Florida, sketches of the topography. civil and natural history. 304 pp. Floridiana Facsimile & Reprint Ser. [ Original ed. printed in 1837.] Univ. Florida Press, Gainesville. YARNELL, R. A. 1964. Aboriginal relationships between cultures and plant life in the upper Great es region. vi + 218 pp. Anthrop. Pap. Mus. Anthrop. Univ. Michigan No. APPENDIX. The Ethnoflora. A. PLANTS NOT PRESENTLY IN CULTIVATION ON CHOKOLOSKEE ISLAND. Aren Acacia farnesion (L.) Willd. (Fabaceae). West INDIAN BLACKTHORN. Cul- tivated for aromatic flowers and medicinal gur Acacia ne (L.) Willd. (Fabaceae). fase ACACIA. Wood used outside Abrus ears L. (Fabaceae). Rosary PEA. Seeds poisonous (Hardin & 974). Albizia lebbeck Bentham (Fabaceae). WoMAN’S oa Cultivated as an orna- mental and for medicine (Morton & Ledin, Alternanthera ramosissima (Martius) Moq. ex ives (Amaranthaceae). CHAFF- known use man. Amaranthus viridis L. (Amaranthaceae). Picwreep. Homovectant. Annona squamosa L. (Annonaceae). SUGAR APPLE. Cultivated for foo Ardisia escallonioides Schlecht. & Cham. (Myrsinaceae). MARLBERRY. ‘Leaves used by the Mikasukis to season their tobacco (Morton, 1968). Argemone mexicana L. (Papaveraceae). MExIcAN poppy. Both herbage and seeds are poisonous (Muenscher, 1951), but the plants have been used as a narcotic Arundo donax L. (Poaceae). GIANT REED. Ornamental. Atriplex arenaria Nutt. (Chenopodiaceae). BEACH-orACH. No known use by man. sees hy minans (L.) L. (Avicenniaceae). BLACK MANGROVE. According to Morton (1968), the seeds are edible when cooke Batis maritima L. ae SALTWoRT. Edible raw or evden (Morton, 1968). Bidens pilosa L. (Asteraceae). ap eee NEEDLES. Leaves edible, juice medicinal on, 1968); leaves used for Blechum brownei Juss. ae nese Borreria ocimoides (Burman f.) DC. (Rubiaceae). No known use by man. Bryophyllum pinnatum Kurz (Crassulaceae), LivE-FOREVER. Medicinal (Mor- ton in, 1952). Bumelia is H.B.K. (Sapotaceae). SAFFRON PLUM. Fruits edible (Mor- ton, 1968). Bursera tae (L.) Sarg. (Burseraceae). GuMBo Limpo. Sap used as bird ne; soft wood used for floats on nets and harpoons; leaves used as tea (Hedrick. 1919) and cattle food (Romans, 1962). ae ta coe (L.) R. Br. (Fabaceae). GRay NICKER. Seeds used as ma- ri y (Morton & Ledin, 1952). ee a. i. corres JAMAICA CAPER TREE. No known use by man, Capparis flexuosa L. (Capparaceae). FLEXIBLE CAPER. No known use by man. Capraria biflora L. (Scrophulariaceae). GoATWEED. Homovectant. 1978] AUSTIN & McJUNKIN, ETHNOFLORA 61 Capsicum frutescens L. (Solanaceae). BIRD PEPPER. Fruits preserved in vine- gar ee a as condiments (T. Smallwood, pers. comm.; D’Arcy & Esh- baugh, Cariospermam Vein Le — HEART SEED. Ornamentals (Long & Lakela, 1971); used as food in Ceylon Carica ee L. (Caricaceae), PAPAYA, PAPAW. Leaves, stems, fruits, and roots en: leaves smoked as a remedy for asthma (Morton, 1968). Cassia obtusifolia L. (Fabaceae), SICKLE POD, COFFEE WEED. Homovectant; sed as coffee substitute. Cenchrus echinatus L. (Poaceae). SANDBUR. Homovectant. Cenchrus incertus M. A. Curtis (Poaceae). Homovectant. Chamaesyce blodgettii (Engelm. & Hitchc.) Small (Euphorbiaceae). SPURGE. Chamaesyce hirta (L.) Millsp. (Euphorbiaceae). SpURCE. Homovectant. Chamaesyce hypericifolia (L.) Small (Euphorbiaceae). SPURGE. Homovectant. Chamaesyce hyssopifolia (L.) Small (Euphorbiaceae). SPURGE. Homovectant. Chamaesyce porterana Small (Euphorbiaceae). SPuURGE. Homovectant. ik alba (L.) A. Hitchc. (Rubiaceae). SNowBerry. No known use by Chysophllum oliviforme L. (Sapotaceae). SaTINLEAF. Fruits edible; wood d (Small, 1933); now an orname ee sinensis (L.) Osbeck (Rutaceae). Swi EET ORANGE (temple variety). Cul- Clerodendrum philippinum Schauer (Verbenaceae). CHRISTMAS ROSE, JAPANESE rnamenta Coccoloba Ci Jacq. (Polygonaceae ). PIGEON PLUM. Fruits eaten by Mikasukis (Morton, 1968). Spe diffusa Burman f, (Commelinaceae). DAYFLOWER. Homovectant. nocarpus erectus L. (Combretaceae). Burronwoop. Wood used in making 1 in the early 1900’s Crotalaria incana L. (Fabaceae). RATTLEBOx. Homovectant. Crotalaria pallida Aiton (Fabaceae). Homovectant. Cynanchum sp. (Asclepiadaceae). Not seen in fruit or flower; no known use by Cyperus ligularis L. (Cyperaceae). Sawcrass. No known use by m Dactyloctenium aegyptium (L.) Richter (Poaceae). EGYPTIAN GRASS. Homo- ctan ve t. a tortuosum (Sw.) DC. (Fabaceae). Tick TREFOIL. Introduced as a rage - now naturalized. ihe assurgens (L.) Juss. (Acanthaceae). No known use by m Dioscorea bulbifera L. (Dioscoreaceae). POTATO VINE. oe ‘used as 00 Distichlis spicata (L.) Greene (Poaceae). SALT Grass. No known use by m Eragrostis ciliaris (L.) R. Br. (Poaceae). LOvE GRASS. ene eet Erythrina herbacea L. (Fabaceae). CORAL BEAN. Seeds poisonous. Eugenia nee re Willd. Crees WHITE STOPPER. rer? edible; sed for cabinets (Small, Ne Eugenia pene Pers. en a ce STOPPER. Perhaps used like £ llaris. Eupatorium odoratum L. (Asteraceae). No known use by m Ficus aurea Nutt. (Moraceae). STRANGLER FIG. Fruits ace “Tatex used by Mikasukis to prepare a masticatory rubber (Neill, 1956). 62 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 Fimbristylis caroliniana (Lam.) Fernald (Cyperaceae). SEDGE. No known use y man. Fimbristylis spathacea Roth (Cyperaceae). SEDGE. Homovectant. sae sack eh aes ae Krug & Urban (Oleaceae). FLORIDA PRIVET. No k se by n creee macreei M. . Curtis (Fabaceae). Mi_k PEA. No known use by man Gouania lupuloides (L.) Urban (Rhamnaceae). CHEW STICK. Stems used as othbrushes; arene to have medicinal properties (Hedrick, 1919). Hamelia patens Jacq. (Rubiaceae). FIREBUSH. Ornamental. Heliotropium angiospermum Murray ( (Boraeinacese ). No known use by man. epee undatus (Haw.) Britton & Rose (= Cereus undatus Haw.) (Cacta- VIGHT-BLOOMING CEREUS. Ornamental. Hyptis. ae (L.) Poiret (Lamiaceae). Homovectant. Ipomoea alba L. (Convolvulaceae). MoonFLOWER. Most often cultivated for flowers, although leaves said to be edible (Morton, 1968). Ipomoea carnea Jacq. (Convolvulaceae), BUSH MORNING GLORY, Ornamental. Ipomoea hederifolia L. (Convolvulaceae). Quamoc.ir. Cultivated and na- turalized. Ipomoea indica (Burman f.) Merrill (Convol vuleceae) MORNING GLORY. No known use in Florida. Cultivated elsewhere Ipomoea macrantha Roemer & Schultes (Convolv cipceae BEACH MOONFLOWER. young stems are used as pig food in Great Inagua, but no known use in Florida. Ipomoea triloba L. (Convolvulaceae). Probably introduced from West Indies in ballast. Tresine diffusa Humb. & Bonpl. ex Willd. (Amaranthaceae). BLoopLEar. No known use by man. Kalanchoé grandiflora A. Rich. (Crassulaceae). LIFE-PLANT. Ornamental. Nat- uralize Laguncularia racemosa (L.) Gaertner f. in Gaertner (Combretaceae). WHITE MANGROVE. Wood occasionally used Lantana involucrata L. (Verbenaceae). BEACH SAGE. No known use by man. Lantana sass Ae (Verbenaceae). LANTANA, Fruits may be poisonous M Lepidium virginicum L. (Brassicaceae), PEPPER-GRASS. Used by some as a con- din We nent (Morton, 1968). Now a weed on Chokoloskee and perhaps homo- eae Ligustrum ovalifolium rick = L. amurense Carr. 2?) (Oleaceae). PRIVET. Per- sistent from cultivation as an ornamental. Lycium cag ae W alter (Solanaceae). CHRISTMAS BERRY. No known use yr Eaten by birds, particularly the mockingbird Tr ee foetidissimum ( Jacq.) Cronquist (cai ceie), MastTIc. Vhile now absent from the island, there were three plants in the early 1900’s (Thelma Smallwood, pers. comm., 1974). Fruits eaten: wood used (MacCauley, Melanthera parviflora Small (Asteraceae). No known use by m — sano L. (Meliaceae). CHINABERRY. Persistent or “cco nat- Ornamental; wood formerly used (Small, 1933 joie Ga L. foes CREEPING CUCUMBER. ae eaten (Mor- 8) although strongly laxative (Hardin & Arena, 1974). — floridana Nutt. (Loasaceae). POOR-MAN’S-PATCH. No known use by aid to be the inspiration for “Velcro.” 1978] AUSTIN & McJUNKIN, ETHNOFLORA 63 Morinda royoc L. (Rubiaceae). CHEESEWEED. No known use by man. Momordica charantia L. (Cucurbitaceae). BALSAM APPLE. Medicinal and poi- sonous (Morton, 1968; Hardin & Arena, 1974), but used as food. ee ee var, dillenii ve L. Benson (Cactaceae). PRICKLY PEAR. s and perhaps pads e ’ Oxalis eee Ps we LaApy’s SORREL. Eaten by some people as greens. The plants are largely homovectant. Panicum maximum Jacq. (Poaceae). GUINEA GRaAss. Introduced as pasture s by the 1780's (Parsons, 1972); now widely naturalized. Parietaria floridana Nutt. (Urticaceae). PELLitory. Probably homovectant. Parthenocissus quinquefolia (L.) Planchon (Vitaceae). VIRGINIA CREEPER. No k . Fruits said to be poisonous. Pectis leptocephala (Cassini) Urban (Asteraceae). Used in a tea. Persea americana Miller (Lauraceae). Avocapo, During the early 1900's groves of these trees covered much of the southern end of the island (Tebeau, 1955). The 1910 hurricane damaged the crop, but fruit continued to be produced commercially. In 1926 a hurricane destroyed most of the trees, market production ceased. There are still scattered trees on oa ee Petiveria alliacea L. (Phytolaccaceae). GUINEA-HEN WEED. Homo Philoxerus vermicularis = ) Beauv. (Amaranthaceae). Morton ( mee os that th s may be n. Phlebodium aureum (L. ’ Sm. (Polypodiaceae). CABBAGE PALM FERN. No known use by ma tee inctsier L. (Arecaceae). DATE PALM. Fruits eaten. Plants persis- te m cultivation ae australis (Cav.) Trin. ex Steudel (Poaceae). REEpD. Often called P. communis Trin., but Adams (1972) and Gillis (1974), among others, have pointed out that this is a later name. Roots eaten raw, and sap used as sugar (Hedrick, 1919). Piscidia ae (L.) Sarg. (Fabaceae). FisH POISON TREE. Bark used to pot- son Pisonia seet L. (Nyctaginaceae). PULL-AND-HOLD-BACK. No known use by ma me unguis-cati (L.) Bentham (Fabaceae). Cat’s-cLtaw. Arils eaten. Plumbago scandens L. (Plumbaginaceae). WILD pPLUMBAGO. No known use, ex- cept possibly ornamental. pe cyathophora (Murray) ae & Garcke (Euphorbiaceae). WILD NSETTIA. No known use by rv ie oleracea (L.) Sarg. ne ey PursLaneE. The seeds used in by the Seminoles (Morton, 1968); the herbage eaten. Still culti- vated in the Bahamas. Psidium guajava L. (Myrtaceae). GuAvA. Introduced for edible fruits; nat- uralized. lag nervosa Sw. (Rubiaceae). WILD re Seeds may have been used coffee substitute if the name is indic aii seen L. (Rubiaceae). ae BERRY. ea (1966) lists the ber- as a dye in the West Indie Rhabdadenia biflora (Jacq.) rai AGE: (Apocynaceae). RUBBER VINE. No e by man eae mangle L. (Rhizophoraceae). — MANGROVE. Wood used for cabi- s; bark used for dying and tannin 64 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Rhynchelytrum repens (Willd.) C. E. Hubbard (Poaceae). NATAL Grass. Intro- duced as a pasture grass (Small, 1933); now widely naturalized. Rhynchosia parvifolia DC. AMebaceae): SNOUT BEAN. No known use by man. Rivina humilis L. (Phyt ). BLoopBerry. No known use by man sets pane ie mis Schlecht. & Cham. (Scrophulariaceae). CORAL PLANT. ntal and persistent. Gee metalic Gérome & Labroy (Liliaceae), BowstRInG HEMP. Culti- ordage. Persistent and naturalize Sapindus ee L. (Sapindaceae). SOAP BERRY TREE. Leaves and fruits used soap substitute. Sco paria no L. (Scrophulariaceae). SWEET BROOM. Leaves used as a medicinal tea in the Philippines (Hedrick, 1919). Sesbania mile Raf.) Cory (Fabaceae). No known use by man. Sida acuta Burman f. (Malvaceae). INDIAN MALLOW. Burkill (1966) stated that has been used as a medicine and a fiber source in the Old World. aes americanum Miller (Solanaceae). NIGHTSHADE. Green fruits polson- ripe fruits edible (Heiser, 1969). Solidave stricta Aiton (Asteraceae). GOLDENROD. No known use in Florida. Sonchus oleraceus L. Cae Introduced and naturalized. Leaves and roots edible (Hedrick, 1919 ees tomentosa L. ates NECKLACE pop. Said to have been used as allucinogen by Indians. ee patens (Aiton) Muhl. (Poaceae). Corpcrass. Although sometimes se are a normal part of the coastal marsh vegetatio ee domingensis (Trin.) Kunth (Poaceae). DropsrEep. Perhaps oe as a for Stachytarpheta ie Satine (L.) Vahl (Verbenaceae). BLUE PORTERWEED. Used the Old World (Burkill, 1966) and to brew a porter in Sas esta Se (L.) B.S.P. (Fabaceae). PENcIL-FLOWER. No known use Suaeda eee (Ell.) Moq. (Chenopodiaceae). SEA BLITE. Leaves edible. meat stans (L.) Kunth (Bignoniaceae). YELLOW ELDER. Medicinal, but per- only ornamental and naturalized on Chokoloskee. ee Sta ‘opine (L.) Solander ex Correa (Malvaceae). SEASIDE MAHOE. Small (1910) recorded the first appearance of this tree on the mainland from a ee in Dade pe ia in 1905, although it had been in the Keys years. No known use by aon: ignee aie tes ee WILD PINE. Cultivated for or- and persis Tillandsia sect oe rose, oS AIR PLANT. Only one plant been found on the island. No known Tillandsia Weeaces Sw. (Bromeliaceae). ot WILD PINE. Only one plant has been found. No known use. Tillandsia setacea Sw. (Bromeliaceae). NEEDLE-LEAF WILD PINE. Persistent from cultivation at an old homesite Toxicodendron radicans (L.) Kuntze (Anacardiaceae). Porson ivy. Used as a medicine by some groups of people EIGHOSERING octandrum (L.) H. W alter ken cotie Small (1910) re- this plant in Florida only on Chokoloskee. George N. Avery (pers. 1972) has found it recently; we ane not seen the plant. Tridax ae L. (Asteraceae). Homovectant. 1978 | AUSTIN & McJUNKIN, ETHNOFLORA 65 Verbesina virginica L. (Asteraceae). FROSTWEED. No known use b an. Vigna luteola (Jacq.) Bentham (Fabaceae). Cow pea. No known use in » Florida. Waltheria indica L. (Sterculiaceae). WALTHERIA. Small (1933) said that the were “reputed to have medicinal properties.” Homovectant. Cn fagara (L.) Sarg. (Rutaceae). WILD LIME. No known use of this specie nee others are known as “toothache tree’ because of numbing proper B. PLANTS PRESENTLY CULTIVATED ON CHOKOLOSKEE ISLAND. Acalypha dia Muell.-Arg. (Euphorbiaceae), CopPERLEAF. Ornamental. Agave americana L. (Agavaceae). CENTURY PLANT. Ornamental. Two varieties occur ao island: var. marginata Trel. and var. variegata Hort. Allium re L. (Amaryllidaceae). ONION. ds Aloé barbadensis Miller = A. vera L.) (Liliaceae). BARBADOS ALOE. Orna- men d medicina Anthurium sp. ee “Ornam tal. Araucaria columnaris (Forster) nae (Araucariaceae). NORFOLK ISLAND PINE. Ornamental. Possibly A. heterophylla ( (Salisb.) Franco also; not seen. Arecastrum romanzofianum Becc. (Arecaceae). QUEEN'S PALM. Ornamental. Asparagus densiflorus (Kunth) Jessop ( (Liliaceae). ASPARAGUS FERN. Orna- mental. Asparagus setaceus (Kunth) Jessop (Liliaceae). ASPARAGUS FERN. Ornamental. Bains variegata L. (Fabaceae). ORCHID TREE. Ornamental. Bombax malabaricum DC. (Bombacaceae). RED SILK-COTTON TREE. Ornamental. Bougainvillea glabra Choisy (Nyctaginaceae). BOUGAINVILLEA. Ornamental Brassaia actinophylla F. Mueller (Araliaceae). UMBRELLA TREE. Coe Brassica oleracea L. (Brassicaceae). CABBAGE. Too Callistemon Lor aan Sw. (Myrtaceae). BOTTLE-E BRUSH. Ornamental. sie odorata Hooker f. E Thomson (Annonaceae). YLANG-YLANG. Orna- al; aromatic flower Ca mitts Lour. ee. FISHTAIL PALM. Ornamental. Casuarina equisetifolia L. (Casuarinaceae). AUSTRALIAN PINE. Ornamental; ind-brea Casuarina glauca Sieb. (Casuarinaceae). BEEFWoop. Ornamental. Catharanthus roseus (L.) Don (Apocynaceae). PERIWINKLE. Ornamental. naturalized in Florida. Cereus pases s Haw. (Cactaceae). HepcGe cactus. Ornamental. Chrysalidocarpus lutescens Wendl. (Arecaceae). CANE PALM. Ornamental. Citrus Hee ea Swingle (Rutaceae). Lime, Key Lime. As an ornamental and for food. Citrus ae L. (Rutaceae). SOUR ORANGE. Ornamental. Citrus paradisi Macf. (Rutaceae). ee For food and as an ornamental. Coccoloba uvifera L. (Polygonaceae). SEA GRAPE. Ornamental. Cocos nucifera L. (Arecaceae). COCONUT. ne an ornamental; rarely for food. Lethal yellowing, a mycoplasma-induced, lethal eee has not yet hit the island, although it is in Everglades City Codiaeum pata (L.) Blume (Euphorbiaceae) Croton. Ornamental. mental. Cordia sebestena L. (Boraginaceae). GEIGER TREE. Ornam Cordyline pane (Forster) Hooker f. (Liliaceae). meee Ornamental. Coleus blumei Bentham (Lamiaceae). COLEUS. Ornamental. Crinum sanderianum Baker (Amaryllidaceae). ve AND WINE LILY. Ornamen- tal 66 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 Cryptostegia madagascariensis Bojer (Asclepiadaceae). PURPLE ALAMANDA. Or- menta Cucurbita moschata Duchesne (Cucurbitaceae). SQUASH. Food. Cyperus alternifolius L. (Cyperaceae). UMBRELLA PLANT. Ornamental. Delonix regia (Bojer) Raf. (Fabaceae). PorncrANA. Ornamental. Dizygotheca elegantissima Vig. & Guill. in becom: (Araliaceae). FALSE ARALIA. Ornamental. Epiphyllum sp. (Cactaceae). NIGHT-BLOOMING CACTUS. Ornamental. Eriobotrya japonica (Thunb.) eae (Rosaceae). Loguat. As an ornamental, and for shade and edible fruit ee grandis Hill ex Maiden (Myrtaceae). BRowN Gum. Ornamental. Eugenia uniflora L. (Myrtaceae). SURINAM CHERRY. Ornamental. Euphorbia lactea Haw. (Euphorbiaceae). CANDELABRA CACTUS. Ornamental. Euphorbia milii Moulins (Euphorbiaceae). CROWN-OF-THORNS. Ornamental. Euphorbia tirucalli L. (Euphorbiaceae). PENCIL TREE. Ornamental. Ixora coccinea L. (Rubiaceae). Ixora. Ornamental. Juniperus conferta Parl. (Cupressaceae). SHORE JUNIPER. Ornamental. Lactuca sativa L. (Asteraceae). Letruce. Foo Lycopersicon esculentum (.) Miller’ (Solanaceae): Tomato. Food. Mangifera indica L. (Anacardiaceae). Manco. For food and as an ornamental. Manilkara zapota (L.) Royen (Sapotaceae). SApopiLLta. As an ornamental and or food. Melicoccus bijugatus Jacq. (Sapindaceae). SPANISH LIME. Ms. Smallwood told us that two trees remained on the island until they were cut in 1974. Orna- l mental. Monstera deliciosa Liebm. (Araceae). CERIMAN. Ornamental; fruit is some- imes eate sae dale: var. oom Kuntze (Musaceae). BANANA. As an orna- rarely for Mating calabura L. ia cae STRAWBERRY TREE. As an ornamental a fruit ea olean ider 7 (Ap yocynaceae). OLEANDER. Ornamental. Pedilanthus Ce re Poit. (Euphorbiaceae). SLIPPER-FLOWER. Ornamental. Pelargonium hortorum Bailey (Geraniaceae). GERANIUM. Ornamental. Phaseolus vulgaris L. (Fabaceae), KIDNEY BEAN. Food. Phoenix roebellinii O’Brien (Arecaceae). DATE PALM. Ornamental. Phoenix begs T. Anderson (Arecaceae). ae A few of these trees were planted near the Smallwood Store by Mr. C. Smallwood. Aas eer (Cav.) C. Chr. ( Seen STAGHORN FERN. Or- namental. Pleomele marginata (Lam.) N. E. Br. (Agavaceae). DRACENA. ala Plumbago capensis Thunb. (Plumbaginaceae). LEapwort. Ornamental. Plumeria rubra L. eae FRANGIPANI. Orna sah ee (Willd.) R. Graham (iintiovbiaree). POINSETTIA. Orna- ntal pan clas filicifol ta (Moore) Bailey (Araliaceae). ARALIA. Ornamental. Polyscias guilfoyle: (Bull.) Bailey (Araliaceae). ARALIA. Ornamental. * According to the Standing Committee on Stabilization of Specific Names (Taxon 24(1): 174. 1975), the correct name for the tomato is Lycopersicon lycopersicum (L.) Karsten ex Farwell. However, Terrell (Taxon 26(1): 129-131. 1977) comes to a dif- ferent conclusion and accepts the name used by the authors of the present paper. Ed. 1978 | AUSTIN & McJUNKIN, ETHNOFLORA 67 Pouteria campechiana (H.B.K.) Baehni (Sapotaceae). CANISTEL. For fruit and as an ornamental. Prunus persica (L.) Batsch (Rosaceae). PEACH. Fruit t ee granatum L. (Punicaceae), POMEGRANATE. For ae and as an orna- ntal. Rnphidoplor aurea (Linden ex André) Birdsey (Araceae). PoTHos. Ornamen- pesee communis L. (Euphorbiaceae). CaAsTOR BEAN. Ornamental. Rosa sp. (Rosa ree Rose. Ornam Roystonea regia (H.B.K.) Cook (Arecaceae). ROYAL PALM. Ornamental. sae eae ae Lodd. ex Schultes (Arecaceae). CABBAGE PALM. When e Smallwood family moved to the island in 1897, there were no palms of ne kind on Chokoloskee. The largest Sabal on the island now is beside the Smallwood Store; C. S. Smallwood planted the tree in 1919, and it now has about ten feet of trunk. Other small palms were planted near the store by Thelma Smallwood. The palms planted were from the mainland, they remain abundant. Ornamental. Sanchezia speciosa Leonard ea. SANCHEZIA. Ornamental. Schinus terebinthifolius Raddi (Anacardiaceae). PEPPER TREE. Ornamental. Widely naturalized in southern Florida. Setcreasea purpurea Boom (Commelinaceae). PURPLE QUEEN. Ornamental. Spathodea campanulata Beauv. (Bignoniaceae). AFRICAN TULIP TREE. Orna- tal. Swietenia mahagoni Jacq. (Meliaceae). Manocany. Ornamental. Syzygium cuminii (L.) Skeels (Myrtaceae), JAMBOLAN PLUM. As an ornamental and for fruit ete indicus L. (Fabaceae). TAMARIND. As an ornamental and for fruit. Tecomaria ae (Thunb.) Spach (Bignoniaceae). CAPE HONEYSUCKLE, Or- nt namental. Terminalia catappa L. (Combretaceae). TROPICAL ALMOND. Ornamental. Thrinax radiata Lodd. ex J. A. & J. H. Schultes (Arecaceae ). THATCH PALM. ee to be native to Giokeloskes but no longer found in the wild (Read, 75). Ornamental. Thuja orientalis L. (Cupressaceae). ARBOR VITAE. Ornamental. Tillandsia usneoides L. (Bromeliaceae). SPANISH MOSS. ae ental. Veitchia ee (Becc.) Moore (Arecaceae). CHRISTMAS PALM. Ornamental. Yucca aloifolia L. (Liliaceae). SPANISH DAGGER. Ornamental. Zea mays L. a. Corn. Food. D.F.A. D. M. DEPARTMENT OF BIOLOGICAL SCIENCES DEPARTMENT OF GEOGRAPHY FLoripA ATLANTIC UNIVERSITY FLortpA ATLANTIC UNIVERSITY Boca RaTON, FLoripA 33431 Boca RATON, FLORIDA 33431 PRESENT ADDRESS: DEPARTMENT OF GEOGRAPHY UNIVERSITY OF CALIFORNIA Los ANGELES, CALIFORNIA 90024 68 JOURNAL OF THE ARNOLD ARBORETUM VOL. 59 DATES OF PUBLICATION OF SARGENT’S SILVA OF NORTH AMERICA * Paut D. SORENSEN FEW WHO HAVE HAD OCCASION to open its pages would disagree that the Silva of North America, by Charles Sargent, ranks high among the most sumptuous botanical works ever published. Its lavish format matches the generous bibliographic, descriptive, and anecdotal information it contains. Its artistic merits extend to nearly all aspects of its production including the superb quality of the line drawings by C. E. Faxon, the skill and crafts- manship of their engraving and printing, and the clarity of the type face. Houghton, Mifflin and ae originally offered the volumes at $25 each, or $300 for the twelve volumes planned. During the eight years spanned in preparing the first twelve es much additional information came to light on new species in the plant families already covered. To accommodate these new species, Sargent added a two-volume supplement, volumes 13 and 14. These also sold for $25 per volume, bringing the price for the full fourteen-volume work to $350 (Sargent, 1902-1913). The high cost (by late nineteenth century standards) required that the Silva depend chiefly on pre-publication subscriptions as the main manner of sale, thereby avoiding expensive inventories of unsold books. As a result the work passed out of print rather soon. It was reprinted (by a photo- offset method) in 1947 by Peter Smith of Magnolia, Massachusetts. The first reprinting was sold by the early 1960's. Mr. Smith, who still owns the copyright to the Silva, has told me he plans to issue another printing of the facsimile edition for release late in 1977. Sargent’s Siva is an important historical reference and one likely to be overlooked owing to its somewhat limited distribution. Its historical im- portance is based on the fact that Sargent was one of the first and most influential champions of nomenclatural priority, and the Silva is the first major treatise covering a large and diverse number of North American plants in which the rule of priority was strictly applied The Silva does not contain any new plant names. It includes only those species first described and published elsewhere; it does, however, resurrect many names from obscurity. Since the date when a certain name, though not new, has been recalled from earlier literature and used in place of an- other may be of some interest or importance, the publication date of the work where such a name reappeared could be critical. When I recently had occasion to consult the Silva, the edition available to me at the time was the 1947 Peter Smith facsimile reprint, in which I discovered that all * Sargent, Charles Sprague. The Silva of North America; a description of the tree which grow naturally in North America exclusive of Mexico. Illustrated with nee and analyses drawn from nature by C. E. on 14 vols. Houghton, Mifflin a : pany, Boston. 4°. 740 plates. 1890-190 1978 | SORENSEN, SARGENT’S SILVA 69 seven volumes (the fourteen volumes of the original bound in seven) bore the same title page which I assumed had been that of the original Volume 1. The entire work thus gave the appearance of having been published in 1890 and contradicted the sequence of issuance according to the account of Sutton (1970). In turning to the original edition for clarification on publication dates, I encountered another discrepancy: the title page of the original Volume 1 carries the date 1891, a year later than the date given in the reprint, but in this volume the copyright is dated 1890, which prob- ably represents the source of the “1890” used on the title page of the re- print. Which is the correct date? In what year did the volume actually appear? att searching for these answers, I eventually learned that in vol- umes 1, 3, and 11, the title page or co pyright dates disagree with the actual dates of beat I have arrived at the determination of correct publica- tion dates, as outlined in TABLE 1, by searching published reviews, library catalogues, and notes by Sargent as well as those of Sutton (1970), his biographer. The details of these sources appear among the notes pertain- ing to each volume presented below. For most of the volumes, corrobora- tion of publication dates comes largely from Tucker (1914), Catalogue of the Library of the Arnold Arboretum. Here the catalogue entry Cpe023) lists exact dates for every volume except Volume 1, which is given merely as 1891. The printing of an exact date for thirteen of the fourteen volumes of the Silva aroused my curiosity as this practice seems not to occur else- where in the Catalogue, or, if it does, to do so inconsistently. I suspect that the exact dates indicate the dates when the volumes arrived at the library and were accessioned. One would expect accuracy from this entry inasmuch TABLE 1. Dates of publication of the Silva of North America by Charles Sprague Sargent ITLE Copy- VOLUME PAGE RIGHT PLATE ADOPTED NUMBER DATE DATE NUMBERS PUBLICATION DATE SOURCE 1 1891 1890 1- 50 October 1, 1890 Sargent, Preface to Vol. 1 Z, 1891 1891 51- 97 May 23, 1891 Tucker 3 1892 1891 98-147 January 21, 1892 Tucker 4 1892 1892 148-197 July 2, 1892 Tucker 5 1893 1893 198-251 December 30, 1893 Tucker 6 1894 1894 252-300 May 18, 1894 Tucker fl 1895 1895 301-355 February 1, 1895 Tucker 8 1895 1895 356-438 September 20,1895 Tucker 9 1896 1896 439-496 March 16, 1896 Tucker 10 1896 1896 497-537. November 28,1896 Garden & Forest (Dec. 9) 1896 ipl 1897 1897 538-592 April 27, 1898 Tucker 12 1898 1898 593-620 January 10, 1898 Tucker 13 1902 1902 621-704 December 15,1902 Tucker 14 1902 1902 705-740 December 15,1902 Tucker 70 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 as the Catalogue comes from the same institution as the Silva. But, as I point out later, in addition to the 1891 date given for Volume 1, at least one other of the dates given by Tucker (1914) lacks accuracy. I hope that my clarifications will aid those who, as I, have been vexed by the prob- lem of publication dates in this work. NOTES ON THE SEPARATE VOLUMES VoLuME 1. Magnoliaceae to Ilicineae | Aquifoliaceae|]. Pp. i-ix, 1-119, plates 1-50. Publication date: October 1, 1890. In the preface to Volume 13 (the first of a two-volume supplement), Sargent (1902) gives the publication date of Volume 1 as October, 1890. Sutton (1970, p. 150) states that, “volume one of Sargent’s Silva appeared in 1891.” She probably derived this date from the title page of Volume 1 as noted above. But a page later (p. 151), she quotes from a review of the Silva which appeared in the January 3, 1891, issue of Gardener’s Chronicle (anon., 1891a). It seems unlikely that a review would appear in a British periodical on the third day of the new year unless the review- er had had the work in hand a few weeks earlier. Additional reference sup- porting an 1890 publication date comes from an article in Garden and Forest (Britton, 1890), of December 3, 1890, announcing the appearance of Volume 1. Since Sargent himself established the date in October of that year, I pro- pose October 1, 1890, as the day of publication. VoLuME 2. Cyrillaceae to Sapindaceae. 117 pp., plates 51-97. Publication date: May 23,1891 (Tucker, 1914). An unsigned review of Volume 2 of the Silva appeared in the June 10, 1891, issue of Garden and Forest (anon., 1891b). There the phrase, “The second volume of this work which has just been issued . . .” corroborates Tucker’s accession date of May 23, 1891. VotuMmE 3. Anacardiaceae to Leguminosae. 141 pp., plates 98-147. Publication date: January 21, 1892 (Tucker, 1914 The words, ‘The third volume . . . which has feccally appeared . taken from a review in Garden and Forest dated February 17, 1892 ce 1892), suggest that the date provided by Tucker comes very close to the actual date of publication, and I have adopted it here. ) VoLUME 4. Rosaceae to Saxifragaceae. 141 pp., plates 148-197. Publication date: July 2, 1892 (Tucker, 1914 ) The celebrated Liberty Hyde Bailey published a review (1892) of Vol- ume 4 in the July 20, 1892, issue of Garden and Forest. The interval be- tween the date of this review and the earlier date of Tucker as noted above argues in favor of adopting her date for the date of publication. 1978 | SORENSEN, SARGENT’S SILVA jal VoLuME 5. Hamamelidaceae to Sapotaceae. 189 pp., plates 198-251. Publication date: December 30, 1893 (Tucker, 1914). The issue of Gardener’s Chronicle dated December 2, 1893, carries a brief notice of the appearance of Volume 5 of the Siva (anon., 1893a). An expanded review of this volume came out in the same periodical a week later (anon., 1893b). In neither place do we find a direct reference to a publication date. Therefore, I adopt the one given by Tucker. VoLumE 6. Ebenaceae to Polygonaceae. 124 pp., plates 252-300. Publication date: May 18, 1894 (Tucker, 1914). A review of Volume 6 was printed in the Gardener’s Chronicle (anon., 1894) nearly four months after the date noted by Tucker. VoLtumeE 7. Lauraceae to Juglandaceae. 173 pp., plates 301-355. Publication date: February 1, 1895 (Tucker, 1914). Neither of two reviews (anon., 1895a, 1895b) reports a date of publica- tion which would contradict that of Tucker. VotuME 8. Cupuliferae [Fagaceae], Quercus. 190 pp., plates 356-438. Publication date: September 20, 1895 (Tucker, 1914). Reviews of Volume 8 (Buchenau, 1895; anon., 1896a) are not in con- flict with the date of publication assigned by Tucker. Votume 9. Cupuliferae [Fagaceae] to Salicaceae. 190 pp., plates 439— 496. Publication date: March 16, 1896 (Tucker, 1914). A copy of the original edition held by the combined libraries of the Arnold Arboretum and the Gray Herbarium, Harvard University, Cam- bridge, bears on the title page an unsigned penciled note giving the Volume 9 publication date as March 16, 1896. This is the same date Tucker has used. Buchenau (1896) wrote a review (dated October 16, 1896) of Volume 9 for a German periodical. VoLumE 10. Liliaceae to Coniferae [Pinaceae sensu lato|. 159 pp., plates Publication date: November 28, 1896 (anon., 1896b). In a review of Volume 10 in the December 9, 1896, issue of Garden and Forest, I find the only precise publication date in periodical literature for any installment of the Silva. It happens also that among the “exact day” dates of Tucker (1914), only this one is in error. Tucker sets the date of publication as November 30. This two-day discrepancy between the date shown in Garden and Forest and that given by Tucker represents the evidence upon which I conclude that the dates Tucker records reflect the times of accession of these volumes to the Arnold Arboretum library. A review of Volume 10 appeared also in Gardener’s Chronicle of January 2, 1897 (anon., 1897a). A second and longer article (anon., 1897b) in the iz JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 same periodical further deals with Volume 10 and adds a general essay on American trees, especially as revealed in the Silva up to that time. VoLuME 11. Coniferae [Pinaceae], Pinus. 163 pp., plates 538-592. Publication date: April 27, 1898 (Tucker, 1914). In this year Garden and Forest ceased publication. The magazine was not an official organ of the Arnold Arboretum, but had served as a vehicle for Sargent to express his views (Sutton, 1970). Many of the unsigned articles such as the reviews cited here were actually written by Sargent (Sutton, pers. comm.). VoLuME 12. Coniferae | Pinaceae except Pinus]. 144 pp., plates 593-620. Publication date: January 10, 1899 (Tucker, 1914). During the years Sargent worked on the Si/va, many new species of woody plants from North America and elsewhere came to light. These Sargent (1902-1913) collected in a two-volume folio work entitled Trees and Shrubs. The original covers of these volumes are bound into the copies held by the library of the Arnold Arboretum (Jamaica Plain). Printed on these covers is an advertisement for the Si/va, accompanied by notes on the scope of the work. The advertisement announces the date of appearance of Volume 12 as January, 1899, thus substantiating the date recorded by Tucker. VoLUME 13. Rhamnaceae to Rosaceae. 184 pp., plates 621-704. Publication date: December 15, 1902 (Tucker, 1914) This is the first of the two-volume supplement to the originally planned twelve-volume work. Both volumes appeared on the same day (see below). Sargent dated the preface in the first “June 1902.” The interval from June to December when Tucker recorded the December 15 date seems a reasonable time to allow for printing. VoLuME 14. Caricaceae to Coniferae | Pinaceae]. 52 pp., plates 705-740. Publication date: December 15, 1902 (Tucker, 1914). ACKNOWLEDGMENTS I wish to thank the Directors of the Arnold Arboretum and the Morton Arboretum for the use of their respective library facilities while searching these matters, and the Council of Academic Deans, Northern Illinois Uni- versity, for a summer research stipend LITERATURE CITED Anonymous. 1891a. Gard. Chron, III. 9: 12. January 3, 1891. ————, 1891 b.. Gard. Forest 42:274..275.. June 1891, 1892. Ibid. 5: 83, 84. February 17, 18 . 1893a. Gard. Chron. III. 14: 690. ran 2, 1893. 1978 | SORENSEN, SARGENT’S SILVA ~I Ww ——. 1893b. Jbid. 720. December 9, 1893. ——. 1894. Ibid. 16: 286. September 8, 1894. —, 1895a. Gard. Forest 8: 119. March 20, 1895. a = S80 5beocard: Chron. Il 372 560.. May:4, 1895. ———. 1896a. /bid. 19: 52. January 11, 1896. ———. 1896b. Gard. Forest 9: 499. December 9, 1896. ————, 1897%a. Gard. Chron. III. 21: 9. January 2, 1897. . 1897b, Zbid. 101. February 6, 1897. BAILEY, L. H. 1892. Gard. Forest 5: 346, 347. July 20, 1892. Britton, N. L. 1890. Gard. Forest 3: 590. December 3, 1890. BUCHENAU, F. 1895. Bot. Zeit. om 53: 265-269. SARGENT, C. S. 1890. Silva ie North America 1: preface, p. viii. Houghton, 1902. Ibid. 13: preface, p. Vv. ed. 1902-1913. Trees and shrubs. 2 vols. Houghton, Mifflin & Co., me Boston. Sutton, S. B. 1970. Charles Sprague Sargent and the Arnold Arboretum. Har- vard University Press, Cambridge, Massachusetts. Tucker, E. M. 1914. Catalogue of the Library of the Arnold Arboretum. Cos- mos Press, Cambridge, Massachusetts. DEPT. OF BIOLOGICAL SCIENCES NORTHERN ILLINOIS UNIVERSITY DEKA Lp, ILLINOIs 60115 74 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 TWO NEW SPECIES AND A NEW SUBGENUS OF CYCLANTHACEAE GEORGE J. WILDER A NEW SUBGENUS AND SPECIES OF DICRANOPYGIUM IN 1954 HARLING ESTABLISHED the genus Dicranopygium, with 16 spe- cles, as a segregate from Carludovica (Harling, 1954). Subsequently, in his Monograph of the Cyclanthaceae, Harling (1958) recognized 44 species in this genus. He later described three more, increasing to 47 the known num- ber of Dicranopygium species (Harling, 1963, 1972, 1973). In his mono- graph Harling also described three subgenera of Dicranopygium, i.e., Dt- CRANOPYGIUM, URIBAN THUS, and GLEASONIANTHUS. While in Panama the present writer discovered a new species of Dicran- opygium, D. Harling G. Wilder, which exhibited a combination of features precluding placement in any of Harling’s subgenera. The species is pres- ently described and placed in its own subgenus, ToMLINSONIANTHUS. Material of Dicranopygium Harlingii was collected by G. Wilder and R. Dressler on April 29, 1973, from a single population in a region called La Eneida or Altos de Pacora, east of Cerro Jefe, Panama, at an altitude of ca. 750 meters. Plants were preserved in FAA and subsequently stored in an aqueous solution of 5 percent glycerine, 50 percent ethanol. All measurements and illustrations of D. Harlingii (except the illustra- tion in Ficure 18) are of this preserved material. After being studied, some of the specimens including the holotype were pressed and placed on deposit at the Field Museum of Natural History, Chicago, Ilinois (Field Museum sheet nos. 1764553-1764555). Aside from indicated exceptions, the description of flowering material is based solely on study of one inflorescence from the holotype (FIGURES 1-3, 1). On this inflorescence all spathes were found bent outward in the manner typical of spathes on mature cyclanthaceous inflorescences, and were intact when material was collected and preserved. This inflorescence bore staminate flowers, which appeared to be at a stage just prior to an- thesis (FiGuREs 6-17), and expanded staminodes (F1cures 2, 3). How- ever, spathes were also counted on 21 additional immature inflorescences (none from the holotype) which had peduncles sufficiently expanded to enable clear distinction to be made betwen spathes and peduncle sheaths (sensu Harling, 1958). The description of fruiting material is based on study of eight infructescences, none from the holotype. During investigation of the holotype, the distal portion of the studied inflorescence was severed from the basal part, as illustrated in FrcureE 2. In addition, the two most terminal spathes were removed from the pedun- cle, and several flowers were excised from the spadix. However, both por- tions of the inflorescence and all spathes were subsequently affixed to the herbarium sheet bearing the holotype. fy 1978 | WILDER, CYCLANTHACEAE rie) Some preserved specimens of Dicranopygium Harlingit (as well as of Lu- dovia Bierhorstii) were stained either with toluidine blue or methyl green before being photographed. In accordance with Harling’s (1958) usage, the term “‘stigma” is here- in applied to the entire upper surface of a cyclanthaceous carpel, not mere- ly to the hair-bearing portion (stigmatic crest). Presently described taxa, like all Cyclanthaceae, produce solely unisexual flowers. Dicranopygium subgenus Tomlinsonianthus G. Wilder, subg. nov. Type species: D. Harlingii G. Wilder. Descriptio subgeneris: receptaculum florum masculorum planum. Peri- anthium verum brevissimum, lobis unilateralibus in latere florum pistilla- torum spectantibus positis. Stamina pauca ad multa, latrorsa et introrsa, thecis antherarum ad basim plerumque distincte divergentibus sed inter- dum plus minusve parallelae et approximatae. Filamentum et bulbus ba- salis plana atque lata. Ubi thecae divergentes, connectivum cum filamento et bulbo basali lamellam plus minusve triangularem formans. Staminate flowers having a distinctly marked, excentrically attached pedicel (i.e., pedicel inserted on the side of the receptacle opposite the associated pistillate flower); pedicel solid; the receptacle flat; a true (sensu Harling, 1958), though extremely short, perianth present, its lobes unilaterally placed on the side facing the associated pistillate flower, the glanduliferous portions extremely reduced and often not outwardly visible; stamens few to many, having both lateral and introrse dehiscence (based on orientations of non-dehisced thecae), anthers with thecae usually dis- tinctly divergent at base but occasionally more or less parallel and ap- proximate. The filament and basal bulb broad, flattened, and not outward- ly discrete from one another. Where thecae are divergent, the connective is broad and flattened: in this case connective, filament, and basal bulb jointly forming a lamella which tends to be + triangular. This subgenus contains one species, a species known only from the single Panamanian locality. The name ToMLINSONIANTHUs is given in honor of Dr. P. B. Tomlinson, who has undertaken numerous important investiga- tions of the anatomy and morphology of monocotyledons, and who original- ly suggested this family as a subject of my research. Dicranopygium Harlingii G. Wilder, sp. nov. Ficures 1-18. Holotype from material preserved in FAA (lacking collection number), collected April 29, 1973, at La Eneida, east of Cerro Jefe, Panama, G. Wilder & R. Dressler (F, no. 1764553). Laminae 7-17 cm. longae, fere ad 1/2-2/3 bipartitae, segmentis lan- ceolatis, acuminatis. Spathae 3-4. Spadix ellipsoidea in statu florifero (in holotypo) et in statu fructifero. Flores masculi 1.6—3 mm. longi; lobi perianthii 2-4, positis unilateralibus in latere florum pistillatorum spec- 76 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 ead 4 eo + G we FIGURE 1. reels of aero es a Harlingii, drawn from preserved speci- men, X 0.5. ow L indic evel below which leaves had already abscissed naturally ont main axis. ‘S and C are basal portions of expanding, imma ture vegetative branches, Numerous roots are inserted on main axis up to lev zl of lowermost intact leaves. I, ii inflorescence (arrow points to spadix) ; Ro, root; S, two lowermost spat 1978] WILDER, CYCLANTHACEAE 77 tantibus, carnosi et maxime breves. Stamina 3-14, connectivo, bulbo ba- sali, et filamento plus minusve triangularibus. Flores feminei 3.0-3.8 mm. lati. Tepala brevissima. Stigmata sessilia, desuper visa plana ad leviter convexa, plus minusve elliptica. Plant terrestrial and colonial. Stem up to at least 25 cm. long, 1.4 cm. in diameter. On fertile stems the leaf blades 7-17 cm. long, bifid to ca. 1/2-2/3 of their length; the segments 16-35 mm. broad, lanceolate, acuminate; petiole and leaf sheath together 3.5—8 cm. long, the margins of the sheath tearing away from the medial portion during ontogeny, this medial portion then appearing petiole-like. Peduncle of flowering speci- men (holotype) 49 mm. long. Spathes 3 to 4, their basal portions white, the distal parts pigmented (at least sometimes red). On holotype 4 spathes are present: the lowermost one 32 mm. long, 15 mm. broad at base, nar- rowly triangular, long attenuate; the second, 29 mm. long, 22 mm. broad at base, broadly triangular, long attenuate; the third, 22 mm. long, 22 mm. broad at base, broadly ovate, cuspidate; the uppermost 17 mm. long, 18 mm, broad at base, broadly ovate, acuminate. Spadix pauciflorous, on flowering specimen (holotype) ellipsoid, 14.5 mm. long, 10 mm. wide meas- ured to tips of staminate flowers, at least sometimes having a very short, fleshy, encircling, collar-like structure at its base, ie., situated between the lowermost flowers and the uppermost spathe; spadix ellipsoid in fruit- ing stage, 9.5-21.5 mm. long, 7.5-11.5 mm. thick. Staminate flowers 1.6- 3 mm. long, receptacle 0.6-1.7 mm. in diameter; perianth lobes 2 to 4, only present on the side facing the associated pistillate flower, fleshy and extremely short, rarely approaching 0.5 mm. long, varying in shape from barely raised, laterally extended ridges to triangular; stamens 3 to 14 (up to 16 in cultivated material), placed in a whorl on the receptacle margin, commonly with one to five additional stamens situated on the receptacle surface, 1.6—3 mm. long, the anthers 0.5—0.8 mm. long, each theca ca. 0.4— .5 mm. broad, the connective, basal bulb, and filament jointly + triangu- lar, 0.3-1.2 mm. long, 0.5-1 mm. broad at base. Pistillate flowers 3.0—3.8 mm. broad; tepals extremely short, appearing as barely raised fleshy ridges along the edges of the flowers; stigmas sessile, seen from above vir- tually flat to slightly convex, more or less elliptic, the hairy portion of each stigma appearing either straight or curved when projected onto the transverse plane; staminodes ca. 3 cm. long. Seeds 0.7—0.9 mm. long, 0.25— 3 mm. thick This species is named for Dr. Gunnar Harling, who has contributed very greatly to our knowledge of cyclanths in his Monograph of the Cyclantha- ceae (1958) and other important works pertaining to the family. Dicranopygium Harlingii clearly belongs to the genus Dicranopygium because it has dispersed rather than distichous foliage leaves, these having bifid, thin laminae; 3 to 4 spathes per inflorescence, these inserted im- mediately beneath the spadix, the two lowermost spathes + triangular; ovaries which extend beneath the level of surrounding tissues, giving the impression that pistillate flowers are connate; four parietal placentae per 78 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 5 FIGURES 2-5. Material of Dicranopygium Harlingii: 2-4, from holotype; _ another specimen from type locality. 2, 3, distal portion of inflorescence “I” i FIGURE 1, X 1, X 1.9, eee Three lo pene spathes intact when ee photographs were made. Note all spathes inserted a a spadix. 4, pistillate flower from inflorescence in Ficures 2 and 3, * 13. Lowermost staminode excised just above its base; yey ee es of remaining staminodes visible. 5, acs x S2 ote proximity of dark-staining 1978 | WILDER, CYCLANTHACEAE 79 ovary; perianth lobes on staminate flowers relatively reduced, never cov- ering the stamens, and borne in one series; anthers lacking secretion globules; and terete, sculptured seeds. Harling (1958) employed only features of staminate flowers to dis- tinguish the three subgenera of the genus Dicranopygium which he rec- ognized, i.e., DICRANOPYGIUM, URIBANTHUS, and GLEASONIANTHUS. These features are listed in the left column of TaBLe 1. The four columns to the right compare his three subgenera with subgenus TOMLINSONIANTHUS according to these features. As indicated in TaBLEe 1, subg. TOMLINSONIANTHUS differs from subg. DIcRANoPYGIUM according to theca orientation, size and shape of con- nective and filament, and degree of discreteness of basal bulb; from subg. UriBANTHUus in shape of receptacle; and from subg. GLEASONIANTHUS in presence vs. absence of a “real perianth.” Harling (1958) stated that “real perianths” of the subgenera D1cRANo- PYGIUM and UrIBANTHUs differ in two ways from the perianth type (“‘pseu- doperianth’’) of subg. GLEASONIANTHUS. First, in a real perianth, the lobes have glandular tips, whereas in a pseudoperianth they never do. Second, Harling indicated that in subg. GLEASONIANTHUS perianth lobes clearly represent “metamorphosed stamens” because abnormal lobes bear rudi- mentary anthers at or near their tips; in contrast, he found no such evi- dence that real perianths of the subgenera DicraNopycruM and URI- BANTHUS are interpretable as staminodes. Among staminate flowers of Dicranopygium Harlingii examined through the dissecting microscope at 50, only one perianth lobe was found which appeared to have a glandular tip, i.e., a relatively large lobe below an un- usual, eight-microsporangiate stamen (FicurE 7, P; the glandular region is too small to be evident in this photograph, however). Despite the gen- eral paucity of such glandular tissue in D. Harlingii, the nature of the aforementioned lobe, together with the total absence of lobes having rudi- mentary anthers, led the present writer to conclude that D. Harlingi (and hence subg. ToMLINSONIANTHUS) has a “real perianth.” The overall scarcity of glandular tissue in this perianth is attributed to extreme re- duction of the perianth itself, a phenomenon emphasized by the minute di- mensions of the perianth in comparison to those of other species of the genus Dicranopygium. In addition, pseudoperianth lobes on staminate flowers of subg. GLEASONIANTHUS are always extremely elongate (Har- ling, 1958, 1963), unlike those of D. Harlingit. Because Dicranopygium Harlingii exhibits features in common with each of Harling’s three subgenera (TABLE 1), a choice was necessitated be- tween assigning D. Harlingii to a new subgenus (TOMLINSONIANTHUS), placing it in one of Harling’s three subgenera, or considering its interme- diate features as a basis for merging two or all of these subgenera. This writer excluded the second alternative since inclusion of D. Harlingt in spathe scars to spadi x. P, tepal; S1, $2, S3, three lowermost spathes on inflores- cence, numbered in acropetal order of levels of insertion on peduncle; St, stami- node. TABLE 1. A comparison of the subgenera of Dicranopygium.* Characters of Subgenus staminate DICRANOPYGIUM URIBANTHUS GLEASONIANTHUS TOMLINSONIANTHUS flower (41 species) (1 species) (4 species ) (1 species ) Occurrence of real perianth ** present present absent present flat or some- deeply hollow, flat flat receptacle of times shallowly funnelform staminate concave Type of anther only seldom introrse mostly introrse lateral and introrse (based dehiscence introrse on orientations of non- dehisced anthers ) Orientation of nearly always distinctly distinctly usually distinctly divergent thecae of stamen parallel and divergent at divergent at at base but occasionally generally pase base more or less parallel approximate and approximate Shape of con- nective and filament Discreteness and shape of basal bulb in most species rather thin distinct; globose to slightly oblong, or sometimes flattened broad and flattened jointly with connective and filament forming a triangular lamella; flattene broad and flattened jointly with connective and filament forming a triangular lamella; flattened filament broad and flattened. connective usually so, ex- cept when thecae are parallel and approximate jointly with filament and usually also connective (where thecae are divergent ) forming a triangular la- mella; flattened + Taformation on all s ** See text for explanation of this t ubgenera ak TOMLINSONIANTHUS is from Harling (1958, p. 275). 08 WOLAYOUUVY GIONYV AHL AO TYVNUNOL 6S “10A] 1978 | WILDER, CYCLANTHACEAE 81 any of Harling’s three subgenera would radically alter our concept of that subgenus, given the importance Harling (1958) placed on features char- acterizing each of the subgenera. The third alternative was likewise re- jected since differences separating Harling’s (1958) three subgenera do not appear sufficiently bridged by D. Harlingii. The decision to place D. Harlingii in its own subgenus was also influenced by Harling’s (1958) con- clusion that his three subgenera are “. . . well characterized and undoubt- edly very natural main groups.” It is believed, however, that discovery of one or more additional spe- cies of the genus Dicranopygium which, similar to D. Harlingii, exhibit particular features in common with various of the subgenera, might pro- vide sufficient justification to combine two or more of the four subgenera into one. Harling (1958) stated that the most primitive of his three subgenera “ . . is certainly Dicranopygium, the staminate flowers of which are in all respects on a lower evolutionary level than those of the remaining groups.” If he is correct, evolution of the subgenera may have proceeded as diagramed in Ficure 19. In all subgenera other than DicrANoPyYGIUM, there evolved a characteristic stamen type typically having divergent thecae and a continuous flattened lamella consisting of modified basal bulb, filament, and connective (FiGURE 19, stage A). In accordance with Harling’s interpretation, there also evolved in subg. URIBANTHUS a deeply hollow, funnelform receptacle and pedicel (FicurE 19, stage B), and in subg, GLEASONIANTHUS a “pseudoperianth” concomitant to loss of a real perianth (FicurE 19, stage C). If we assume that the subgenera differ mainly according to the features considered, subg. ToOMLINSONIANTHUS would thus differ less from subg. DicrANopyciuM than from either subg. URIBANTHUS or subg. GLEASONIANTHUS, and would thus represent a morphological if not a phylogenetic intermediate between DIcRANOPY- GiuM and the other two subgenera. As a member of subgenus TOMLINSONIANTHUS, Dicranopygium Har- lingi differs from all three species of the genus Dicranopygium described since publication of Harling’s (1958) monograph, i.e., D. Campi Harl. (Harling, 1972), D. macrophyllum Harl. (Harling, 1963), and D. robus- tum Harl. (Harling, 1972). Two of these species, D. Campi and D. ro- bustum (based on Harling’s (1972) description), belong to subgenus Dt- CRANOPYGIUM, while D. macrophyllum is included in subgenus GLEASON- IANTHUS (Harling, 1963). A NEW SPECIES OF LUDOVIA In his Monograph of the Cyclanthaceae, Harling (1958) recognized eleven cyclanthaceous genera, but stated that only two of these genera, Lu- dovia and Pseudoludovia, bear consistently entire adult leaves. He rec- ognized two species of Ludovia (L. lancifolia Brongn. and L. integrifolia (Woodson) Harl., described, respectively, in Brongniart (1861) and Wood- son and Schery (1942)), and one of Pseudoludovia (P,. Andreana Harl., 82 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 FicuRES 6-11. Staminate flowers from holotype of Dicranopygium Harlingii, < 14. All flowers photographed from above while inserted on spadix. Trows not Beene by symbols indicate boundary of flower described in correspond- 1978 | WILDER, CYCLANTHACEAE 83 described by Harling (1958)). No new species of these genera have been reported since publication of Harling’s monograph. The present description of Ludovia Bierhorstii G. Wilder is limited to one clone of this species under cultivation, although sterile plants that may belong to L. Bierhorstii were preserved in the field. This limitation pre- cludes description of a mixed collection; it is necessitated by indications (detailed herein) that criteria for distinguishing between sterile plants of Ludovia and Pseudoludovia are both inadequate and suspect Cyclanthaceae having solely entire adult leaves and not necessarily in- cluding only Ludovia Bierhorstii were collected by G. Wilder and E. Her- nandez between July 28-30, 1974, in Narino, Colombia, in wet, relatively level forest situated along the north side of the road between Altaquer and Junin. In this region palms as well as other Cyclanthaceae (including species of Asplundia, Cyclanthus, Dicranopygium, and Sphaeradenia(?) ) were common. At the time of collection in 1974, it was assumed that all entire-leaved cyclanths at the collecting locality were conspecific, so no records were kept regarding specific habitats occupied by individual specimens. Because this assumption is now uncertain (for reasons to be discussed), a precise description of habitat(s) of Ludovia Bierhorstii cannot be provided. How- ever, the range of habitats collectively occupied by all entire-leaved cy- clanths observed at the collecting locality is described below. These plants typically grew as epiphytes on dicotyledonous trees bearing abundant bry- ophytes. The cyclanths were more plentiful on large, thick-stemmed trees than on small ones. Individuals situated on upright tree stems grew either upward or downward and were usually rooted at their basal ends in bry- ophytes and organic matter on the tree stems; their distal portions usually extended outward from the trees. However, some entire-leaved cyclanths were also found at ground level on and around a fallen tree. These plants tended to grow horizontally and upright. Numerous specimens were preserved shortly after collection, and cut- tings were also retained in living condition. Preservation was accomplished with either FAA, a formalin-ethanol solution, or 10 percent formalin, and plants were later stored in an aqueous solution of 5 percent glycerine, 50 percent ethanol. Herbarium specimens were prepared from living material ing figure legend. 6, flower bearing three stamens, all inserted on receptacle mar- in. 7, flower bearing eight stamens, ail on receptacle margin. Median stamen anther Sp reatly because of ee Triangular perianth member beneath trated from side and from below, respectively, in FicurEs 12, 13. 8, flower bearing seven stamens on receptacle margin, two stamens on receptacle surface. r u ur ine stamens on receptacle margin, five on receptacle ee) "Flower appears diamond- a ed. 84 JOURNAL OF THE ARNOLD ARBORETUM [voOL. 59 Ficures 12-17. Staminate flowers from holotype of Dicranopygim Harlingit, x 16. 12, 13, flower from FicuRE 7 seen from a facing eee pistillate flower (FIGURE 12); from below (FicureE 13). Three perianth me only the central one appearing triangular. ve in FicureE 13, ere inserted 1978] WILDER, CYCLANTHACEAE 85 while near the collecting locality and also, later on, from preserved speci- mens. Sheets prepared at the collecting locality are on deposit at the her- barium of the Universidad de Narifo in Pasto, Colombia. Sheets pre- pared from sterile specimens preserved in liquid in Colombia, as well as from specimens of cultivated Ludovia Bierhorstit preserved in FPA, are on deposit at the Field Museum of Natural History (F). Cuttings of entire-leaved Cyclanthaceae from the collecting locality were brought into cultivation at Fairchild Tropical Garden, Miami, Florida (This material has remained sterile and therefore cannot be identified with certainty.), and subsequently at the University of Illinois at Chicago Circle, All material at the latter institution belongs to Ludovia Bierhorstit. The holotype of Ludovia Bierhorstii initially consisted of a leafy shoot bearing two outwardly visible inflorescences (FIGURE 20, I). However, during investigation of the holotype, the distal portion of each inflorescence was severed from the basal part. Subsequently, one of the inflorescences was completely dissected so that only one remains. On the remaining inflorescence, before preservation, the three spathes had unfolded; however, staminodes were intact and anthers had not yet dehisced. The herbarium material of the holotype exhibits this inflorescence (its staminate flowers having been largely removed from one side), its spathes, and part of the leafy shoot. In the following description of Ludovia Bierhorstii, all characterizations of vegetative axes and foliage leaves (excluding measurements of maximum length of sheath, petioles, and lamina, which were done on material pre- served in liquid) are based on study of living material, as is all mention of fragrance and color. In contrast, all description of structure and size of the peduncle, spadix, spathes, peduncle sheaths, flowers, and floral parts is based on specimens preserved in FPA. To avoid inclusion of ob- viously juvenile features, only foliage leaves 10 cm. or more long were described. With staminate flowers, measurement of receptacle diameters was made from bottom (rather than top) view of the receptacles, since this method was more accurate. During this investigation the writer studied one inflorescence each of Ludovia lancifolia and L. integrifolia for comparative purposes. Obtained, respectively, from plants cultivated at Fairchild Tropical Garden and the University of Illinois at Chicago Circle, both inflorescences, preserved in liquid, were at anthesis. The cultivated plants of L. integrifolia were originally collected as cuttings by R. Dressler and G. Wilder between mid- April and early May, 1973, near the El Llano-Carti highway, from 8 to 20 km. north of El Llano, prov. Panama, Panama. on side of receptacle opposite perianth (therefore also opposite associated eae flo tilla late flower). 14-16, flower seen from side facing associated pistillate flower (FicuRE 14), from opposite side (FicurE 15), and from side Se aaa ange to the other two (FIGURE 16). Flower resembles that in Frcures 12 and 13 but lacks teratological stamens; it bears three tepals, visible only in FicuREs 14 and 16. 17, flower viewed fro m side facing associated pistillate flower, eee only two tepals. P, perianth member. Pe, pedicel. 86 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 FIGURE 18. ee ioe cal a at the type locality. Plants of this species, indicated by arrows, growing in dried- up creek bed amidst boulders, logs, and fallen leaves. 1978 | WILDER, CYCLANTHACEAE 87 URIBANTHUS GLEASONIANTHUS A cal / / / / / / ‘i TOMLINSONIANTHUS | ze sw | / ~~ / =s- —iSlage B— /*—Stage Mg i / <——— Stage A 19 DICRANOPYGIUM FicurE 19. Diagrammatic representation of putative phylogeny of subgenera of Dicranopygium. Ludovia Bierhorstii G. Wilder, sp. nov. Ficures 20-44. Holotype from cultivated material (lacking collection number) pre- served in FPA, originally collected July 28-30, 1974, near Altaquer, Colombia, G. Wilder & E. Hernandez (¥, no. 1764556). Laminae 10-27 cm. longae, 5.5—18 cm. latae, majores late ovatae et obo- vatae ad plus minusve rhombicas, distincte longocuspidatae. Spathae 3, affixae ad sextam partem superam (plerumque ad decimam partem super- am) pedunculi. Spadix in statu florifero anguste cylindricus, ad apices florum masculorum mensus 14-36 mm. longus et 8-11 mm. crassus, 4.5—6 mm. crassus, non flores masculi includentes. Flores masculi 3-3.5 mm. longi; receptaculum planum, infra visa orbiculare, vel aliquantum elonga- tum: lobi perianthii breves acuminati ad apiculatos. Stamina 36—-100-- ; bulbis basalibus staminum marginalium in aspectu abaxiali plerumque obovatis ad obovato-depressos, plus minusve obcordatis, desuper visis ra- dialiter ad tangentialiter complanatos; bulbis basalibus staminum cen- tralium brevior et tenuior, sinu distali absenti, desuper visis tangentialiter complanatis ad circulares. Flores feminei in statu florifero 5.5-8. longi, 2.7-3.5 mm. lati; tepala brevissima; stigmata (in floribus tetracar- pellis) ovata, rhombica ad elliptica. Stem at least 60 cm. long, 8.5-18 mm. in diameter, varying from red- dish green to dark red (depending on concentration of small red flecks on a green background); the internodes elongate, ca. 6-9 cm. long. Leaf blades 10-27 cm. long, 5.5-18 cm. broad, the larger ones broadly ovate and obovate to -_ rhombic, distinctly long-cuspidate. Sublaminar portion of leaf 5-11.8 cm. long, typically consisting entirely of leaf sheath, some- times also exhibiting a petiole up to 4 mm. long, this petiole with two 88 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 FIGURES 20, 21. Ne Bierhorstii. 20, holotype prior to removal of the two inflorescences, X 0 Foliage leaves shown with adaxial sides outermost. Upper inflorescence has ae to unfold; lower one with all spathes intact. 21, adaxial view of basal portion of leaf subtending lower inflorescence in FIGURE 1978] WILDER, CYCLANTHACEAE 89 ridges, each ridge comprising one margin of the adaxial surface; if petiole is absent, the membranous margins of the leaf sheath continuous with margins of the lamina and sometimes red-tinged; each margin of the sheath commonly ending distally in a small lobe ca. 1-6 mm. high. Inflorescence fragrant at anthesis. Peduncle during anthesis 34-77 mm. long, light to dark green, bearing one or two peduncle sheaths. The lowermost peduncle sheath (i.e., the prophyll) relatively short and greenish white on the out- side, the next higher one longer and green on the outside. Spathes 3, all inserted on the upper 15 percent (commonly upper 10 percent) of the peduncle, each one partly enveloping the spadix; the lowermost one 59- 82 mm. long, 21-28 mm. broad, lanceolate, cymbiform, terminating in a rudimentary, plicate lamina, + acute, fleshy, on the outside having a laterally extended green band on its basal-most part, the lower half mainly light brown or yellow-brown, the upper half yellow-brown or mainly ex- hibiting green spots on a yellow background, white or brown and brownish white on the inside, the next higher spathe 55-70 mm. long, 23-30 mm. broad, ovate, cymbiform, terminating in a + obvious, rudimentary, plicate lamina, acuminate, somewhat fleshy, on the outside having a laterally ex- tended green band on its basal-most part, the lower part (including the majority of this spathe) mostly light brown or yellow-brown, the upper part light brown, yellow-brown, or mainly yellow, white on the inside; the uppermost spathe 46-59 mm. long, 24-29 mm. broad, ovate, cymbiform, shortly subcuspidate, rather membranous, on the outside having a lateral- ly extended green band on its basal-most part, otherwise light brown, white on the inside. Spadix narrowly cylindrical at anthesis, measured to the ends of the staminate flowers 14-36 mm. long, 8-11 mm. thick, when not including the staminate flowers 4.5—-6 mm. thick. Staminate flowers 3-3.5 mm. long; their pedicels glistening white and of two general types, the first type belonging to staminate flowers inserted lateral to their as- sociated pistillate flowers (relative to the long axis of the inflorescence), each such pedicel being compressed parallel to the lateral margin of the associated pistillate flower, the surfaces of this pedicel which face toward and away from the pistillate flower being, respectively, gently concave and convex, this pedicel terminating basally in a gently curved ridge (FIGURE 29), the second type of pedicel belonging to staminate flowers inserted dis- tally or basally with respect to their associated pistillate flowers, each such pedicel typically less compressed (and then generally parallel to the radius of the associated pistillate flower) or not so, the surfaces of this pedicel which face toward and away from the pistillate flower generally being, respectively, pronouncedly concave and convex at their basal ends, the pedicel generally terminating basally in a + distinct V-shaped ridge which opens toward the pistillate flower (F1curEs 26, 28); receptacle flat, or- bicular as seen from below, 2—2.5 mm. in diameter, or somewhat lengthened, 3-5.5 2-3 mm.; perianth lobes white to transparent, secreting a clear, 20, sheath appearing artificially flattened here; petiole absent (arrow indicates boundary between lamina and sheath), * 0 49, I, inflorescence; V, vegetative branch. 90 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Ficures 22-24, Distal portion of remaining inflorescence from holotype of Ludovia Bierhorstii. 22, 23, portion of inflorescence with all spall intact, 0.78, X 3.7, respectively. Note proximity of all three spathe scars to spadix. 24, portion of spadix from which staminate flowers have been ieseicd away, 1978 | WILDER, CYCLANTHACEAE 91 very viscous, sticky substance, 0.5-0.7 mm. long, 0.3-0.7 mm. broad, short, acuminate to apiculate, reaching or virtually reaching the basal bulbs of the stamens, in side view each basal bulb appearing entirely exposed or mainly so; stamens 36 to over 100 although usually well within these limits, densely crowded, the most centrally situated ones differing sub- stantially in form from those at the margin; anthers yellow, those of mar- ginal stamens 0.5-0.7 mm. long, 0.3-0.6 mm. broad, those of the most centrally situated stamens 0.7-0.8 mm. long, 0.5 mm. broad; filaments dull white, those of marginal stamens tending to be virtually absent or extremely short, typically not reaching the maximum length of 0.2 mm., filaments of the most centrally situated stamens always obvious, ca. 0.1— 0.2 mm. long; basal bulbs dull white, those of marginal stamens in abaxial view tending to appear obovate to depressed obovate, + obcordate, in top view varying from radially to tangentially flattened, basal bulbs of the most centrally situated stamens much shorter and less thick, lacking a distal sinus, in top view appearing tangentially flattened to circular. Pistillate flowers during anthesis 5.5-8.5 mm. long, 2.7-3.5 mm. broad; tepals green basally, white distally, very short, consisting (together with the light green, more peripherally situated tissue of the same flower) of numerous closely packed, glandular(?) protuberances, the apices of these tepals thus variable, e.g., distinctly trilobed or bilobed, emarginate, or truncate, pressed against the stigmas; stigmas light green except for the hairy, white stigmatic crests, commonly 4, sometimes 5 and (by branching of one stigma) 6, somewhat laterally compressed, tending to be distinctly uncinate, seen from above (on tetracarpellate flowers) varying from ovate, or rhombic, to elliptic, projecting between the tepals; staminodes ca. 8-12 cm. long, white. Fruit and seeds unknown. — Ludovia Bierhorstii is named for Dr. David William Bierhorst, who has contributed greatly to our knowledge of anatomy and morphology of vas- cular plants. Ludovia Bierhorstii clearly belongs to the genus Ludovia because it has larger leaves which are entire, unicostate, distichously arranged, distally crenellated, and cuspidate; a peduncle less than 12 cm. long; three spathes per peduncle, these distichously arranged, each lanceolate or ovate; a cylindrical spadix; on staminate flowers large numbers of evenly distrib- uted, glanduliferous perianth lobes; on pistillate flowers extremely small peneach lobes; stigmas commonly uncinate, lacking styles; ovaries ex- tending beneath the level of surrounding tissues, giving the impression of connate pistillate flowers; four parietal placentae per ovary which ap- proach but do not reach its basal-most part. However, Ludovia Bierhorstii exhibits the following features which do not conform to Harling’s (1958) circumscription of Ludovia: the unfolded, larger leaves are somewhat plicate (FicuRE 20), the spathes are rather pane ee ee ee x 4.2. Long axis of each pistillate flower parallels long axis of _ nee. Pi, Sie: flower; S, spathe scars; Sf, staminate ee St stam oe 92 JOURNAL OF THE ARNOLD ARBORETUM VOL. 59 FIGURES 25-30. Staminate flowers from inflorescence illustrated in F1GURES 22-24, X 9.9. 25, 26, staminate oF bilaterally symmetrical, from above (FIGURE 25) and below (FicurRE 26). 28, staminate flower, more or less or- bicular, from above (FIGURE 27) and ice (FIGURE 28). 29, 30, two staminate 1978 | WILDER, CYCLANTHACEAE 93 crowded beneath the base of the spadix, in examples studied being con- fined to the upper 15 percent (commonly, the upper 10 percent) of the peduncle. In contrast, Harling (1958) stated that for both L. lancifolia and L. integrifolia, the lowermost spathe is “. . . placed at about the middle of the peduncle.” Additionally, in L. Bierhorstii each spathe partly encloses the spadix, although Harling (1958) indicated that in both other species of Ludovia the lowermost spathe does not enclose the spadix. In the following account, except where specified, all characterizations of Ludovia lancifolia and L. integrifolia are based on Harling’s (1958) de- scriptions of these species. Apparent differences in size between each of these species and L. Bierhorstii are not indicated where it is suspected that these differences may merely reflect differences in the ways the materials were preserved. Specifically, Harling (1958) apparently made some if not all of his measurements on dried specimens, while the present writer utilized solely materials which were living or preserved in liquid. Accord- ingly, present measurements would appear to be less reflective of shrink- age than were those of Harling (1958). Ludovia Bierhorstu differs from L. lancifolia as follows: (1) Maximum ae Z riage in L. Bierhorstii and L. lancifolia are, respectively, 27 cm. and 110 cm. Even the largest leaf observed which might belong to ia Lorne (an unidentifiable specimen preserved at the type locality) measured only 43 cm. long. However, in L. Bierhorsti and L. lancifolia maximum leaf diameters are comparable (respectively, 18 cm. vs. 16 cm.). In addition, foliage leaves of L. lancifolia differ in shape from those of L. Bierhorstit. (2) During anthesis spadices of L. Bierhorstii and L. lancifolia are, respectively, 4.5-6 mm. and 12-16 mm. thick. (3) On staminate flowers of LZ. Bierhorsti and L. lancifolia, receptacles are, respectively, flat vs. distinctly and broadly concave.! 4) In L. Bierhorstit perianth lobes on staminate flowers are 0.5—0.7 mm. long and in side view fail substantially to cover basal bulbs of the marginal stamens (FicurREs 35-37). However, in L. lancifolia, the lobes are larger (0.8-1 mm. long), in side view substantially covering basal bulbs of the marginal stamens (based on Harling’s fig. 89a as well as on present observation of L. lancifolia). parent contradictions exist in literature regarding Ludovia lancifolia. Sat hee described the staminate flower of L. lancifolia as having a “. . . receptacle distinctly and broadly concave a feature verified during my own ae Brongniart (1861, fig. 4), however, illuetrated a longitudinally sectioned flower, pur- portedly of L£. lancifolia, which exhibited an apparently flat receptacle. In RAdiCon; Brongniart emphasized that oes flowers of ZL. lancifolia lack sepals (‘calyx nul- 3 does not depict the obvious broad peripheral region of protuberances which occurs on pistillate flowers of ZL. lancifolia. flowers, from below. Flower in FicureE 29 approximately orbicular; that in FIGURE 30, irregular. P, tepals; Pe, ridge at basal end of pedicel; Sta, stamens; each arrow in FIGURES 26, 28 indicates one arm of V-shaped ridge. 94 JOURNAL OF THE ARNOLD ARBORETUM VOL. 59 Ficures 31-38. One whole staminate flower and portions of other staminate flowers from inflorescence illustrated in FIGURES 22-24. 31, 32, largest flower of Ludovia Bierhorsti, from above (F1cGuRE 31) and be slow (FIGURE 32), both This flower, from basal end of spadix, bears over 100 stamens; its long axis 1978 | WILDER, CYCLANTHACEAE 95 (5) On staminate flowers of L. Bierhorstii, anthers from both margin- ally and centrally situated stamens tend to be shorter than their counter- parts in L. lancifolia, respectively, on marginal stamens 0.5—-0.7 mm. vs. 0.6-0.9 mm., and on centrally situated stamens 0.7—-0.8 mm. vs. 1.4-1.7 mm. (6) In L. Bierhorstii filaments vary from virtually absent to 0.2 mm. long; in L. lancifolia filaments are 0.1-0. (7) In L. Bierhorstii basal bulbs of marginal stamens commonly ap- pear obcordate in side view (Ficures 35-37), whereas in L. lancifolia this is not so (based on Harling’s fig. 89d, f, g, and on present study of L. lan- cifolia). Ludovia Bierhorstii differs from L. integrifolia as follows: (1) Maximum lamina width in L. Bierhorstii greatly exceeds that in L. integrifolia (respectively, 18 cm. vs. 9 cm.), although ranges of lamina length in the two species are roughly comparable. In addition, laminae of the two species differ substantially in shape, those of L. integrifolia being elliptical to oblanceolate (however, see later discussion regarding leaf shape (2) Maximum numbers of stamens per staminate flower in L. Bier- horstit and L. integrifolia are, respectively, over 100 vs. 3) In L. integrifolia filaments of all stamens are ca. 0.2 mm. long, in contrast to the condition in L. Bierhorstit. (4) Contrary to the condition in 1. Bierhorstii, in L. integrifolia basal bulbs on marginal stamens are not obcordate in side view (based on Har- ling’s fig. 91a—c, e, f, and on present study of L. integrifolia). (5) In L. integrifolia stigmas are distinctly laterally compressed, where- as stigmas of L. Bierhorstii tend to appear somewhat less compressed and, in addition, substantially more connate to one another (based on Har- ling’s fig. 91g, together with present study of both species). (6) In L. Bierhorstii protuberances are abundant and pronounced on that portion of a pistillate flower peripheral to its tepals (FIGURES 39-43, Pr); this is not so in L. integrifolia (based on Harling’s fig. 91g as well as present study of L. integrifolia). However, comparable pronounced pro- tuberances have been observed by the author in ZL. lancifolia and, indeed, their presence in L. lancifolia is suggested by Harling’s fig. 89h. Also, in ee curved. ue ee in ee arcuate ridge. 33, half of a flow- : ; , exposing basal b situated basal bulbs appearing much smaller than peripheral ones (dark-stain- i everanc ) anther). 35-37, abaxial views of stamens situated Sarin ai Via me 15, respectively. Basal bulb of an apparently single stamen illustrated at left in Ficure 37 bears two anthers. 38, centrally situated stamens from flower shown in Ficure 33, illustrated in face view, X 22. A aida stamen note relatively long filament and short basal bulb. A, anther ; Fr. filament; Bb, basal bulb; P, tepal; Pe, ridge at basal end of pedi cel. 96 JOURNAL OF THE ARNOLD ARBORETUM ‘VOL. 59 FIGURES 39-44, Pistillate flowers and portions thereof from region of spadix illustrated 1 in FIGURE 24. 39, 40, Have Gots flower, from above (FIGURE 39) and from side (FicurE 40), X 7.2, X 13, respectively. 41, pistillate flower with eee stigmatic crests, from ae x 7.2. Note crest at upper left, branched 1978] WILDER, CYCLANTHACEAE 97 L. Bierhorstii protuberances at least partly comprise each tepal (FIGURES 42, 43). Marginal protuberances of each pistillate flower appear to comprise a discrete group and eats at least positional equivalents of tepals on staminate flowers (e.s mpare FIGURE 42 with FicurEs 26, 28-30, and 3 he marginal eeieeances may appear either more or less connate or distinct (e.g., note a few distinct protuberances on the small portion of flower illustrated at bottom of FicurE 42). As illustrated in FIGURES 42 and 43, protuberances comprising tepals also appear connate to vary- ing extents. All protuberances might represent modified stamens. If such homology could be demonstrated, our concept of the nature and phylogeny of cyclan- thaceous flowers (particularly of their tepals) would be profoundly af- fected, especially because among cyclanths Ludovia appears primitive in various respects (Harling, 1958). Indeed, Garcin (1958) concluded, based on morphological and anatomical studies of various Carludovicoideae (not including Ludovia), that each tepal on pistillate flowers of these taxa rep- resents a group of completely sterilized, more or less concrescent stami- nodes. In light of these possibilities, study of floral development in Lu- dovia, particularly including either L. Bierhorstii or L. lancifolia, now ap- pears Imperative The present description of at least vegetative aspects of Ludovia Bier- horstii might not accurately reflect morphology of this species in nature, for the following reasons. First, specimens were described from ealliva tion and they could have developed differently than in nature. Second, because these specimens were genetically homogeneous (a clone), they could not have reflected any genetically based variation which L. Bier- horstii as a whole might exhibit. Finally, the material may or may not have possessed an atypical genotype (and hence phenotype) relative to those of most naturally occurring plants of L. Bierhorstii. Such uncer- tainties, together with lack of fertile specimens of L. Bierhorstii from na- ture, preclude definitive conclusions as to whether the described material is conspecific with any or all other previously collected entire-leaved cy- clanths from Altaquer and vicinity. Indeed, entire-leaved cyclanths from this region collectively exhibit leaves of diverse morphology and, as dis- cussed below, Harling (1958) included one sterile, entire-leaved specimen from Altaquer in a species of another genus, Pseudoludovia Andreana, Nevertheless, it appears significant that presently described specimens (FicurEs 20, 21), together with all plants I preserved at the collecting locality (including specimens bearing leaves appearing essentially identi- cal to those illustrated by Harling (1958, plate 86) under the name Pseu- doludovia Andreana (FicurES 45—48)), have the following features in toward periphery of flower. 42, 43, pistillate flowers, from side, both * 18. 44, pistillate flower, from above, 14. Perianth member bs bottom of figure oa a three-lobed (each arrow indicating one lobe). epal; Pr, protuber- an on portion of pistillate flower es rerpheial - et Sc, stig- aes crest; St, staminode or staminodial sca enol OF THE ARNOLD ARBORETUM FIGURES 45-48. Material of entire-leaved Cyclanthaceae of uncertain affinity preserved in fluid at Altaquer, Colombia, near type locality of Ludovia Bier- horstu. Each arrow not accompanied by symbol indicates oe between sheath and petiole of leaf. 45, portion of leaf illustrated in FrGuRE 47, at higher magnification and in adaxial view, * 1.9. Note vinbed ae mis 1978 | WILDER, CYCLANTHACEAE 99 common. All plants exhibit elongate internodes and distichous phyllotaxy (FicurEs 20, 48). All specimens bear small, varyingly concentrated, red flecks both on stems and sublaminar portions of leaves. The leaves col- lectively exhibit a virtual continuum of laminar shapes including elliptical, ovate, oblanceolate, obovate, and more or less rhombic. On virtually every leaf, in the sequence listed, lamina, sheath, and petiole (where present) oc- cur in decreasing order of size (FIGURE 49). On the petiole, two adaxial ridges continuous with margins of the lamina occur either along its entire length (where short; FicurE 45) or on its distal end (where long) ; such ridges also occur on the specimen in Harling’s (1958) plate 86. Large leaves each apparently form a cuspidate tip, although this tip is commonly torn away (Ficures 20, 47, 48) However, described specimens differ from plants preserved in liquid at the type locality (hereafter called non-described specimens). First, on de- scribed plants petioles are typically absent (Ficures 20, 21, 49) but vary up to 4 mm. long; on non-described specimens petioles are typically but not always present, varying up to 75 mm. long (FIGURES 45-49). How- ever, in being typically non-petiolate (and perhaps in other ways), leaves from described plants may merely exhibit a persistent juvenile feature. Second, on described plants leaves 140 mm. or more long frequently ex- hibit broadly ovate laminae (F1iGuRE 20), whereas among non-described specimens leaves in this size range are less commonly ovate. Large, more or less oblanceolate and obovate leaves are common among non-described specimens (FicurEs 47, 48), yet are absent in described material. Third, laminae from described specimens tend to exhibit a higher ratio of width to length (in the investigated size range) than do those from non-described specimens (FIGURE 50). The discovery that Ludovia Bierhorstii represents a new species has cast doubt on the distinctiveness, if not the existence, of Pseudoludovia An- dreana. This is partly because of the similarity between leaves of L. Bier- horstii and those of P. Andreana as described by Harling (1958). It would be desirable to collect material of entire-leaved Cyclanthaceae at the type locality of P. Andreana in Quindio, Colombia, in order to resolve this prob- lem. If future studies indicate that P. Andreana was described on the basis of a mixed collection consisting of sterile material of Ludovia Bierhorstiu and fertile material of another cyclanth (perhaps a species of Sphaera- denia), the epithet “Andreana” would more appropriately be assigned to the fertile material than to L. Bierhorstit. I ae ge eg ee a a wing continuous with margin of lamina and sheath. 46, portion of shoot tip, s 0.6. Folded, expanding adult leaf at center of illustration enclosed by two unfolded, clearly petiolate, adult leaves. 47, adaxial view of relatively large, obovate, petiolate leaf resembling those illustrated in Harling’s (1958) plate 86 as Pseudoludovia Andreana, 0.3. 48, shoot bearing obovate, petiolate leaves comparable to those illustrated by Harling (1958, plate 86), X 0.16. Ro, roots. 100 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 280 : ete) ra) e e 240 © @ LAMINA LENGTH . O m SHEATH LENGTH 4 ee e - A PETIOLE LENGTH o® e oe 200 Oo ®@ ae £08 | ie) _ 0° Cot e = 160+ ® = & - @ eS * wa 5 fe) ° 2 120- on 0 = oe "s oo 28 = oe ] = Os ' é r] a e Og of gs Or ee ae, et oe 7 A 2 ay A E i) 1, On a2 _ A oh geo a z a A i A A 40 aa , y A i Q A the A is A aA A a A aA OL ah Apll a RADA AAA MA AN 120 160 200 240 280 320 360 400 440 49 LEAF LENGTH (MM) 10 times as long as cell width v (very oblique end wall >5 to = 10 times as long as cell width o (oblique) end wall >2 to =5 times as long as cell width s (slightly oblique) end wall >1 to =2 times as long as cell width t (transverse) end wall perpendicular to lateral wall 108 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 59 . oa ome - : +e ett ete: x 82. eral wide vessels in metaxylem. 4, root of Trithrinax brasiliensis, « 75. Several rows of wide vessels. Medullary vessels present (mv). 5, root of Mauritiella pacifica, X 82. One row of wide vessels. (c, cortex; e, endodermis; f, fibers; g, n, narrow tracheary elements; nm, narrow ; 0, obliterated tracheary elements of proto- xylem; p, tracheary elements of protoxylem; ph, phloem; v, parenchyma of vas- cular bundle; w, wide vessel of metaxylem. ) 1978] KLOTZ, PERFORATION PLATES 109 The distinction between transverse and slightly oblique is somewhat subjective but is adhered to rather strictly. In many samples, especially those from roots, end walls range from apparently perpendicular to the length of the cell to only a few (=15) degrees from the perpendicular —ie., they are very slightly oblique. The overall or net slope of curved end walls was estimated. Ligules (FIGURES 36-38) were not included in determining slope. Two methods were devised to summarize the data. First, proportional frequencies were determined for the five categories of slope for each organ in each major group. Each species was given equal weight, even if more vessel elements had been obtained from some specimens than from others. The proporonal frequencies are presented as histograms (FIGURE 41). Second, ‘‘slope indexes’? (TABLE 1) were calculated for each organ ac- cording to the formula: slope index & 100 = 0 (% tracheids) + 1 (% extremely oblique end walls) + 2 (% very oblique end walls) + 3 (% oblique end walls) + 4 (% slightly oblique end walls) + 5 (% transverse end walls) Average slope indexes were calculated for the organs in each of the major groups, and “total slope indexes” were calculated by adding the average slope indexes for the three organs. In this way, a single numerical value was obtained to summarize the average slope of the end walls in each of the major groups. In TaBte 1, the total slope indexes of the major groups are listed in order of increasing value (or decreasing obliqueness). In the iriarteoid major group, the average slope index for the aerial roots is lower than that for the subterranean roots. Since the two kinds of roots are not from the same species, total slope indexes for the iriarteoid palms are calculated both ways (TABLE 1) “SPECIALIZATION VALUES.” “Specialization values” are assigned to the samples of perforation plates of the wide vessel elements of metaxylem according to the following system, which includes categories for describing the relative spacing of bars of secondary wall in multiple perforation plates. This system differs from the one invented by Cheadle for other groups of monocotyledons (e.g., Cheadle & Kosakai, 1975) in order to express the data for palms in more detail. O — tracheids only — definite vessels not observed 1 — multiple perforation plates with “pitlike’ perforations that span <1 vessel—parenchyma pit on the adjacent lateral wall (Fuic- URES 6, 7 2— multiple perforation plates with ‘narrow’ perforations that span >1 to =2 vessel—parenchyma pits (FIGURE 8) 3 — multiple perforation plates with “medium-size” perforations that span >2 to =5 vessel-parenchyma pits (Ficures 9, 11, 23) 4 — multiple perforation plates with “wide” perforations that Sor >5 vessel_parenchyma pits (FIGURE 15), or with = 3 bar 110 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 MANNA NAL bat WU eee "v7": TY,TTY T°. TS eer geeee ee Cee oe ee we Pda ye bore ton tna tl fs boi cman: I'IGURES 6-13. Isolated wide tracheary elements of metaxylem. 6, petiole of Pinanga sp. (H. E. Moore 9112), * 240. Arrow indicates end wall with aper- osely spaced than oe of adjacent lateral wall. 7, stem of W endlandiella ae > 240. Arrow indicates end wall with apertures about as closely spaced as pits of adjacent lateral wall. 8, root of Physokentia rosea, 240. End wall with each perforation oe an two pits of adjacent ores wall. Arrow indicates narrower aperture ssibly intervessel pits, in distal part of end wall. 9, stem of ene oe x 140. End wall in which perforations, separated by wide bars, each span about five pits of adjacent 1978 | KLOTZ, PERFORATION PLATES Eh TABLE 1, Total average slope indexes for the major groups. Mayor GROUP _* SLOPE INDEX x. ~ Podococcoid Ve IX. Chamaedoreoid: smaller species 7.4 1S Chamaedoreoid: total 8.1 XV. Phytelephantoid * 8.2 V. Nypoid 8.6 LV Geonomoid 8.9 VIII. Ceroxyloid 9.1 X. Iriarteoid: including aerial roots 9.1 XII. Arecoid 9.3 Pe Iriarteoid: including subterranean roots 9.4 IX. Chamaedoreoid: larger species 10.0 II. Phoenicoid 10.1 IV. Lepidocaryoid: non-lianas 10.9 WALT, Pseudophoenicoid Wi2 ie Coryphoid a ee SELLE, Cocosoid 11.2 III. Borassoid 11.4 VI. Caryotoid Wes Vi Lepidocaryoid: total 120 IV. Lepidocaryoid: lianas 12.6 * Including Ammandra, with only tracheids in ieee stem. 5 —-mixture of multiple perforation plates (as in #4) and simple per- foration plates 6 — simple perforation plates only (FicurE 13) Samples that span two adjacent categories are assigned values of “1%”; for example, “31%” expresses samples of multiple perforation plates with ars of “medium” to “wide” spacing, and “514” expresses samples in which the end walls bear one or no bars. Mixed samples that span a wider range of categories are designated with “a”: 24a — ee plates in sample range from 1 to 4 3a — from 2 to 4 34a — from 2 to 5 (or 6) 4a — from 3 to 5 (or 6) lateral wall. 10, stem of sh ca humilis, * 100. Vessel elements with slightly crooked or scalloped (arrow) lateral walls. Perforation plates have ee bars. 11, stem of EHuterpe ee sx 750. Detail of vessel element with each perforation (p) spanning about five pits of peor an wall. 12, stem of Rhapidophyllum oe s< 120. Vessel element with one bar in each of three perforation plates. 13, stem of Calamus mueller pa 100. Vessel ele- ment with simple Sia plates. Arrow indicates broad flange of secondary wall around perfora 112 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 A similar procedure was applied to Cheadle’s system of specialization values (Cheadle, 1955), which considers only the relative abundance of multiple perforation plates versus that of simple perforation plates: O — tracheids only 1 — multiple perforation plates only 2 — =1/3 simple perforation plates 3 — >1/3 simple to =2/3 simple 4— >2/3 simple 5 — simple perforation plates only mixed multiple and simple per- foration plates The values refer only to the wide tracheary elements of the metaxylem. For values 2 through 4, Cheadle does not specify exact percentages but lets 2 stand for “mostly scalariform”; 3, for “about equal scalariform and simple”; and 4, for “mostly simple” (Cheadle & Kosakai, 1975). Within the taxonomic groups of palms, the means of the specialization values were calculated for each organ and were plotted as points (FIGURES 42,43). A vertical line through a point indicates the range of specialization TABLE 2. Total average specialization values for the major groups. AUTHOR'S SYSTEM CHEADLE’S SYSTEM Chamaedoreoid: small species 5.1 Nypoid 3.0 Chamaedoreoid: small species 3.0 Iriarteoid: aerial roots 6.2 eo total 6.3 Chamaedoreoid: total 3.7 Podococcoid 6.5 Iriarteoid: aerial roots 3.8 Iriarteoid. subterranean roots 7.1 Jriarteoid: subterranean roots 5.0 Nypoid 8.0 nemaedersold: large species 5.5 Ceroxyloid 8.3 Arecoid 58 Geonomoid 8.6 Pseudophoenicoid 6.0 Arecoid 9.0 Podococcoid 6.0 Phytelephantoid * 9.0 Geonomoid 6.0 Ceroxyloid 6.5 Chamaedoreoid: large species 9.7 Phytelephantoid * G7 Pseudophoenicoid 10.5 Caryotoid 7.6 Cocosoid 7.8 Coryphoid 11.9 Phoenicoid 8.0 Caryotoid 11.9 Lepidocaryoid: non-lianas 8.6 Phoenicoid 12,0 Coryphoid 8.8 Cocosoid 12.1 Borassoid 10.0 Lepidocaryoid: non-lianas 13:2 Borassoid 13.4 Lepidocaryoid: total 10.8 Lepidocaryoid: total 145 Lepidocaryoid: lianas 11.9 Lepidocaryoid: lianas 15.0 * Including Ammandra, with only tracheids in the stem. 1978] KLOTZ, PERFORATION PLATES 113 values for the organ in the group. The three points on each chart (for petiole, stem, and root) are connected by lines merely to make the pat- terns easier to perceive. The specialization values were further summarized by adding the aver- aged values for petiole, stem, and root in the major groups of palms (Ta- BLE 2). In the iriarteoid major group, the average value for the aerial roots is lower than that for the subterranean roots. Since the two kinds of roots are not from the same species, the total average specialization values for the iriarteoid palms are calculated both ways. OBSERVATIONS KINDS OF PERFORATION PLATES. The wide tracheary elements of the metaxylem are vessel elements, except possibly in the rhizomatous stem of the phytelephantoid palm Ammandra decasperma. Multiple and/or simple perforation plates occur in the wide vessels.? Most of the multiple perforation plates are basically scalariform. The most tracheid-like va- riety contains many bars on steeply sloping end walls in which the per- forations are less widely spaced than the adjacent vessel-parenchyma pits of the lateral walls, but more widely spaced than the intervessel pits (FicurE 6). This primitive variety grades into perforations that are about as closely spaced as the adjacent vessel-parenchyma pits of the lateral walls (FicurE 7). In successive forms there is a gradation toward fewer, more widely spaced bars (Ficures 8-11, 14-18); to only one bar (Fic- URE 12); and finally, in the advanced extreme of the series, to no bars (Ficure 13). Paralleling this gradation is a general decrease in the slope of the end wall. Where perforations are narrow, the bars of the perforation plates are usually spaced evenly; but with wider spacing of bars, the bars tend in- creasingly toward irregular spacing (Ficures 14-18). Bars of perfora- tion plates frequently branch and anastomose to varying degrees (FIGURES 23, 28, 29) and rarely end blindly (Ficures 28, 31). The gyres of secondary wall that compose the bars of a perforation plate sometimes appear incompletely “fused” to varying degrees (F1GuRE 30). The apertures between such “unfused” gyres are often no wider than the pits of the lateral walls, but sometimes they approach the width of the major perforations. Reticulate perforation plates are the other basic type of multiple per- foration plate (Ficures 19-22, 24). An extreme variation of this type consists of only one bar, which is oriented longitudinally rather than trans- versely in the perforation plate. In another distinctive variation, a large, central perforation is surrounded by smaller apertures that are larger than the pits of the lateral walls (FrcurE 24). Reticulate perforation plates in- tergrade with scalariform perforation plates through a great diversity of forms. These intermediate, ‘“scalariform-reticulate” configurations range ? The terminology for perforation plates is from Esau (1965). 114 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 iitay; mig coe NIE Bree rire "4 a aa Op hee 4 e~ bit ai wayyy ae | : me, ee i i « l'icurEs 14-24. Vessel elements of metaxylem. Figure 15 is narrow vessel ele- ment; the others are wide vessel elements. 14-18, scalariform perforation plates: 14, stem of Pseudophoenix sargentii, X 82, few bars, unevenly spaced; 15, root of 0, few i ate); 21, stem of Lepidocaryum gracile, < 66; 22, ages of Daemonorops andis, X 300. 23, scalariform-reticulate perforation plates: stem of Clinostigma savaiiense, > 75, branching and anastomosing of bars rather aoe (upper ves- 1978 | KLOTZ, PERFORATION PLATES T5 from highly regular, lattice-like arrangements of many bars spaced rather closely (FiGuREs 25, 26) to arrangements in which the spacing of the bars is more sparse and/or irregular (FrcurEs 23, 27, 29, Glee Perforation plates also vary in other ways. Bars of the end wall may be thicker (Ficure 9), similar in thickness (FIGURE 7), or thinner (Fic- uRE 8) than the gyres of the lateral wall. The widths of the bars may be uniform (Ficure 8) or variable (Ficure 15) throughout the length of a perforation plate. A flange or rim of secondary wall frames many simple perforation plates (FIGURE 13) and some multiple perforation plates, especially those with few or widely spaced bars (FIGURES 20231)... Per foration plates may extend the entire length (F1curE 26) or only part of the length of the end wall. In the latter category, there is a gradual or abrupt transition from perforations to intervessel pits at the proximal end of the perforation plate (FicurEs 33, 34), at the distal end (FIGURES 8, 9), or at both ends. This character is often highly variable within a sam- ple. End walls also occur in which the perforations do not extend to the lateral margins of the end wall (FicurE 35). Vessel elements with inter- vessel pitting on the end wall distal to the perforation plate intergrade with vessel elements in which an extension of the cell continues distally beyond the end wall but is not part of it. Such extensions, called ligules, vary in shape and relative size (FIGURES 36-38). End walls are usually quite straight, but occasionally they are curved or even “‘saddle-shaped”’ (Ficure 39). Perforation plates rarely occur on the lateral walls. END WALL SLOPE. The histograms of proportional frequencies (FIc- URE 41) indicate that slopes of end walls are generally greatest in the petiole, intermediate in the stem, and lowest in the root. The only signifi- cant exception to this trend is the phytelephantoid 4 mmandra decasperma, in which clearly discernible vessels were not found in the stem. Some of the major groups are rather heterogeneous in slopes of end walls (TABLE 1, FicureE 41). End walls tend to be less oblique in the larger chamaedoreoid species than in the smaller chamaedoreoid species. End walls also tend to be less oblique in the lepidocaryoid lianas than in the other lepidocaryoid palms. “SPECIALIZATION VALUES.” The diagrams of average specialization values of perforation plates in the major groups of palms indicate that bars in the perforation plates of the wide tracheary elements of the metaxylem are generally most closely spaced in the petiole, intermediate in the stem, and most widely spaced (or fewest) in the root (FIGURE 42). The only significant exception to this tendency is Ammandra decasperma, in which clearly discernible vessels were not observed in the stem. The diagrams according to Cheadle’s system (F1GURE 43) show that multiple perfora- tion plates are most frequent in petioles, intermediate in stems, and least frequent in roots. sel element) or irregular (lower vessel element). 24, reticulate perforation plate: root of Reinhardtia simplex, * 340, single large perforation (arrow) sur- rounded by smaller perforations. 116 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Ficures 25-31, Wide vessel elements of metaxylem. 25-27, scalariform- reticulate perforation plates: 25, petiole of Polyandrococos caudescens, X 240 bars connect numerous crossbars, both sets of bars very evenly be rpos right angles to form screenlike configuration; 27, petiole of Arenga undulatifolia, X 180. 28, scalariform perforation plate: stem of Clinostigma savatiense, < 88, bars much branched, especially toward sides of perforation plate; some narrow bars end blindly (arrows). 29, scalariform- reticulate perforation plate: rachis of Eremospatha sp., X 100. 30, scalariform perforation plate: stem of Mauritia vinifera, < 65, gyres of secondary wall that r in : siphonospathus, 140, some narrower bars ending blindly (arrows) and peripheral flange of secondary wall. 1978] KLOTZ, PERFORATION PLATES 117 The positions of the major groups in TABLE 2 are generally similar in the two schemes. The major groups are clustered along the continuum of numerical values. At the lower end are the chamaedoreoid (especially the small species) and iriarteoid palms. Above them is a cluster that includes the ceroxyloid, arecoid, geonomoid, and phytelephantoid palms. The larger chamaedoreoid species and Pseudophoenix are also in or near this cluster. Nypa and Podococcus belong to one of the two lower clusters of groups in both schemes. The next cluster in the series includes the cory- phoid, phoenicoid, caryotoid, and cocosoid palms. The nonscandent lepi- docaryoid palms also belong in or near this cluster in both schemes. Fi- nally, at the higher end of the spectrum are the borassoid palms and the lepidocaryoid lianas. The spacing of bars in perforation plates (F1curE 42) and the distribu- tion of multiple and simple perforation plates (FicurRE 43) show trends that parallel those shown by the slopes of end walls (Ficure 40). aki oreo seer | a s[tfe[vfots{[tfefvyo][s]t LCORYPHOID MAJOR GROUP (36- pie V.NYPOID MAJOR GROUP (I-I-!) a 53 16 . e 3 50 47], 60 40 30] 10 90] 10] I¢ a dao 4 4 J jer 4 LA. TRITHRINAX ALLIANCE (18-15-15) Ol ncARYOTOID MAJOR GROUP (6-5-4) walt--2 24 - 17 79 3 5 60 35 5 50 43 2 2 42 54 2 60 pode pay Qa I.B.LIVISTONA ALLIANCE (14-9-9) = 5 56 9 - 8 19 il i inl | / CEROXYLOID MAJOR GROUP (3-4-2) 0 67 3 r ae “SP ECIES (5-2-4 WITH LIANAS ix “CHA AMAED JOR. EOI! ) MG SMA ALL SPECIES (9-9. 9-9) (20-24-8) BQ 49 i PHYLOGENETIC IMPLICATIONS. The most primitive xylem in palms is not found in the coryphoid major group, which contains the greatest num- chamaedoreoid major groups ( a 1973), the small chamaedoreoid palms have the most primitive xylem, yet include the genera most advanced 1978 | KLOTZ, PERFORATION PLATES 121 | EM | PET | ROOT | a Piette me [s[ aa ees EEE COEE EET e[v[eo]s]t] PODOCOCCOID MAJOR GROUP (I-I-1) (AERIAL YTELEPHANTOID MAJOR GROUP (3-3*-2) go [loo 60 40 6 70. 104 10 1 ' 45 55 20F x fe) ~|XILARECOID MAJOR GROUP (45-45-4 28 32 8 (Z ol 2b <= ao © oOo 9° Postel slealesl Pte Jour ae iasloal- glottal ahd at 1 n ny 2 oO T £ XH. COCOSOID MAJOR GROUP (37-!9- 3 24 24 5 ap) a ad fas ct a Gal cab a Ficure 41. End wall slopes of wide tracheary elements of metaxylem: hi sto- e in each group for petiole, stem, and root in parentheses. * Including Ammandra = 0). in other characters. Pseudophoenix, the most primitive genus in this evo- lutionary line, has xylem that is much more advanced than that in the small chamaedoreoid palms. On the other hand, within the coryphoid major group, the most primitive xylem occurs in the Trithrinax alliance, which does contain the greatest number of other primitive features. Nypa fruticans, the sole species of nypoid palm, possesses a mixture of primitive and advanced features, in- cluding primitive xylem. The iriarteoid, podococcoid, and arecoid palms generally have more primitive xylem and are more primitive in other ways than are the cocosoid palms. There is little resemblance between Moore’s (1973) systematic sequence of the evolutionary lines and major groups and the sequence in the pres- ent study based on the average level of evolution of the xylem. This result is not surprising, for different selective pressures would be ex- erted on xylem than are exerted, for example, on the reproductive parts of the plant such as the inflorescences, flowers, and fruits. Evolution in the xylem appears to have progressed independently in the various tax- onomic groups of palms. 122 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 ECOLOGICAL IMPLICATIONS REGARDING HABITAT. Carlquist (1975) hy- pothesized that perforation plates with few or no bars generally represent a more favorable adaptation to environmental conditions of greater aver- age moisture demand than do perforation plates with more numerous bars that are more closely spaced. To some extent, the palms tend to display this correlation between form of perforation plates and degree of moisture demand in the environment, at least as a general trend among the major groups. For example, the coryphoid, phoenicoid, borassoid, pseudophoeni- coid, and cocosoid palms include many species adapted to habitats that are very exposed (e.g., savannas) or that have low or strongly seasonal precipitation (Moore, 1973). These groups tend to have fewer, more widely spaced bars in the perforation plates than do some of the groups that are mainly adapted to more mesic habitats such as rain forest, cloud forest, or forest understory — e.g., ceroxyloid, chamaedoreoid, iriarteoid, po- dococcoid, arecoid, geonomoid, and phytelephantoid palms. Nypa fruticans, which inhabits estuaries in the humid tropics of the Old World, also fits with the latter groups regarding form of perforation plates (Carlquist, 1975). Most lepidocaryoid and caryotoid palms are adapted to tropical re- gions of high rainfall (Moore, 1973), but the moderately to very advanced perforation plates that characterize these species may be related to the fact that many of them inhabit forests at low elevations and must therefore tolerate the greater moisture demand that would accompany a warmer climate. However, the discrepancies that various species and higher groups of palms present with regard to these trends indicate that the relation- ship between form of perforation plates and adaptation to the moisture re- gime of the environment is not simple. For example, the lianas of the lepidocaryoid and cocosoid major groups are adapted to similar kinds of habitats yet differ markedly in the form of their perforation plates, especial- ly in the stem. Among the arborescent species of wet lowland tropical forests, certain coryphoid palms (e.g., Cryosophila albida, Itaya amicorum, Licuala muelleri) have more advanced perforation plates (especially in stem and root) than do iriarteoid palms of comparable or greater size. HABIT CATEGORIES. Few consistent relationships occur between the habits of palms and the forms of their perforation plates.4 Among the species of lianas, the sequence of kinds of perforation plates in the organs of the plant corresponds to the modal or average condition of this charac- ter within the major group to which the species belongs. For example, only multiple perforation plates occur in the wide vessel elements of the scandent species Chamaedorea elatior, as in the nonscandent small chamaedoreoid palms. Species of the cocosoid genus Desmoncus are lianas with only mul- tiple perforation plates in the wide vessel elements of the shoot and simple perforation plates in the wide vessel elements of the root. The sequence of kinds of perforation plates in Desmoncus falls within the lower part of the range of this character exhibited by the cocosoid major group. The “In contrast, there are relationships between the habits of palms and the dimensions of their vessel elements, especially the diameters of the wide vessels in the stem (Klotz, 1977, 1978 1978 | KLOTZ, PERFORATION PLATES igs lepidocaryoid lianas are very advanced in the form of their perforation plates (Ficures 41-43). They fall within the upper part of the range of this character in the lepidocaryoid major group. The same tendency exhibited by the lianas is also shown by the rhi- zomatous palms, which occur in several of the major groups. Perforation plates of the wide vessel elements of the metaxylem are all multiple in the rhizomatous stems of Sclerosperma mannii (arecoid), Calyptrogyne brachystachys (geonomoid), and Phytelephas microcarpa (phytelephan- toid). They are mixed multiple and simple in the coryphoid palms Rhapi- dophyllum hystrix, Licuala sp., and Serenoa repens; and in the cocosoid palms Allagoptera arenaria and Elaeis oleifera. Finally, they are mostly simple in the lepidocaryoid Eleiodoxa conferta. The remaining category includes all palms with erect, nonscandent stems. It is thus a very diverse group of arborescent, shrubby, and dwarf species. In this group, the small, slender species of the forest understory usually have the most primitive perforation plates within their respective major groups. This tendency is exhibited among the coryphoid palms (RfAapis excelsa), the lepidocaryoid palms (Lepidocaryum spp.), the chamae- doreoid palms (Synechanthus spp., Chamaedorea spp., Wendlandiella polyclada), the arecoid palms (Reinhardtia simplex, Pinanga sp., /guanura sp., and others), the cocosoid palms (Aiphanes sp., Bactris simplicifrons), and the geonomoid palms (Pholidostachys pulchra, Geonoma spp.). Con- versely, the most advanced perforation plates in some of the major groups occur among the larger species of those major groups — for example, Piga- fetta filaris (lepidocaryoid); Hyophorbe indica (chamaedoreoid); Mani- caria saccifera, Roystonea spp. (arecoid); and Welfia georgit (geono- moid). DATA OF OTHER AUTHORS. Cheadle (1942) found only scalariform per- foration plates in the metaxylem of leaves (22 species of palms) and in- florescence axes (6 species). In aerial stems, he found 15 species with scalariform perforation plates only and 8 species with a mixture of scalari- form and simple perforation plates. In roots, all 21 species that he ex- amined had a mixture of scalariform and simple perforation plates in the metaxylem, although the scalariform perforation plates were generally re- stricted to the early-matured metaxylem. The present study reports pet- ioles that contain simple perforation plates and roots that lack simple per- foration plates. Thus Cheadle’s (1942) findings are similar to those of the present study except that he found fewer kinds of perforation plates within leaves and roots of palms. Mahabalé (1959) observed the range of numbers of perforation plate bars in stems of 14 species of palms, including 8 species of Phoenix. Most of his data fit within the limits of the variation observed for the major groups in the present study. Tomlinson’s (1965) brief summary of his unpublished findings on perforation plates in palm stems presents a sequence for the taxonomic groups according to their average level of evolution in this character. His |. CORYPHOID X.IRIARTEOID MA-|XI. PODOCOCCOID | XII. ARECOID MA- LA TRITHRINAX V.NYPOID MAJOR 1V.A,B,G. LEPID IV.C-F,H.OT MAJOR GROUP ALLIANCE CARYOID ALLI - Se eiboenavoi0 (1-1-1) JOR GROUP ges GROUP JOR aad (36-27-27) (18-15 -15 ) ANCES WITH ALLIA wa (9-8-5-2) (I-1-1) ( 41) 1 -_—oO i SUBTERRANEAN seriall ah yn ow ff a i . LIANAS 20-24- 0 ( 4-8) | B.LIVISTONA 1.C. CORYPHA 1.0. SABAL ALLI- Vi. CARYOTOID VIL. PSEUDOPHOE - | VIIl.CEROXYLOID XILB ELAEIS AL- ALLIANCE ALLIANCE ANCE (2-1-1) MAJOR GROUP NICOID MAJOR OR GROUP LIANCE (3-1-1) (2-2-2 (6-5-4 GROUP (2-1-1) (3-4-2) (14-9-9) ene ees 1 ! 1 1 1 1 XULC eacTRs XIV. GEONOMOID XV.PHYTELEPHAN- ll. PHOENICOID lll. BORASSOID IV. LEPIDOCARYOID IX. CHAMAEDORE- |IX.LARGER CHA- |IX.SMALLER CHA- arias GROUP MAJOR GROUP |} MAJOR GROUP OID MAJOR | MAEDOREOID MAEDOREOID rea MAJOR GROUP TOID MAJOR -1) (6-3-2 (33-32-14) GROUP (/4-I1-13)| SPECIES SPECIES (14 es (1-8-8) GROUP (4-3-2) 6 = =|6 (5-2-4) | (9-9-9) 6 " | P 5 5+ 5 y | 4 | f) a 4 la 3 a 1 i 3 \ | eo a 2 abt ea E 9 4 2 J | cs i ae | | ° 0 o L 1 L 4 1 1 1 e) L 1 1 1 L 1 1 PETIOLE STEM ROOT P s R P Ss R PETIOLE STEM ROOT P Ss R P Ss R PETIOLE STEM ROOT P Ss R P Ss R él WOLAYOIUY GIONYV AHL JO TWNYAOL 6S “10A| FIGURE 43. 43 ce er a mixed i PETIOLE STEM ROOT 1. CORYPHOID lelges a 7) Il. PHOENICOID MAJOR GROUP (2=b-19 |.A.TRITHRINAX ALLIANCE (18-15-15) CES eee nee Il. BORASSOID AJOR GROUP Sone, ° bee esmeed hens IV. es F,H. OTHER LEPIDOCARYOID Fo L 0 Ez n IV. A,B,G. LEPIDO- CARYOID ALL. WITH LIANAS | ALLIANCES | ——o _~—L (20-24-8) ld Ss R |.B. LIVISTONA -LEPIDOCARYOID MAJOR GROUP ( 33-32-14) ° -NYPOID MAJOR GROUP (lit) theses under the name of ea group. VI, CARYOTOID MAJOR GROUP (6-5-4) i ao} o x ; ! tit mut 0 ie IX. CHAMAEDORE- Vil. PSEUDOPH HOE- NICOID MAJOR GROU P(2-1-1) ee Sees ee See IX. LARGER CHA- a GROUP -2) ae ee: eee Hees SMALLER CHA- Vill “CEROXYLOI ie) mixed XII. COCOSOID MAJOR GROUP (37-19-14) . | XIIl. A. COCOS ALLIANCE (20-7-6) —o ye ° XU C. BACTRIS [pag Ll L XV. PHYTELE - Se een) eee Ix XIV. SEONOMOID | OID MAJOR GR MAEDOREOID MAEDOREOID MA GROU PHANTOID (14-11-13) SPECIES ECIES peas) MAJOR ono sip (5-2-4) r (9-9-9) si i 3-3=2) Oh i be OE L o o x OR FE - x be - =e "a I ac L es mo. L oo o—o muk o ° =< [ss L joes eee | [ees centers Meee | ! L im To nl [aomea| Ece! Saeaet SL X. IRIARTEOID X|.PODOCOCCOID | Xil. ARECOID PETIOLE STEM ROOT P S R MAJOR GROUP MAJOR GROUP MAJOR GROUP (9-8-5-2) (1-1-1) (45-45-41) SUB- FF TERRANEAN fe nenssy/| uk pee °o mr 4 ! nl PETIOLE STEM ROOT 1 P Ss R preenas SUE?) ee ce values of wide tracheary elements of metaxylem. 42 (above), according to author’s method; headle’s method. Numbers of specimens examined of petiole, stem, and root, respectively, in paren- [SL6I SALVId NOILVYOINAd “ZLOTY — bo 126 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 sequence of groups is similar to the ones that are set forth in the present study (TABLES 1, 2) except that he lists the phytelephantoid palms as lacking vessels in the stem. He found the chamaedoreoid and iriarteoid palms to be “least specialized”; the “Arecoid” (including geonomoid), nypoid, and phoenicoid palms to be “unspecialized”; the caryotoid, coco- soid (including ‘“‘Cocoid” and ‘“Bactroid”’), coryphoid (as ‘‘Sabaloid’’), and nonscandent lepidocaryoid palms to be ‘‘moderately specialized”; and the borassoid and scandent lepidocaryoid palms to be “most specialized.” Tomlinson’s data on the forms of perforation plates in various species of palms (Tomlinson, 1961; 1966; 1969; unpublished data) generally agree with the trends expressed for the major groups in the present study. TYPES OF VASCULAR BUNDLES. In their survey of monocotyledons, Cheadle and Uhl (1948) found that as the wide vessel elements of the metaxylem became more advanced, ‘‘they became larger in relation to the remaining elements in the metaxylem.” In petioles of palms, the most primitive wide vessel elements (i.e., those with the narrowest perforations) occur in bundles of types IITA and IV rather than type I, which is con- sidered to be the most primitive bundle type. The species with type I bundles predominating in the petiole have wide vessel elements of metaxy- lem in which the bars of the perforation plates tend to be slightly more widely spaced (Klotz, 1977). However, this exception to the relationship stated by Cheadle and Uhl (1948) is relatively minor. The present ob- servations in palms generally support the conclusions of these authors. SUMMARY Perforation plates of the wide vessels of metaxylem from the petiole (209 species), the central part of the stem (169 species), and the root (136 species) were examined in representative species from all of the major groups of palms. In nearly all of the species examined, the bars of the perforation plates are most closely spaced in the petiole, most wide- ly spaced in the root, and intermediate (or else similar to either petiole Or root) in the stem. End walls of the wide vessel elements are most oblique in the petiole, least oblique in the root, and intermediate in the stem. Differences in the form of the perforation plates among the tax- onomic groups are mostly average or partial rather than absolute. Nypa has multiple perforation plates in all three organs. Iriarteoid and most chamaedoreoid palms have multiple perforation plates in all organs, but some species also have simple perforation plates in the roots. Ceroxyloid, geonomoid, most arecoid, and most phytelephantoid palms have either simple or mixed multiple and simple perforation plates in the roots and only multiple perforation plates in the shoots. Caryotoid, most cocosoid, and most coryphoid palms have simple perforation plates in the roots, multiple or mixed multiple and simple perforation plates in the stems, and multiple perforation plates in the petioles. Lepidocaryoid palms have multiple, mixed multiple and simple, or simple perforation plates in the stems, and multiple or mixed multiple and simple perforation plates in the 1978] KLOTZ, PERFORATION PLATES lar petioles. The form of the perforation plates is not strictly correlated with habit. For example, the lepidocaryoid lianas all have simple perforation plates in the stems, but only multiple perforation plates occur in the stems of the cocosoid and chamaedoreoid lianas. The form of the per- foration plates shows only partial or inconsistent correlation with en- vironmental moisture regime and with evolutionary advancement in other characters. ACKNOWLEDGMENTS This study is part of a dissertation submitted in partial fulfillment of the requirements for a Ph.D. degree at Cornell University. The author is grateful to Drs. M. V. Parthasarathy and H. E. Moore, Jr. for their gui- dance and assistance during the course of the project, and to Drs. N. W. Uhl, J. M. Kingsbury, P. B. Tomlinson, J. B. Fisher, and V. I. Cheadle for valuable discussion. The study was supported in part by funds from Hatch Project #407. Dr. H. P. Banks permitted the use of his Zeiss Photomicroscope for some of the photomicrography. LITERATURE CITED CarLguist, S$. 1975. Ecological strategies of xylem evolution. xi + 259 pp. niv. California Press, ele CHEADLE, V. I. 1942, The occurrence and types of vessels in the various or- gans of the plant in the Monocotyledoneae. Am. Jour. Bot. 29: 441-450 _ 1943a. The origin and certain trends of specialization of the vessel in the Monocotyledoneae. /bid, 30: 11-17. . 1943b. Vessel specialization in on metaxylem of the various or- gans in the Monocotyledoneae. /bid.: 484-490. & H. Kosakal. 1975. Vessels in Junaes II. Centrolepidaceae and Restionaceae. Am. Jour. Bot. 62: 1017-— & N. UHL. 1948. Types of ene bundles in the Monocotyle- doneae and nee as to the late metaxylem conducting elements. Am. Jour. Bot. 35: 486-496. Esau, K. 1965. aote anatomy. ed. 2. xx + 767 pp. John Wiley & Sons, New Vo rk. FIsHER, J. B. 1975. Environmental impact of lethal yellowing disease of co- conut palms. Environ. Conserv. 2: 299-304. D. A. 1940. Plant microtechnique. xi + 523 pp. McGraw-Hill, — Kiorz, L. H. 1975. Specialization of late metaxylem tracheary elements in palms. P. 45 im Abstracts of papers to be presented at the meetings of the Botanical Society of America and certain affiliated groups at Oregon uae Say Gack Corvalis. | Abstract. | 977. A systematic survey of the morphology of tracheary elements in an vili -++ 221 pp. Unpubl. Ph.D. Thesis, Cornell University, Ithaca, . 1978. Observations on diameters of vessels in stems of palms. Prin- cipes (in press). ManasaLf, T. S. 1959. Resolution of the artificial palm genus, Palmoxylon: a new approach. Palaeobotanist 7: 76-84 128 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Moore, H. E., Jr. ee Pi major groups of palms and their distribution. Gent, Herb, 11: 27- & N. W. URL. a Palms and the origin and evolution of monocoty- ledons. Quart. Rev. Biol. 48: 414-436. PARTHASARATHY, M. V., & L. H. Kiotz. 1976. Palm “wood.” I. Anatomical aspects, Wood Sci. Tech. 10: 215-229 TOMLINSON, P. B. 1961. Anatomy of the monocotyledons, II. Palmae. (C. R. METCALFE, ed.) xv + 453 pp. Clarendon Press, Oxford. 1965. Trends in cell dimensions in palms. Am. Philos. Soc. Yearb. 1965: 354, 355 . 1966. Notes on the vegetative anatomy of Aristeyera spicata (Palmae). Jour. Arnold Arb. 47: 23-29 . The anatomy of the vegetative organs of Juania australis (Palmae). Gent. Herb. 10: 412-424. M. ZIMMERMANN. 1967. The “wood” of monocotyledons. Int. oc. Wood Anat. Bull. 1967(2): 4-24 nt M. H. 1978. Structural peqiiements for optimal water conduc- tion in tree stems. /#: P. B. Tomiinson & M. H. ZIMMERMANN, eds., Tropical trees as living systems. Cambridge Univ. Press, New York (in SECTION OF GENETICS PRESENT ADDRESS: DEVELOPMENT AND PHYSIOLOGY HARVARD FOREST CORNELL UNIVERSITY PETERSHAM, ITHaca, New York 14853 MASSACHUSETTS 01366 1978 | STEVENS, XEROTEAE 129 GENERIC LIMITS IN THE XEROTEAE (LILIACEAE SENSU LATO) P. F. STEVENS THE GENUS Lomandra Labill., variously placed in the Liliaceae Juss. (e.g., Engler, 1888; Krause, 1930). Juncaceae Juss. (Bentham, 1878, 1883), Xanthorrhoeaceae Dumort. (Melchior, 1964; Airy Shaw, 1973; Hutchinson, 1973), or Lomandraceae Lotsy (including Xanthorrhoea; Lotsy, 1911), currently contains about forty-three species. Forty-one of these species are endemic to Australia, growing mostly in the south and east, one species, L. banksii, is known from Australia (Queensland), New Caledonia, and southern New Guinea, while a single species, L. papuana, has hitherto been known from a single collection made in northeast New Guinea. During a collecting trip with Y. Lelean (Division of Botany, Lae, Papua New Guinea) in the mountains of New Britain in 1973, the author was fortunate enough to collect a rather inconspicuous liliaceous plant with the facies of a species of Dianella, but with asymmetric, capsular fruits; an- other collection was made independently by J. Croft and P. Katik. This plant could not readily be placed in a genus, but Professor C. G. van Steenis, of Leiden, drew my attention to Aet & Idjan 24, from W est New Giines, obviously the same taxon as the New Britain plant, and iden- tified as Lomandra. The New Britain and West New Guinea plants were found to be conspecific with L. papuana, a species hitherto known only from a single fruiting collection made in the Madang Province of Papua New Guinea. Lauterbach (1913) tentatively assigned L. papuana to sec- tion EULOMANDRA (correctly LoMANDRA) series Sparsiflorae. The un- usual characters of L. papuana, which had caused my initial difficulties in naming the plant, prompted a survey of the differences between Lo- mandra and related genera, the results of which are presented below. GENERA ASSOCIATED WITH LOMANDRA Seven genera have often been associated with Lomandra in whatever family it has been placed: Acanthocarpus Lehm. (1 species), Chamae- xeros Bentham (3 species), Baxteria R. Br. (1 species), Calectasia R. Br. (1 species), Dasypogon R. Br. (2 species), Kingia R. Br. (1 species), and Xanthorrhoea J. E. Sm. (ca. 15 species). Bentham (1878) divided these genera among three tribes in the Junca- ceae. The Xeroteae Bentham, including Xerotes (a synonym of Loman- dra), Chamaexeros, and Acanthocarpus, were characterized by the small perianth, usually scarious or hyaline, rarely petaloid; dorsifixed anthers; and 3-celled ovary with one laterally attached ovule per loculus. The Xanthorrhoeae Bentham included Xanthorrhoea and Dasypogon, and were distinguished by the perianth, of which the outer whorl was glumelike and 130 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 the inner whorl thin and scarious or petal-like; the dorsifixed anthers; the variable ovary; and the stem, which was either short, thick, and hard, or elongated and woody. The Calectasieae Bentham, including Kingia, Bax- teria, and Calectasia, were distinguished by the rigid, sometimes colored, perianth; the basifixed anthers; and the ovary, which had three ovules. Subsequent classifications are variations of this arrangement. Bentham (1883) merged the Xanthorrhoeae with the Xeroteae. The classification of Engler (1888; in the Liliaceae, subfamily Asphodeloideae) , followed that of Bentham, except that Dasypogon was the sole genus in a new tribe, the Dasypogoneae, characterized by a 1- or incompletely 3- locular ovary, basally attached ovules, and a single-seeded, indehiscent fruit. In the Lomandreae, which had the circumscription of the Xeroteae sensu Bentham (1883), Chamaexeros was included in Acanthocarpus; the Calectasieae were unchanged. Krause (1930) followed Engler, but rein- stated the genus Chamaexeros: Fitzgerald (1903) had earlier removed Xerotes turbinata Endl. to a monotypic genus, Hensmania Fitzgerald, which Krause (1930) included in the Liliaceae-Asphodeloideae- Johnsonieae. Fahn (1954, 1961) examined the leaf anatomy, and, less comprehen- sively, the anatomy of the stem and root, of a great majority of the species in the eight genera, and Suggested that the genera fell into four groups: Lomandra-Chamaexeros-Acanthocar pus, Xanthorrhoea, Dasy pogon- Calec- tasia, and Kingia- Biiena: However, he included Hensmania in Loman- dra (as L. turbinata), but did not find anything distinctive in its anatomy, so his conclusions should be treated with caution. A recent palynological survey of forty-two species from all eight genera (Chanda & Ghosh, 1976) shows a very considerable range of variation of the pollen both between and within (especially in Lomandra) the gen- era. Chanda and Ghosh also recognized four groups of genera; these were, however, mostly differently circumscribed from those of Fahn. Although they placed Lomandra, Chamaexeros, and Acanthocarpus together, they noted that L. micrantha, L. leucocephala subsp. leucocephala, and L. end- licheri should be excluded from the Lomandraceae (i.e., the rest of ee dra plus the two other genera) on palynological grounds (op. cit., p. 550), and placed with Aphyllanthes of the Liliaceae, an advanced member of the family (see also Erdtman, 1966). They did not consider morphological characters, but noted that the groups that they recognized did not correlate very well with those recognized by Fahn (1954, 1961) on anatomical grounds. Other recent work on these genera is that by Lee (1966, 1972), who de- scribed a few taxa in Lomandra and Xanthorrhoea, and Kuchel (1976), who described a third species in Chamaexeros. Burbidge (1963) and Airy Shaw (1973) recognize the same genera as did Bentham (1878). Hamann (1961, pp. 712-714) briefly discusses the relationships of the Xanthorroeaceae. All eight genera agree in their xeromorphic facies and their often rath- er tough perianth parts; in addition, the stem tends to be stout and erect. However, there is considerable variation in inflorescence, in details 1978] STEVENS, XEROTEAE 131 of the perianth, and in stamens, ovary, and fruit, as well as in anatomy (Fahn, 1954, 1961) and palynology (Chanda & Ghosh, 1976). These genera may not all be particularly closely related, but Lomandra, Chamae- xeros, and Acanthocarpus are three genera which appear to be related and which have a number of characters in common. Plants in these three genera are usually rather small, although of varied habit. The inflorescence is many flowered. The tepals are small, often rather tough (but not woody), usually similar in shape and size, and persistent at the base of the capsule, becoming indurated. The six anthers are dorsifixed. The completely 3-locular ovary has one, rarely two, ovules attached to the middle of the axis of the ovary (in one species at the apex) ; the style is not inflated. The 1- to 3-seeded capsule is loculicidal. The seeds have a thin, dull, brown to yellowish brown testa which is often finely reticulate because of the prominence of the anticlinal cell walls (FIGURE 5), and the endosperm, although oily, is tough. The micropyle faces the bottom of the ovary loculus, 90° from the point of attachment of the seed to the placenta (this position will be called basal, since the micropyle is at the bottom of the long axis of the seed). The embryo is straight or slightly curved. There is also substantial agreement between the genera in lamina anatomy. The palisade mesophyll, when present, is bifacial, and there are raphides in some cells. The fibers surrounding the vascular bundles join the epidermis on at least one side of the lamina in all except two species, the vascular bundle as a whole being dorso-ventrally elongated in all species. There is no lignified hypodermis (details from Fahn, 1954, 1961; Kuchel, 1976). Xanthorrhoea, one of the genera that has been included in the same tribe as Lomandra, differs from the Lomandra group in several characters. Xanthorrhoea is often quite a large plant, and the base of the leaf widens abruptly (rare in the Lomandra group). The inflorescence structure is very complicated (Waterhouse, 1967), but the ultimate bracts lack the regular arrangement of those of species of the Lomandra group (see be- low). The inner perianth whorl is somewhat larger and more petaloid than the outer, a condition rare in the Lomandra group, although occur- ring there (e.g., in L. juncea). The seed has a black, crustaceous, almost shiny, and minutely bullate testa; the embryo lies almost horizontally across the seed, being only slightly curved in the vertical plane. The anatomy of the lamina is very distinctive. The hypodermis is sclerified, and the center of the leaf is filled with pithlike tissue. In this pith are radially arranged (i.e., as in the stem), V-shaped vascular bundles with phloem occupying the arms (see Fahn, 1954, 1961). Chromosome numbers in Xanthorrhoea are consistently n = 11; those of Lomandra are more vari- able (Briggs, 1966), being based on » = 7 and m = 8 (Lee, 1966). Chromosome numbers in other genera of the Xeroteae are unknown. Dasypogon, the other genus which has been included in the Xeroteae, has a capitate inflorescence superficially similar to that of some species of Lomandra, but with a different arrangement of bractlike structures and flowers (compare FicuRE 2, b, with Ficure 4, g, h). There are major 132 JOURNAL OF THE ARNOLD ARBORETUM [ VOL: 59 differences in the flower: the tepals are large, the outer three being con- nate and woody toward the tips while the inner three are free and rather delicate (in Lomandra, any connation of the tepals affects both series) ; the ovary is imperfectly 3-locular, with 3 basal ovules: the style is swollen. The fruits are reported to be 1-seeded and indehiscent (Bentham, 1878). In anatomy, Dasypogon differs from members of the Lomandra group: It lacks raphide cells in the mesophyll, it has only abaxial palisade tissue, and it has vascular bundles in groups of three surrounded by a common fiber sheath, the whole being more or less circular in transverse section and not reaching the epidermis (Fahn, 1954, 1961). The other genera which have consistently been placed near Lomandra, but not in the same tribe, are also markedly different from the Lomandra group in anatomy (Fahn, 1954, 1961) and in many details of inflorescence, flower, and fruit. Hensmania turbinata (Endl.) Fitzgerald, once included in Xerotes (see above), differs from the genera once included in the Xan- thorrhoeaceae in having a fleshy perianth, only three stamens with the thecae separated by a broad connective, and black, nitid, carunculate seeds with a soft endosperm. The monotypic genus AphAyillanthes (A. mon- speliensis L. is the only species), which Chanda and Ghosh (1976) con- sidered to be similar in pollen type to Lomandra mucrantha, L. leucocephala subsp. leucocephala, and L. endlicheri, is also a very different plant from both the Xanthorrhoeaceae as a whole and the Lomandra group in partic- ular. The individual flowers of the solitary to 3-flowered, capitate inflo- rescence are surrounded by a spiral involucre of 4 to 5 “bracts” that are connate at the base. The perianth is large, brightly colored, and fleshy, and the black seeds have a minutely bullate testa and a soft endosperm. Aphyllanthes also has bisexual flowers, a long style, and a stigma that has three spreading (rather than recurved) lobes, all additional differences between it and the three species of Lomandra mentioned, as well as most other species of the genus (see below). Lomandra papuana, however, agrees with all the details in the descrip- tion given above for members of the Lomandra group, except that its em- bryo is strongly curved, and so further discussion will be restricted to a comparison of L. papuana with Lomandra, Chamaexeros, and Acanthocar- pus. Unqualified mention of Lomandra may be taken to include all species currently included in the genus Lomandra except L. papuana: the Xero- teae will be used to refer only to these four taxa, essentially the same circumscription as when it was originally used (Bentham, 1878). VARIATION OF CHARACTERS AT THE GENERIC LEVEL WITHIN THE XEROTEAE The main morphological differences between Lomandra papuana, Cha- maexeros, Acanthocarpus, and the other species of Lomandra are sum- marized in TABLE 1. The characters are discussed below, and the anatomy of L. papuana is described. TABLE 1, Variation of some characters in Lomandra, L. papuana, Chamaexeros, and Acanthocarpus.* CHARACTER Lomandra Chamaexeros Acanthocar pus Lomandra papuana LEAF MARGIN prickly, rarely scarious prickly smooth scarious or almost smooth LEAF BASE AURICULATE rarely ~ + a FLOWER CLUSTERS IN INFLORESCENCE + — — + as CLEARLY CYMOSE FLOWER TYPE ON SINGLE PLANT usually [ 3 ], ae | fe) 3 9) rarely [U],[¢ 2], [% d],or[g 2] STIGMA PROMINENTLY RECURVED-TRIFID _ — ~ — STYLE LESS THAN TWICE AS LONG AS OVARY + — -- + NUMBER OF OVULES PER LOCULUS 1 2 1 2 POSITION OF OVULES IN LOCULUS median median near apex median SEED FOVEATE - _ _ + EMBRYO + STRAIGHT, MICROPYLE BASAL + + + = FRUIT WITH SPINES — = a [=] * Explanation of symbols: +, character state as mentioned; proaching that mentioned. —, character state other than that mentioned; [—], character state ap- AVALOWAX ‘SSNAAULS [8261 eft 134 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 LEAF MARGIN. Lomandra papuana has a leaf without papillae, spines, or scarious margin (this last is present only at the base). All species of Chamaexeros have a white, scarious margin to the leaf that runs the length of the leaf, as does L. leucocephala (especially prominent in subsp. leuco- cephala — the decurrent leaf base of Lee, 1966), but there is an important difference between them. Chamaexeros lacks fiber strands in the leaf margin. The margins of C. fimbriata and C. macranthera consist of polygonal cells adjacent to the green part of the leaf and an irregular fringe of cells elongated at right angles to the leaf; the margin of C. serra consists almost entirely of the latter type of cells. Lomandra leucocephala, on the other hand, has fiber strands in the leaf margin, and the rest of the cells are elongated parallel to the long axis of the leaf. The closest approach to the Chamaexeros type of leaf margin in Lomandra perhaps occurs in species like L. sororia, That species has small, earlike flaps of tissue along the leaf edge; these flaps have cells elongated at right angles to the long axis of the leaf. LEAF BASE, Auriculate leaf bases are rare in Lomandra, although they oc- cur in L. obliqua (Lee, 1972, figure 1); normally the leaf widens only gradually toward the base. INFLORESCENCE TYPE. No previous attempt has been made to interpret the inflorescences in the Xeroteae, mainly because of their great reduction and numerous parts. The diagrams given here must be considered merely as a preliminary contribution to the subject, based as they are on rather limited observations of herbarium material; developmental studies are sorely needed. The nomenclature of the parts is a rather difficult question. Lee (1966) used the term “cluster bracts’ for pointed, leaflike bracts subtending both branches and clusters of flowers, and “bracts” for all the other struc- tures. This is somewhat of an oversimplification. In a large cluster of flowers, there may be all intermediates between leaflike bract structures (cluster bracts) and broad structures rounded at the apex (see also Lee, 1966, p. 16). These broader bractlike structures may split down the mid- dle (see also Lee, 1966), resulting in two apparently independent bracts; Arber (1925) gives details of the effect that pressure during development may have on the shape of prophylla in monocotyledons, although it should be noted that Stebbins (1974, and references therein) has a different in- terpretation of prophylla. In some cases the terminal cluster of flowers may lack cluster bracts, while the clusters along the stem may be sub- tended by them (e.g., Lomandra banksii, L. multiflora). The terms “bract” and “bracteole” are impossible to use confidently for similar rea- sons; there may be a variable number of “bracts” surrounding the flower, and the ‘“‘bract and bracteole” of a flower may turn out to be the lowest two bracts surrounding the single flower of a reduced cymose inflorescence. There are sometimes irregular changes in the position of insertion of the 1978] STEVENS, XEROTEAE 135 Ficure 1. Inflorescence Soe a a, b, Lomandra papuana: axillary flower cluster (VGF 4 b, axillary cluster sho owing origin of a eer branch Ag se 41283). C, ae fimbriata, wena cluster (Salis- bu 1 , e, Lomandra hastilis d, axillary cluster near top of inflorescence yd (George s.n., ti July 1961); e subcluster from axillar ot near base of in- florescence (Canberra Botanic Gardens 22976, coll. Philips, 23 Oct. 1972). C (Filled crescents = cluster bracts; unfilled crescents = bractlike structures; filled circles or ovals = flowers; unfilled circles = inflorescence axis; parallel lines = axis of cluster; arrow = inflorescence branch; x = structure unclear.) various bractlike structures (see Ficures 1, a; 2, a), but these must be treated with caution because of distortions in the dried material. This being said, it has been possible to establish some taxonomically important variation in inflorescence type. The terms “bract,” “bracteole,” and “cluster bract’’ are used despite the caveats entered above. The term “cluster bract’’ denotes a bractlike structure that subtends a cluster of flowers along the stem, although the cluster may be reduced to a single flower Lomandra papuana Lauterbach. The inflorescence axis is fairly long (about 30 cm.) and is irregularly branched. There are small flower clus- ters both at the branching points and along the unbranched axis, especial- ly toward the apex; the clusters are subtended by cluster bracts. These clusters are clearly cymose (FicurE 1, a) and consist of bracts which successively subtend and enclose both the terminal flowers and the rest (the younger part) of the cluster. In the clusters dissected, the flowers arose along three axes; although alternation between the axes was not al- ways regular, this may have been caused by distortion of the dried ma- 136 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 FIGURE 2, Inflorescence structure (diagrammatic). a, b, Pig eecane ie sit: a, complete terminal inflorescence (Drummond 304); b, lowermost flower cluster of well- et iceet inflorescence oe 304). asypogon hookeri, flower cluster (Wilson 274). (Filled crescents = gies ‘acts, unfilled cres- cents = bractlike structures: crosshatche ce crescents = outer whorl of tepals; filled circles = flowers; unfilled circles = inflorescence axis; arrows = vegeta- tive shoots developing from axillary buds.) terial. Branches with racemose growth bearing further cymose clusters may develop from within the cluster (FicurE 1, b). Each cluster is clearly long-lived, flowers probably developing successively for more than 1978 | STEVENS, XEROTEAE ise six months, with pedicels of fallen fruits, ripe fruits, and flowers occurring in the one cluster. Acant} pus. The inflorescence in the sole species of the genus, Acanthocarpus preissii Lehm., is basically similar to that of Lomandra papuana. However, the inflorescence axis is very short, and what appears to be a single, terminal cluster is in fact made up of seyeril clusters (F1c- URE 2, a), each with a cymose structure. The cymose structure of the fadividual clusters is clearly seen in the larger inflorescences (FIc- URE 2, b). The inflorescences are not notably long-lived. As in L. papuana, the whole inflorescence is racemose, and the youngest clusters are borne in the distal part of the inflorescence. Chamaexeros. Chamaexeros macranthera Kuchel, C. serra (Endl.) Bentham, and C. fimbriata (F. Mueller) Bentham have similar inflores- cence structures, although superficially they are very different. The flow- ers are aggregated into clusters, each flower being surrounded by two bractlike structures, possibly bract plus bracteole, inserted opposite each other (FicurE 1, c). No cluster bracts associated with the flower clusters were seen, but there are bracts along the inflorescence axis of C. serra which subtend the branches of the inflorescence. Lomandra.! Lomandra section CEPHALOGYNE (Bentham) Engler. The inflorescence structure in species of this section is very uniform. Each flower is sub- tended by two bractlike structures, “bract’’ plus broad, opposite or sub- opposite “‘bracteole,”’ with or without associated cluster bracts. Pistillate inflorescences consist of a single, terminal head, and the cluster bracts subtend 1 to 4 flowers (F1GURE 4, c—e; note orientation of flower in FIGURE 4, e); in staminate inflorescences there are several clusters of flowers along the inflorescence axis, generally with associated cluster bracts (F1c- URE 4, h), while the terminal clusters of flowers may lack cluster bracts. Both the clusters and the inflorescence axis in section CEPHALOGYNE ap- pear to be racemose. Species examined: Lomandra elongata (Bentham) Ewart, L. glauca (R. Br.) Ewart subsp. glauca and subsp. collina (R. Br.) A. Lee, L. obli- gua (Thunb.) Macbride, L. rupestris (Endl.) Ewart, L. suaveolens ( Endl.) Ewart. ‘Bentham (1878) described Xerotes section SCHOENOXEROS Bentham (= Lo- mandra section SCHOENOLOMANDRA Engler) because it was thought that the barren stems of the included species (L. spartea, L. juncea) were rushlike, while the fertile stems had only sheathing scales at the base. These so-called stems are leaves (see also Fahn (1954), who was unaware of speculations as to the nature of these organs, but who found no anatomical characters leading him to doubt that they were leaves), and often have a small sheath at the base. Similar terete leaves may also be found in L. cylindrica, L. soror d L. micrantha (also noticed b Lee, pers. c Lomandra spartea will be discussed under section LoMANpDRA; L. jumcea under section TyYPHOPSIS oe L. spartea and L. juncea have inverted vascular bundles in the (unrelated) sors also have such inverted bundles. Lomandra dura, with flat leaves, has inverted bundles. 138 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Lomandra section MAcCROSTACHYA (Bentham) Engler. The inflorescence is a ‘spike’ bearing numerous condensed spikelet-like clusters. Each spike- let is subtended by a single cluster bract (FrGuRE 1, d), and the flowers near the top of the spikelet appear to have a bract and a bracteole as- sociated with them. However, even in the spikelets near the top of the inflorescence, the lower flowers are associated with more numerous bract- like structures (FiGURE 1, d), and lower down the inflorescence the bot- tom flower of a spikelet is replaced by a group of flowers which have a cymose arrangement of parts (FicuRE 1, e), so the “bract plus bracteole” represents a reduced, probably cymose, partial inflorescence. Species examined: Lomandra hastilis (R. Br.) Ewart (the only species in the section). Lomandra section TypuHopsis (Bentham) Engler. Elucidation of the inflorescence structure of the two species included here is made difficult by the fact that the bractlike structures are divided almost to the base into numerous hairs. Both species have flowers aggregated into dense heads. In Lomandra leucocephala the cluster bracts are prominent and subtend a number of flowers: each flower is completely surrounded by hairs which apparently represent two or more bractlike structures. In L. juncea the cluster bracts are less prominent, and there appear to be two bractlike structures surrounding each flower. Species examined: Lomandra leucocephala (R. Br.) Ewart subsp. leu- cocephala and subsp. robusta Lee, L. juncea (F. Mueller) Ewart (pos- sibly only superficially similar to the preceding species). Lomandra section LoMANDRA (Eulomandra Engler, Euxerotes Ben- tham). There are about thirty-two species belonging to this section. Bentham (1878) placed the eighteen species that he recognized in three series, based on whether or not the flowers were in clusters and whether or not they were sessile. Most of the species that were included in Xerotes section EUXEROTES (properly XEROTES) series Fasciculatae Bentham and series Sparsiflorae Bentham have clusters which appear to be similar in structure (group A), although varying as to how they are disposed on the inflorescence axis (singly or in whorls). Members of series Glomeratae Bentham, as well as Lomandra spartea and L. multiflora (the last was placed in series Fasciculatae), have a somewhat different inflorescence structure (group B). roup A. The flowers are borne singly (or mostly so) along the axis (Sparsiflorae: F1cuRE 3, a—m), or in clusters which are in turn in whorls (Fasciculatae: FicURE 3, o-w). The inflorescences of a number of the species in the former series appear to be simply racemose, with the brac- teoles laterally placed (the exact position of the bracteole is uncertain — compare FIGURE 1, a, with figures 78 and 83 in Eichler, 1875). How- ever, further bractlike structures, and even seats may develop in asso- ciation with these “bracteoles” (FIGURE tole m, n). This suggests that the racemes and once-branched panicles of these species are derived from more complexly branched inflorescences perhaps with cymose partial in- 1978] STEVENS, XEROTEAE 139 (NSW 49111, staminate plant). d-f, L. filiformis, single flowers: d, usual; e, rare; f, with inflorescence branch (Clemens 44542, staminate plant). g, h, L. effusa, axillary flower (clusters): g, usual; h, rare (Pritzel 392, staminate plant). i-l, L. fibrata, axillary flowers (AD 95941052, pistillate plant). m,n, L. micrantha, cluster I ] ! =e reis- sii, axillary flower cluster (Wilson 3728, staminate plant). w, L. rpurea, axillary flower cluster (Andres s.n., Oct. 1903, staminate plant). x, L. montana, terminal cluster (NSW 88924, pistillate plant). (Filled crescents = cluster bracts; unfilled crescents = bractlike structures; crosshatched crescents = outer whorl of tepals; filled circles = flowers; unfilled circles = inflorescence axis; arrows = inflorescence branches.) 140 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 florescences. The flower clusters of species in series Fasciculatae seem to be similar to, but more complex than, those of series Sparsiflorae (F1c- © (e aN 2 © O ary cluster (Clemens a, terminal cluster (Evans 2689, pistillate plant); b, axill nt). c-e, L. iqua: c, d, flower clusters from a smal ad; e, flower lacking cluster bract (White 10237, pistillate plant). f, L. confertifolia, whorl of small clusters around the stem ( ” 93970, staminate plant). g, L. longifolia, axillary cluster (Clemens 42449, staminate plant). h L. glauca, axillary cluster (Gunn s.n., staminate plant). i, L. banksi cluster (Brass 18849, pistillate plant). (Filled crescent crescents = bractlike structures; crosshatched crescents = outer whorl of tepals; filled circles = flowers; open circles = inflorescence axis; x = structure unclear. ) flower he si, terminal 1978] STEVENS, XEROTEAE 141 URE 3, o—w, especially o-r). In species of group A as a whole, the indi- vidual flowers are not completely ensheathed in bractlike structures, and the bracts are rather narrow and have acute tips. This contrasts with the often very broad, sheathing, innermost bractlike structures in section CEPHALOGYNE and in group B, below. Species examined: Lomandra bracteata Lee, L. brevis Lee, L. es Lee, L. densiflora Black, L. endlicheri (F. Mueller) Ewart,’ L. effusa (Lindley) Ewart, L. fibrata Black, L. filiformis (Thunb.) J. Britten, L. micrantha (Endl.) Ewart, L. pauciflora (R. Br.) Ewart, L. preissit (Endl. ) Ewart,2 L. purpurea (Endl.) Ewart,? L. sororia (F. Mueller) Ewart. Group B. Species in this group have flowers in clusters grouped in whorls along the inflorescence axis. Prominent, pointed cluster bracts oc- cur in many species, and in all species each flower is completely sur- rounded by ae structures, usually two or more in number (FIc- uRrES 3, x; 4, a, b, f, g, i). Terminal clusters in Lomandra longifolia, L. banksii, etc., ae lack cluster bracts, although these are associated with the cistern: Alone the axis of the same inflorescence. The number of bract- like structures surrounding the flower may vary within a species (e.g., in L. longifolia). Species examined: Lomandra banksii (R. Br.) Lauterbach, L. dura (F Mueller) Ewart, L. confertifolia (F. M. Bailey) Fahn, L. longifolia Labill., L. montana (R. Br.) Fraser & Vickery, L. multiflora (R. Br.) J. Britten, L. ordii (F. Mueller) Ewart, L. patens A. Lee, L. rigida Labill., L. spartea (Endl.) Ewart, ZL. spicata A. Lee. Summary of inflorescence variation in the Xeroteae. The in- florescences examined seem to belong to two types. I. Acanthocarpus and Lomandra papuana have clearly cymose flower clusters, the single bract associated with each flower completely ensheathing both that flower and the younger part of the cluster. At least toward the base of the large and apparently complex inflorescence of Lomandra macrostachya, there are clearly cymose clusters of flowers; these clusters are borne on racemose side branches of a racemose main axis. This type of inflorescence, with cymose ultimate branch units borne on a racemose (monopodial) main axis, is common in monocotyledons (Tomlinson, 1970 —see also diagrams in such books as Eichler, 1875; Stebbins (1974, p. 314) considers monocotyledons to have basically racemose or spicate in- florescences). It is also possible that the species of Lomandra in group A of section LoMANDRA have an inflorescence of a similar type, but much more reduced. II. Those species in which each flower is surrounded by two or more bractlike structures (Chamaexeros and the rest of Lomandra) are not so easily related to an inflorescence with cymose clusters borne racemosely along an axis, although it is possible that each flower represents such a cymose cluster. ? Species with flower clusters in whorls. 142 JOURNAL OF THE ARNOLD ARBORETUM VOL. 59 “IGURE 5, Capsule and seed type in the Neroteae. a-e, ae a, Lomandra papuana (Aet & Idjan 24); a L. cylindrica (Wilkes s.n.); Aged raexeros fim- briata (Helms 5... MEL 837) ; d, L. Mose ee ee 1893 30) Acanthocar- pus preissti (E. H. Wilson 3 oa f-k, seeds: . papuana | (Aet Idjan 24) 1978] STEVENS, XEROTEAE 143 FLOWER TyPE. Lomandra has often been considered to be a genus of dioecious plants, but see Lee (1966). Perfect flowers, or flowers of the type other than that predominating on the plant, are to be found quite frequently in the otherwise staminate or pistillate inflorescences of some species (e.g., L. bracteata), and L. hermaphrodita was described from plants with perfect flowers, as its name implies. Lomandra papuana, Acan- thocarpus, and Chamaexeros have flowers which appear to be perfect. STIGMA AND STYLE. The stigma of Lomandra is characteristic, being prominently trifid with recurved lobes. The style is usually shorter than the ovary, although somewhat longer in species like L. leucocephala. Lo- mandra papuana, Chamaexeros, and Acanthocarpus have slender, usually long styles (but only about as long as the ovary in L. papuana) and very small stigmas which are not obviously 3-lobed. OvuLE position. All species in the Xeroteae have 3-locular ovaries. The single, pendulous ovules of Acanthocarpus are borne near the top of the loculus. The paired ovules of Lomandra papuana are also clearly pendulous, and one is borne somewhat above the other. The ovules in Chamaexeros and Lomandra are less obviously pendulous; the paired ovules in the former genus are collateral. Fruit. The fruit of Acanthocarpus is densely covered by rather blunt spines (FIGURE 5, e); minute spines also occur near the top of the im- mature fruit of Lomandra papuana. The mature capsule of L. papuana is smooth (FicurE 5, a), and in both species the young ovary is also smooth. All other taxa lack spines on both fruits and young ovaries, al- though the fruits may become more or less transversely wrinkled when ripe (FicurE 5, b-d). SEED SURFACE. Lomandra papuana has prominently foveate seeds, with pits penetrating the endosperm (Ficure 5, f). The testa of a few other members of the Xeroteae may be irregularly wrinkled (F1cuRE 5, g-k), but then the surface of the endosperm is smooth in all cases. Acanthocar- pus, Lomandra, and Chamaexeros thus have similar seeds, but it should be noted that seeds of only twenty-two species of Lomandra and one species of Chamaexeros were examined (this species of Chamaexeros, C. fimbriata, is apparently the only species in the genus of which the seeds are known — Kuchel, 1976). Emsryo, The embryo of Lomandra papuana is vertical and central in position in the upper part of the seed, but then it becomes curved so that the micropyle is in the center of the outer part of the seed. In all other L. cylindrica (Wilkes s.n.); h, L. multiflora Cees 18939); i, A. preissti i Wilson 300); j, C. fimbr iata (Helms 5.m., MEL 8378); k, i een Cerys White 11612). (Scale in millimeters.) 144 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 FicurE 6. Anatomy of Lomandra papuana: a, transverse section of lamina, X 230 (arrow points to crystal mass); b, transverse section of rhizome, 140 (cortex top right); c, transverse section of root, X 300. (All from LAE 58742: photographs courtesy of I. A. Staff. 1978 | STEVENS, XEROTEAE 145 species of the Xeroteae in which seeds have been seen, the embryo is more or less straight for its whole length, the micropyle being near the base of the seed. Anatomy. Dr. I. A. Staff (La Trobe University, Australia) examined the anatomy of the leaf blade and rhizome of Lomandra papuana (LAE 58742), and the description below is taken from notes that he sent. The specimens were swollen by heating in water, embedded in araldite resin, sectioned at about 2 pm., and observed with a Zeiss Standard Universal microscope. Leaf (Ficurr 6, a). Thickness about 150 »m.; more or less isobilateral. Epidermal cells thick-walled, rounded. Stomata on both surfaces; guard cells and accessory cells thick-walled, the latter with a small beak border- ing the pore. Hypodermis present. No clearly differentiated palisade mesophyll; the central mesophyll cells large and rounded, occasionally with a mass of crystals (birefringent in polarized light when viewed under crossed nichols) embedded in an amorphous matrix. Vascular bundle with ad- and abaxial sclerenchymatous girder extending to, but not including, the hypodermis; sheath cells noticeably smaller than the central meso- phyll cells. Rhizome (Ficure 6, b). Epidermis of small, tightly packed, almost rectangular cells. Cortex parenchymatous. Endodermis with U-shaped thickenings. Vascular bundles most numerous just interior to the endo- dermis, amphivasal, although the xylem elements sometimes not complete- ly surrounding the phloem. Ground parenchyma with numerous intercel- lular spaces. No secondary thickening observed. Root (FicuRE 6, c). Mature root decorticates down to the innermost two layers of the cortex; these are heavily lignified and persistent. Endo- dermis heavily thickened on all walls except the outer tangential walls. Vascular tissue embedded in heavily lignified tissue close to the endoder- mis, with alternating areas of xylem and phloem. Pith of angular, iso- diametric parenchymatous cells with numerous small intercellular spaces. Summary of anatomical variation in the Xeroteae. From the in- formation given by Fahn (1954) and confirmed by P. B. Tomlinson (pers. comm.), it seems that there are no major anatomical groupings recognizable within the Xeroteae; the Xeroteae themselves are separable from the other genera associated with them mainly in not having the charac- ters that distinguish these other genera (see p. 131). Although cambium has been found in a number of species of Lomandra (Fahn, 1954; Staff, pers. comm.) and in Acanthocarpus (Tomlinson, pers. comm.), the taxonomic significance of its absence in L. papuana is unclear. The failure of the sclerenchymatous vascular bundle girders to reach the epidermis in L. papuana is uncommon in the Xeroteae, although it has been reported from L. pauciflora and Chamaexeros (Fahn, 1954). In other anatomical details L. papuana is similar to other Xeroteae. 146 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 TAXONOMIC POSITION OF LOMANDRA PAPUANA The relationships between Lomandra papuana, Lomandra, Chamaexeros, and Acanthocar pus appear to be reticulate (see especially Taser 1). Each taxon can easily be distinguished from the others separately, but when one taxon is arias with the other three simultaneously, the distinction is less clear cut. All genera have a xeromorphic facies, and parallelism and superficial convergence obscuring taxonomic relationships in such plants are notorious. Acanthocarpus and Lomandra papuana are similar in flower cluster structure, the spines or papillae on the (young) fruit, and the clearly pen- dulous ovules. Acanthocarpus can readily be separated from all other taxa in the Neroteae by its fruit, which is spiny when mature, while L. papuana can be recognized by its foveate seeds and curved embryo. In facies the two are very different. canthocarpus, with its short. recurved leaves, long stem, and sessile inflorescence, approaches some members of Lomandra section CEPHALOGYNE (LL. mucronata (R. Br.) A. Lee has been compared with Alcanthocarpus (see Lee, 1972); Macbride (1918) con- sidered it to be a species of that genus), although this growth form is un- common in Lomandra. Lomandra papuana, with its long leaves serrate at the apex, its short stem, and its long, branched inflorescence, is similar to the majority of species of Lomandra in these respects. Lomandra and Chamaexeros are similar in facies and inflorescence, and the diagnostic type of scarious leaf margin of the latter genus may be de- rived from a margin like that of some species of Lomandra (e.g., L. so- roria). Variation in leaf type is comparable in the two genera: in Cha- maexeros Closely related species have either terete or flattened leaves, while in Lomandra there is infraspecific variation in leaf type; in both genera there are species with inverted vascular bundles in their leaves (Fahn, 1957). However, the stigma with three well-developed recurved arms which occurs in all species of Lomandra known to me is unique in the group; in stigma and style Chamaexeros approaches L. papuana and Acanthocarpus. Chamaexeros is usually considered to have a single ovule per loculus (see, for example, Bentham, 1878: Kuchel, 1976), but I have consistently found two ovules per loculus in all three species. All pistillate or bisexual flowers of Lomandra dissected had but a single ovule per loculus. The considerable variation in pollen type within Lomandra does not greatly aid the definition of generic boundaries. Chamaexeros (C. serra, C. fimbriata) and Acanthocarpus have similar boat-shaped, sulcate pol- len grains, the sulcus dividing each grain into equal halves; there are no supra-tectal processes (Chanda & Ghosh, 1976). In twenty-four of the twenty-seven species of Lomandra that Chanda and Ghosh examined, the sulcate grains were either more or less spheroid, with supra-tectal pro- cesses, and divided by the sulcus into two equal halves, or disc-shaped. acking processes, and divided by the sulcus into two unequal halves. However, three species, L. micrantha, L. endlicheri, and L. leucoce phala, — Tae — _ Distribution of the Xeroteae in Papuasia and adjacent Queensland. Squares, Rommnalda papuana; circles, Lo- mandra multiflora; triangles, L. banksii. SL6T ‘SNAAALS AVALOYWAX Lyt 148 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 59 had spiraperturate or spiraperturate-type grains which Chanda and Ghosh (op. cit.) considered to be very different from the others. However, | know of no morphological or anatomical evidence to suggest that those three species should be placed in a genus other than Lomandra. (There was too little material of LZ. papuana for palynological examination in the current study. he variation pattern in the Xeroteae is best represented formally by recognizing four genera, Lomandra papuana being removed from Lo- mandra and becoming the type of the monotypic genus Rommnalda (this generic name was chosen to emphasize the apparently reticulate nature of the relationships in the Xeroteae). Chamaexeros, Lomandra, an Acanthocarpus are maintained as genera: all three are natural taxa. It may be noted that if Ewart’s (1916) reduction of Chamaexeros to Acan- thocarpus is followed (see also Mueller, 1878: Engler, 1888), then there is no reason to keep either Acanthocarpus sensu lato or Romnalda sep- arate from Lomandra, but then the resulting heterogeneity in the latter genus would not be taxonomically satisfactory. Romunalda, from the tropical rainforests in the northern part of Papuasia (Map 1), appears to be most closely related to Acanthocar pus, a genus growing in very much drier and more open conditions in sclerophyll vege- tation in southwest Australia. Romnalda is clearly not immediately re- lated to Lomandra, two species of which occur in relatively drier and open conditions in the Western Province of Papua New Guinea (Map 1). The restriction of the Xeroteae within Malesia to the Papuasian region might be expected of such a xeromorphic, Australian-centered group, but there is clearly much work to be done before the rather puzzling relation- ships between members of this group are understood. ARTIFICIAL KEY TO GENERA OF THE XNEROTEAE 1. Plants usually dioecious; stigma - oe or the rare bisexual flowers with three prominent, recurved a en oe .. Lomandra. 1. Plants hermaphroditic; stigma minu es + pin-shaped. 2. Leaves with prominent scarious margins along their lengths (sometimes broken off in old leaves); each flower surrounded by a “bract plus brac- HOON haces aah oy ecu are matbernsedinn salam. Chamaexeros. Leaves lacking scarious margins except at their bases: each flower sur- rounded by only a single “bract.” 3. Inflorescence axes less than 3 mm. long: ripe capsules spiny; ovules ulus bo one per loculus. 9.0... Acanthocar pus. 3. Inflorescence axes ae 20 cm. long; ripe capsules smooth: ovules two DED OCUIUS.. o:dciy a enn hats oe pak nee anee dk dane oo _.. Romnalda. Romnalda P. F. Stevens, gen. nov. A Acanthocarpo, Lomandra et Chamaexeros in seminibus foveatis et embryone curvato differt, a Acanthocarpo, quo aliter simile est, in loculis ovarii biovulatis, non uniovulatis, axe inflorescentiae elongato, non brevi, et capsula matura laevi, non echinata: a Chamaexeros in inflorescentia et 1978] STEVENS, XEROTEAE 149 margine foliae, et a Lomandra in stylo stigmateque et numero ovularum differt. Herba caule breve. Folia longa, apice serrata. Inflorescentia axe elon- gato; flores in cymulis densis aggregatis. Flos hermaphroditus; tepala tenacia, persistentia; stamina 6, antheris dorsifixis; ovarium 3-loculare, loculis ovulis duabus pendentibus plus minusve superpositis provisum; stylus longus; stigma parvum. Capsula loculicida 1—3-seminalis; semen foveatum, endospermio oleoso tenaci, embryone curvato, micropylo pagina extus inter medium basinque testae praesenti. Typus (et species sola): Romnalda papuana (Lauterbach) P. F. Stevens. THE PAPUASIAN XEROTEAE Romnalda papuana (Lauterbach) P. F. Stevens, comb. nov. Lomandra papuana Lauterbach, Bot. Jahrb. 50: 294. 1913; Krause, Bot. Jahrb. 59: 554. 1925. Type: [Papua New Guinea, Morobe Province] Go- romia, 250 m., 8 March 1908, Schlechter 17396 (holotype, B; isotype, WRSL), a Herb, stem prostrate or ascending, to 5 cm. long and 2-3 mm. thick. Leaves linear, apparently spirally arranged, congested, 22-38 cm. long and 4.4-5 mm, wide, acute at the apex, with 5 to 12 small teeth, pale dull green when dry, plane, nerves 12 to 15, margin and surface smooth, at the base widening gradually and with a somewhat scarious edge break- ing irregularly and + persisting as fibers. Inflorescence to 20 cm. long, sparsely and irregularly branched, with single clusters at the branching points and a few along the inflorescence, 1 to 3 together at the end, first cluster often only 1 cm. or so above the base; cluster consisting of a large cluster bract to 3 cm. long subtending numerous bracts plus flowers, each flower terminal, it and the younger part of the cluster ensheathed by a bract; pedicel flattened, 3.2-4 mm. long. Flower apparently hermaphro- ditic; tepals 6, the outer three ovate, greenish, ca. 3.5 mm. long and 1 mm. wide, the inner tepals white, + elliptic, ca. 3.75 mm. long and 0.75 mm. wide: stamens 6, the filaments connate for the basal 0.5 mm., flattened, narrowly triangular, filaments of stamens opposite the outer whorl of tepals ca. 2.2 mm. long, + adnate to the tepals for the basal 0.5: mm, filaments of stamens opposite the inner whorl of tepals ca. 2.5 mm. long, adnate to the tepals for the basal 0.7-1 mm., the anthers ca. 1 mm. long, dorsifixed, dehiscence latrorse; ovary + ovoid, smooth, ca. 1.5 mm. long, 3-locular, ovules 2 per loculus, pendulous, + superposed; style ca. 1 mm. long: stigma small, capitate. Fruit a loculicidal capsule, surrounded by the persistent tepals at the base, green, usually asymmetrically ovoid, 4— 4.7 mm. long and ca. 3.7 mm. across, pointed at the apex, smooth when mature, minutely spiny when young; seeds 1(to 3), slightly flattened- ellipsoidal, brownish, ca. 4.7 mm. long, 3.8 mm. wide and 3 mm. deep, foveate, the testa finely reticulate, the embryo curved, the micropyle facing outward. 150 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 DISTRIBUTION. Papuasia, scattered in the northern part (Map 1). ADDITIONAL SPECIMENS SEEN. Papuasia. IRIAN JAYA. Geelvink Bay: Jap- pen-Biak, Saroerai near Seroei, Aet & Jdjan (exp. L. J. van Dijk) 24 (a, K, L). Papua NEw GuINEA. New Britain: Mt. Klangal, 40 km. N.N.E. of Gasmata, Kandrian subdistrict, West New Britain, 800 m., NGF 41283 (A, B, LAE); map- ping site at edge of Mengen Massif, Pomio subdistrict. oe New Britain, 960 m., LAE 58742 (A, B, LAE). EcoLocy. Romnalda papuana is a very locally abundant plant of col- line rain forest, growing from 250 to 960 meters altitude. The collections from New Britain were made in forest with much Castanopsis acuminatis- sima, in a creek (NGF 41283) or on a dry, stony ridge (LAE 58742). On New Britain Romnalda papuana was growing in the same area as a species of Alpinia section MyRiocraTER, which also has long-lived in- florescences. KEY TO THE PAPUASIAN SPECIES OF LOMANDRA 1. Shrublet to 2 m. tall; leaves dehiscing just above the base, the part remain- ing becoming recurved and persisting. ...................... banksii. 1. Herb, stem a short underground rhizome; leaves ‘not dehiscing aeue the base, but the whole leaf gradually withering. satya te 5 _.. L. multiflora. F. von Mueller was the first to record Lomandra from Papuasia (Muel- ler, 1876; as Xerotes banksii). This record was based on a sterile speci- men collected by J. Orkney (properly Orknie) near the Baxter River (the Mai Kussa, in the Western Province of Papua New Guinea). Lau- terbach (1913) and Krause (1925) repeated this report, but it was not until the 1936-37 Archbold Expedition that the species was re-collected by L. J. Brass. In 1966 a second species, here identified as L. multiflora, was collected from the Western Province. Lauterbach (1913) recorded Lomandra longifolia Labill. from Java, but Lomandra is not included by Backer and Bakhuisen f. (1968) in their Flora of Java and is probably not native there. The two species of Lo- mandra from Papuasia appear to be the only two at present known from Malesia. Lomandra banksii (R. Br.) Lauterb. Bot. Jahrb. 50: 294. 1913; Krause, Bot. Jahrb. 59: 554. 1925; Xerotes banksii R. Br. Prodr. 1: 263. 1810: F. Mueller, Descr. Notes Papuan Pl. 1. 3: 45. 1876. Type: Australia, Queensland, Endeavour River, Banks & Solander anno 1770 (holotype, BM) Undershrub to 2 m. tall; stem 6-7.5 mm. across. Leaves distichous, rather scattered, linear, 25-40 cm. long and 3.5-4.7 mm. across, with (1 to)3 obscure teeth at the apex, pale dull green when dry, plane, nerves 33 to 45, margin minutely scaberulous, broadening somewhat toward the base, at the base with a subpersistent scarious margin splitting irregularly, ca. 0.5 cm. of the leaf persisting and becoming strongly recurved after 1978] STEVENS, XEROTEAE 151 the rest abscises. Plants dioecious: staminate and pistillate inflorescences similar, apparently terminal (growth of the stem continued by an axillary bud developed just below the inflorescence), paniculate, once or twice branched, smooth, lowest whorl of branches borne 5—15 cm. above the base, usually at least 2 cm. between whorls, branches to 12 cm. long, perhaps rather shorter in pistillate than in staminate plants. Staminate plants: flower clusters along the stem each subtended by a cluster bract to 1 cm. long, 1- to 7-flowered, the terminal clusters 10- to 20-flowered, apparent- ly lacking cluster bracts; flowers sessile, each ensheathed by two further bractlike structures up to 1 mm. tall; tepals 6, free, the outer three tepals ovate, 2.2-2.5 mm. long and ca. 1.5 mm. wide, the inner three tepals ovate, ca. 3 mm. long and 1 mm. wide; stamens 6, the filaments flattened, fila- ments of stamens opposite the outer whorl of tepals ca. 1.7 mm. long, ad- nate to the tepals for ca. 0.5 mm., filaments of stamens opposite the inner whorl of tepals ca. 2.1 mm. long, adnate to the tepals for ca. 1.5 mm., the anthers ca. 0.5 mm. long, dorsifixed, dehiscence latrorse; pistillode prom- inent, ca. 1.5 mm. long, stigmas not developed. Pistillate plant: flower clusters along the stem each subtended by a thick cluster bract ca. 5 mm. long, 1- to 3-flowered, terminal clusters 5- to 8-flowered, outer flowers sub- tended by “cluster bracts,” inner flowers lacking them; flowers sessile, usu- ally with a single additional ensheathing bractlike structure, sometimes with two; tepals free, the outer tepals ovate, 4.3-4.5 mm. long and ca. 3.2 mm. across, the inner tepals + triangular, 4.6-4.8 mm. long and ca. 2.8 mm. across, margins incurved; stamens very much as in staminate flowers, but anthers rudimentary; ovary ovoid, 3-angled, ca. 2.7 mm. long and 1.9 mm. across, 3-locular, one ovule per loculus, ovary continuous with the style; style ca. 1.3 mm. long, with three longitudinal ridges; stigma papillate, with three short, retrorse arms. Fruit a oe capsule, surrounded by per- sistent tepals at the base, ovoid, 8-9 m ong and 5—7 mm. across, usu- ally 3-seeded, strongly polnved at are apex, the valves with 4 to 5 fine transverse wrinkles per 1 mm.; seeds ellipsoid, slightly curved or not, orange when dry, ca. 5 mm. long and 3—3.5 mm. across, the testa coarsely and irregularly reticulate, the embryo almost straight, the micropyle basal. DisTRIBUTION. Papua New Guinea, the Western Province, and Aus- tralia (Queensland) (Map 1); ?New Caledonia. SPECIMEN FROM PAPUASIA EXAMINED: Papua New Guinea (Papua). WESTERN : Tarara, Wassi Kussa River, Brass 8700 (pistillate and carpellate plants: A). Ecotocy. Lomandra banksii is reported as being an occasional sub- shrub in dense savannah forests at low altitudes. It was in flower and young fruit when collected in January. The description of the fruit and seed was taken from Brass 18512, collected in the Cape York Peninsula, Queensland, Australia. Agreement between Australian and Papuasian specimens is excellent, and they grow in similar habitats. The staminate inflorescences of Brass 8700 tend to have alternate, rather than whorled, branching, but this is probably an 152 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 abnormality, part of one inflorescence being weakly fasciated. The re- ports of Lomandra banksii from New Caledonia (e.g., Mueller, 1876) should be re-examined. Lauterbach (oc. cit.) was the first to transfer Xerotes banksii to Lo- mandra, making the combination three years before Macbride (1916), who is usually given as the author of the combination. Lomandra multiflora (R. Br.) J. Britten in Banks & Solander, Ill. Bot. Cook’s Voy. 3: 95. pl. 313. 1905: Xerotes multiflora R. Br. Prodr. 1: 262. 1810. Type: Australia, Banks & Solander s.n. (n.v.). ?’Lomandra Ridsdale, Papua New Guinea Sci. Soc. Trans. 8: 22. 1969. Rhizomatous herb, the rhizome near the surface of the ground, ca. 3 mm. across. Leaves + distichous, aggregated, linear, up to 72 cm. long and (0.6—)2.5—4 cm. across, rounded to obtuse at the apex, when alive yellowish green, when dry dull greenish brown, inrolled, the nerves (7 to) 15 to 21, the surfaces and margins of the leaf minutely scaberulous, at the base gradually expanded, the margins dark brown, + scarious, splitting irregularly and not persistent. Plants dioecious. Staminate inflorescence: 50-65 cm. long, scaberulous, especially toward the apex, twice branched the aaa whorled, 2—5.5 cm. between the whorls, the lowest whorl 20-30 from the base of the inflorescence, with 6 to 12 branches, the Race subtending the branches 2-10 mm. long, the branches to 11 cm. long; clusters of flowers borne in whorls of 3 toward the apex of the branches, the cluster bracts to 3 mm. long, the terminal cluster sur- rounded by ca. 3 cluster bracts; flowers 2 to 4 per cluster, each flower surrounded by 2, very rarely 3, “bracts,” the pedicels up to 1 mm. long; flowers greenish yellow; tepals ovate, 6, almost free, the outer three tepals ca. 1 mm. long and 0.7—-0.8 mm. across, the inner three tepals ca. 0.8 mm. long and 0.5 mm. across, thicker; stamens 6, the filaments adnate to the tepals for ca. 0.2 mm., the free part of the filaments opposite the outer whorl of tepals very short, that of the fila- ments opposite the inner whorl up to 0.1 mm. long, the anthers oblong, 0.35—0.45 mm. long, dorsifixed: pistillode minute. Pistillate inflorescence: 28-30 cm. long, scaberulous, unbranched; clusters of flowers in whorls of up to 6, 1-2.5 cm. distant, the cluster bracts to 7.5 mm. long; flowers 2 to 6 per cluster, “bracts” apparently as in staminate flowers; pedicels absent; tepals ovate (persisting at the base of the fruit), the three outer ca. 3.1 mm. long and 2.9 mm. across, the three inner ca. 2.5 mm. long and 1.5 mm. across; ovary and staminodes unknown, but fruit crowned by three persistent, recurved stigmatic lobes. Fruit a loculicidal capsule, at maturity generally with a single seed, asymmetrical, 6-6.5 mm. long, 3.3— 4 mm. across and ca. 5 mm. deep, grayish brown when dry (the sutures brown), with distant transverse wrinkles: seeds slightly curved, ellipsoid, 5.5-6 mm. long, 3.3-4 mm. across and 3-3.5 mm. deep, the testa minutely reticulated, the anticlinal cell walls being prominent, the embryo slightly curved, the micropyle basal. — 1978] STEVENS, XEROTEAE ays DISTRIBUTION. Papua New Guinea, the Western Province (Map 1); widespread in Australia. SPECIMENS FROM PAPUASIA EXAMINED. Papua New Guinea (Papua). WESTERN: near Weam, 30 m., NGF 33573 (staminate: CANB, K, LAE), NGF 33574 (pistil- late: A, CANB, K, LAE ); ca. 5 km. W. of Morehead River at lat. 8° 48’ S., 15 m Paijmans 304 (staminate: CANB). Ecotocy. Lomandra multiflora grows in open savannah; specimens with both flowers and fruits have been collected in August. Although Lomandra multiflora is a very variable species, the specimens described above at first sight look completely different from the common form of the species. This latter has a smooth inflorescence axis, pedicels up to 10 mm. or more long, and staminate flowers over 2 mm. long. How- ever, small-flowered specimens from Queensland (e.g., Brass 18939, Cape York Peninsula) have pedicels about 3 mm. long, and other specimens with larger flowers may have scaberulous inflorescence axes. Hence the broad circumscription of L. multiflora adopted by Bentham (1878) and Lee (1966) is followed. The description above is taken entirely from the Papuasian specimens. ACKNOWLEDGMENTS I am very grateful to the directors of the Adelaide, Lae, Melbourne, Perth, Sydney, and Wroclaw herbaria for sending on loan material of the Xeroteae, and to the directors of the Berlin, British Museum, Canberra, and Kew herbaria for permission to examine material of Lomandra held by their institutions. A. T. Lee and J. T. Waterhouse (Sydney) perused a draft of the manuscript and made some very valuable suggestions. I am also grateful to E. A. Shaw (Gray Herbarium) for reading the manu- script. I. A. Staff (Melbourne) examined the anatomy of Romnalda papuana; J. Jernstedt (Davis, California) examined pollen of some Xeroteae; P. B. Tomlinson (Harvard Forest) discussed aspects of the anatomy of the Xeroteae and their putative relatives. R. Lefberg drew Figures 1 to 4. R. H. Kuchel (Adelaide) and H. Streimann (Canberra Botanic Gardens) kindly sent material of Chamaexeros for examination. LITERATURE CITED ARBER, A. 1925. Monocotyledons: a morphological study. xvi + 258 pp. Uni- versity Press, Cambridge. Backer, C. A., & R. C. BAKHUIZEN VAN DEN BRINK, JR. 1963. Flora of Java. Vol. 3. vii + 761 pp. Wolters-Noordhoff, Groningen — F. M. 1902. The Queensland flora. Vol. 5. 1373-1700 + xi pp. Queens- and Government, Brisbane ae G. 1878. Juncaceae. 7n: Fl. Australiensis 7: 92- . 1883. Juncaceae. /u: G. BENTHAM & J. D. eae Cen. Pl. 3(2)5 861-869. 154 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 BrIccs, re G. 1966. Chromosome se a some Australian monocotyledons. r. New S. Wales Natl. Herb. 4: eee N. T. 1963. Dictionary of ee plant genera. xvil + 345 pp. Angus & Robertson, Sydne CHANDA, S., & K. GHosH. 1976. Pollen morphology and its evolutionary sig- nificance in the Xanthorrhoeaceae. Pp. 527-559 in L. K. FERGUSON & J. MULLER, eds., The evolutionary significance of the exine. Linn. Soc. Symp. 5 _ xii + 591 pp. Academic Press, London. EICHLER, A. W. 1875. Blithendiagramme construirt und erlautert. Part 1. iv + 347 pp. Wilhelm Engelmann, Leipzi ENGLER, - Liliaceae. Ju: A. ENGLER & kK. PRANTL, Nat. Pflanzenfam. II. 5: 10- ERDTMAN, c ee Pollen oe and plant taxonomy. Angiosperms. xii + 553 pp. Hafner, New Ewart, A. J. 1916. ane acre . the flora of Australia, No. 23. Proc. Roy. Soc. Victoria, n.s. 28: 216-222. Faun, A. 1954. The tee aie structure of the Xanthorrhoeaceae Dumort. Jor: Lint. Se. 55: 158-184. 1961. The ea structure 5 eo paca Dumort. and its taxonomic position. Rec. Adv. Bot FITZGERALD, W. V. 1903. Descriptions iy some new species of West Aus- tralian plants. Proc. Linn. Soc. New S. Wales 28: 104-11 HAMANN, U. 1961. Merkmalbestand und Verwandschaftbeziehungen der Fari- nosae. Willdenowia 2: HUTCHINSON, J. 1973. The ilies of flowering plants. ed. 3. xviii + 968 pp. Clarendon Press, Oxford. KRAUSE, K. 1925. Die Liliaceen eke Il. Jn: C. LAUTERBACH, Beitrage zur Flora von Papuasien XII. t. Jahrb. 59: 547-5 67. 1930. Liliaceae. Ju: A. Bie & K. PRANTL, Nat. Pflanzenfam. ed. 2. 15a: 227-386. KUCHEL, R. H. 1976. A taxonomic revision of the genus Chamaexeros Benth. (Xanthorrhoeaceae). Nuytsia 2: 118-12 LAUTERBACH, C, 1913. Die Liliaceen Papuasiens. In: Beitrage zur Flora von Papuasien III (20). Bot. Jahrb. 50: 290-300. LEE, A. T. 1962. Notes oe Lomandra in New South Wales. Contr. New S. Wales Natl. Herb. 3: -164. =, 19 Flora of a South Wales No. 34. Xanthorrhoeaceae. /bid. Flora Series 34: 1- 1972. Lomandra glauca and some related taxa. /bid. 4: 251-261. Lotsy, 7 P. 1911. Vortrage iiber botanische Stammgeschichte. Vol. 3. 1855 pp. Gustav Fischer, Jena. Macesripe, J. F. 1918. Further new or otherwise interesting Liliaceae. Contr. Gray Herb. 56: 1-20. ME tcuior, H. 1964. Liliaceae. Jn: A. ENcrER, Syllabus Pflanzenfam. ed. 12. —666. MUELLER, F. von. 1876. Descriptive notes on Papuan plants. 1. 3: 35-50. 1878. Xerotes ordii. In: Fragmenta phytographiae Australiae 11(88): STEBBINS, G. L. 1974. Flowering plants: evolution above the species level. xviii + 399 pp. Belknap Press, Cambridge, Ma TOMLINSON, P. B. 1970. Monocotyledons — towards an Ot aie of their anatomy and morphology. Adv. Bot. Res. 3: 20 ~ 1978] STEVENS, XEROTEAE 155 WATERHOUSE, J. T. 1968. Some aspects of the status of the family Nanthor- rhoeaceae Hutchinson. iii + 103 pp. Unpubl. Ph.D. thesis, University of New South Wales. WILLIs, J. C. 1973. A dictionary of the flowering plants and ferns. ed. 8. (Re- vised by H. K. Atry SHAW.) xii + 1245 + Ixvi pp. University Press, Cambridge. ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 156 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 RHIZOPHORA IN AUSTRALASIA — SOME CLARIFICATION OF TAXONOMY AND DISTRIBUTION P. B. ToMLINSON THE WESTERN Paciric is known as an area of floristic discontinuity where many species with a wide distribution in the Indo-Malayan region terminate their range (Thorne, 1969). Consequently, precision is required in mapping plant distribution in this area. The important pantropical genus Rhizophora is an example of particular interest because it has long been recognized (Salvoza, 1936; Ding Hou, 1960) that in the western Pacific there is an area of overlap between the group of four taxa with n “Indo-Pacific” distribution (i.e., R. apiculata Bl., R. « lamarckii! Montr., R. mucronata Lamk., R. stvlosa Griff.) and the three species with n “Atlantic” distribution (R. harrisonii Leechman, R. mangle L., racemosa G. F. W. Meyer). The overlap is nacasioned by the extension of the range of the ane species into the western Pacific via a form of Rhizophora mangle sufficiently distinct to be recognized and named as either a variety (var. samoensis Hochr.) or even a separate species (R. samoensis (Hochr.) Salvoza). A consequence of this overlap is the pro- duction of hybrid populations in the form of a taxon here recognized as . selala. Problems have arisen because the diagnostic features which are clear in the field (although small) are often obscured in herbarium specimens, resulting in frequent misidentification of dried material. Also, the area falls outside that covered by Flora Malesiana, so the account by Ding Hou (1958) of the genus in Malesia does not provide details of dis- tribution in the area considered here. Recent field work in this area, partly summarized here, now allows fur- ther precision to be brought to generalized statements and confirms earlier evidence of hybridization (Guppy, 1906; Tomlinson & Womersley, 1976) which is of phytogeographic and evolutionary interest. Direct observation of wild populations in several countries has been supplemented by study of herbarium materials at A, BRIS, GH, K, SUVA, NoU. Measurements re- ported later were all made on fresh or fluid-preserved specimens. Since New Caledonia is one of the areas visited, and as it is particularly rich in Rhizophora species, some of which have been confused in earlier accounts, this island has been given special consideration. One taxon new to its flora, R. * selala, is recognized here. Fiji is also considered in some detail. My present understanding of the distribution of these taxa is sum- marized in Maps 1-3 : name Rhizophora & lamarckii (Montr., Mém. Acad. Sci. Lyon 10: 201. 1860, pro sp.) indicates the hybrid status of this entity (Tomlinson & Womersley, 1976). R.samocnsis A Rusclala * R, slylosa a 1. Distribution of Rhizophora in Australasia. A-C, R. stylosa: A, eralized distribution in the southeastern part of its range; B, details of ie eee fr New Caledonia; C, details of distribution in Fiji. D-F, R. samoen- e : D, distribution in the south-central Pacific (New Caledonia to Samoa); E. details 5 distribution i in New Caledonia; F, details of agus in Fiji. G-I, R. x sela otal known distribution: H, distribut n New Caledonia; I, distribution i in Pili (Most localities refer to nereae Gea, pecan seen by the author, but supplemented by field observations in Fiji, New Caledonia, Ones land, and New Guinea 158 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 TYPIFICATION AND DISTRIBUTION OF RHIZOPHORA & SELALA Guppy (1906) described and illustrated a form of RAizophora in Fiji which was characterized by the absence of fruits and, consequently, of viviparous seedlings. He used the name “‘Selala” to refer to this plant, from the native name signifying empty (“lala”) flowers (‘‘se”), but did not validly describe it. He considered it likely that this entity was a hybrid between the two fertile species he recognized in Fiji (which he called R. mucronata and R. mangle). More recent work supports this interpretation, although with changes in Guppy’s nomenclature. On the basis of Guppy’s descriptions, Salvoza (1936) used the name R. mucronata var. selala. This is unfortunate because Guppy’s use of the name mu- cronata is incorrect. Modern work shows that he confused R. stylosa and R. mucronata. It seems certain that R. mucronata does not occur in Fiji (cf. Richmond & Ackermann, 1975), and that the one parent of the hybrid is actually R. stylosa. The other parent may be referred to most conveniently as R. samoensis. Guppy discussed the difference between Fijian and American populations of R. mangle and concluded that they were morphologically distinct. This is supported by the present work, es) PF Rhizophora xoclala FIGURE 1. Floral features and diagnostic details of Rhizophora x selala. A-C, just-opened flower, < 2: A, side view; B, median longitudinal section; C, from bove. D-F, oral eae x 2: D, style; E, dehisced stamen; F, petal. G, H, inflorescences, 1 ae lee, H, eyes: and with regular dichotomy. I, floral diagram; a o show one of inflorescence branching. (All drawn from P. B. Tomiinson & a eine 18.X1.74 A, from Kinoya, Viti Levu, Fiji 5 1978] TOMLINSON, RHIZOPHORA 159 and the name R. samoensis provided by Salvoza (1936) recognizes this distinction. In contrast, neither Ding Hou (1960) nor Breteler (1977), on the basis of herbarium studies, considers the small differences between the two forms sufficient to merit segregation of R. samoensis from k. mangle. The further problem of Rhizophora in Pacific America can only be solved by extensive field work and the establishment of further di- agnostic differences, which is not within the scope of the present account. Most recently, R. > selala has been shown to occur abundantly in New Caledonia. This suggests that the name used by Guppy should be given formal recognition. Rhizophora x selala (Salvoza) Tomlinson, comb. et stat. nov. FIGURE 1. “Selala” Guppy, Obs. Nat. Pacific 2: 445, 487. frontisp. 1906. cee eee Lamk. var. celal Salvoza in Bull. Nat. Appl. Sci. Univ 5: 219. 1936. Pose ae Tomlinson & Womersley in Contr. Herb. Austral. 19: 9. 1976. Nomen illegit. A Rhizophora mangle differt: foliis mucronatis mucrone demum recur- vato (FIGURE 4); axe inflorescentiae crassiore longiore et floribus (2- 9) plus quam R. mangle (2-6) praedito; gemmis maturis albis ad 14 mm. longis, pedicellum versus angustatis, in sectione transversali rotundis; bracteola manifesta sub flore omni; ovario apice conico in stylum ca. i mm. longum producto. Tyrer. Frontispiece (and p. 445). “Selala, Vanua Levu, Fiji,” Guppy. 1906 OTHER COLLECTIONS. Fiji: Kinoya Village, near Suva, Viti Levu, ?. B. Tom- linson & T. Richmond 18.X1.74 A (A). New Caledonia: common in association with its putative parents in mangrove swamps on both coasts (e.g., Teremba, P. Morat, J.-M. Veillon, & P. B. Tomlinson 18.VII.77 B (a, Nou); Poindimie, P. Morat, J.-M. Veillon, & P. B. Tomlinson 19.V11.77 C (A, NOUV). This species differs from Rhizophora samoensis in having a leaf apex with an appreciable but ephemeral mucro, which becomes recurved with age (FIGURE 4). The inflorescence axis (F1GURE 1, G, H) is longer, thicker, and often with more flowers (2 to 9, sometimes more). The flower buds at maturity are white (not yellow), rounded (not angular) in cross section, up to 14 mm. long, and gradually (not abruptly) narrowed to the pedicel. The bracteole below the flower is distinct (2-3 mm. long) and not obscure as in R. samoensis. The top of the ovary is conical, with a distinct style scarcely 2 mm. long (FicureE 1, D), i.e., shorter than in R. stylosa (Fic- URE 3, C) but longer than in R. samoensis, which has a steeply conical ovary (Ficure 2, E). Populations of R. * selala in New Caledonia and Fiji seem indistinguishable. With fresh material of mixed populations at hand, it is easy to allocate a given specimen to one of the three species; to provide comparative in- 160 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Fic Z, Comparison of Rhizophora samoensis and R. mangle. Above: nee samoensis. A-C, just-opened flower, & 2: A, side view; B, median longitudinal section; c rom above. D-G, details of flowers > D dehisced sta- x 2; tyle, x 2; F, relation between style and dehisced stamens, in median longitudinal view; fer ‘petal, x 2. H, 2-flowered inflorescence, x 1 and x Ws. I, 3-flowered inflorescence at the same magnifications; flowers removed in larger figure to show obscure bracteoles. J, floral diagram (A rawn from P B. Tomlinson & T. Richmond 18.X1. rom Queen Eli abeth Drive, Suva, Viti Levu, Fiji.) Below: Rhizophora mangle, details of inflorescence to show diagnostic features. A— -, 5-, and d inflorescences, respectively, all x %. D, E, details of 2 inflorescences, , showing obvious bracteoles. F, ex- first node. (All drawn from P. B. Tomlinson s.n., Fairchild Tropical Garden, Miami, Florida.) 1978 | TOMLINSON, RHIZOPHORA 161 formation, a series of illustrations emphasizing certain diagnostic features is included in Frcures 1-3. Apart from the morphology of the leaf tip (Ficure 4), R. selala also can be recognized because it often has a trichotomy of the first node of the inflorescence, whereas branching in R. stylosa is exclusively dichotomous. Flower buds in R. x selala are white rather than yellow as in R. samoensis. In addition, illustrations of Rhizo- phora mangle from South Florida show how it differs from R. samoensis in its conspicuous bracteoles (FIGURE 2). Guppy (1906) referred to two forms of Rhizophora selala, the sec- ond distinguished from the one described above by its larger numbers of flowers (up to 24) because the inflorescence is branched 4 to 5 times. I have not seen this second form. RHIZOPHORA IN AUSTRALASIA With the above clarification, it now becomes possible to provide a diag- nostic key to Rhizophora species in this area of overlap in the western Rhijophora slylosa FicuRE 3. Rhizophora stylosa, habit and diagnostic features. A, leafy shoot with axillary inflorescence, X 1%; inset, leaf apex. B, same with leaf blades re- moved to show extensively branched inflorescence. C, style, X 3. D, construc- tional principle of inflorescence. ( rawn from P. B. Tomlinson 30.X.74 from Bootless Bay, Port Moresby, Papua New Guinea.) 162 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Pacific. Distributional information summarized in the key is shown in detail in Maps 1-3. These are drawn up from limited sampling of her- barium material together with field study. Undoubtedly, the range of species needs to be expressed with further precision in the future, but the present summaries are more accurate than the generalized map of Ding Hou (1960). In the following key, measurements represent ap- proximate values and may be exceeded in individual specimens. Only Rhizophora stylosa, R. samoensis, and R. X selala are illustrated; de- tailed drawings of R. apiculata and R. « lamarckii are given in Tomlin- son and Womersley (1976). R.omucronala © Map Rhizophora mucronata. Distribution in Australasia (continued), based on ‘erbartun specimens seen by the author. Its absence from New Cale- onia Fiji is supported by field observation. It is likely to occur in the Seen Archipelago and the Solomon Islands. KEY TO RHIZOPHORA IN AUSTRALASIA a) Apex of leaf with a persistent extended mucro 2-3 mm. long (eg., FIc- (RE 4, D-F); leaf blade usually 10-20 cm. long; inflorescence branching exclusively by dichotomy (e.g., Ficure 3, B, D), never trifurcated, pe- duncle scarcely flattened. Flowers never in 3’s. Mature flower buds white or greenish yellow (never yellow), 15 mm. or more long, rounded in cross section, pedicel of individual flowers often short or obscure. Sta- ens 8, 12, or 16; petals glabrous or hairy. (‘“Indo- — group’ F species group ranging from East Africa to the Western Pacific. Apex oF leaf without a gee hese mucro (e.g., sere 4, A- é G-M); leaf blade usually <10 . long. Inflorescence axis flattened and frequently Henge at fe ne (rarely subsequent) nodes. Pe- duncle often flattened. Flowers commonly in 3’s. Mature flower bud white or yellow, somewhat angular in cross section, <15 mm. long. Pedicel of individual flowers always distinct. Stamens 8: petals always aan ion 1978] ~ = as rat} mn © a & az TOMLINSON, RHIZOPHORA 163 woolly-hairy. (“Atlantic group”; species group ranging from New Cale- Gonias tO WES AICI << 8e Sauda eke sdnd cite eee Ree BRO 5: Stamens 8; petals conspicuously woolly-hairy on the margins. Leat blade usually less than 15 (to as short as 8) cm. long. Mature flower buds white, ca. 15 mm. long. Inflorescence long- -stalked, peduncle usually longer than 2.0 cm., usually branched to more than 2 orders. Flowers usually more than 4 per inflorescence; pedicel evident. Apex of ovary extended abruptly into a style ca. 3 mm. long. (Common and widely ranging throughout the ee Pacific to Samoa and Tonga CNIRE ASO aaa ete mana oe Rhizophora stylosa (FIGURE 3). Apex of ovary bluntly conical, lacking a distinct style. (Extending into the South Pacific as - as the New Hebrides, apparently absent from inconspicuous a eee hairs. Leaf blade commonly 15 cm. long (some- times more). r buds greenish yellow, never white, 15-20 mm. long at maturity. Ly eas short-stalked, peduncle 1-2 cm. long, never branched to more than 2 orders. Flowers 2 to 4 per inflorescence: pedicel short and more or less obscured by bracteoles. ............ 4. Stamens usually 12, petals glabrous. Flowers usually 2 per inflorescence, peduncle short (usually <1 cm.). eee ae developing slowly and at maturity borne below the leafy crown. Bracts and bracteoles thick, rough, corky. Style absent. Plants fertile, with ee fruits and vivip- arous seedlings on the older parts of the twigs. (Common and wide ranging, but not further east than the New Hebrides; in New Caledonia restricted to the northeast coast (Map 3, A, B).) ................... pty Wig tp reed st OS ty gga ee ee hd Rhizophora ‘apiculata. Stamens varying in number, but based on 16; petals sparsely hairy on the margins. Flowers usually 4 per iadecescence peduncle moderately long (1-2.5 cm.). Inflorescence borne within the leafy crown. Bracts and bracteoles not corky. Style distinct. Plants sterile, virtually fae fruits and viviparous seedlings. (Scattered in Papuasia (M always in association with its putative parents, e.g., in New see restricted with R. apiculata to the northeast ar (Map 3, D). seid oh heey fois 4 4a eats eine bags Se ee Rhizophora * gare oe a F, hybrid, R. stylosa X apiculata; Tomlinson & Womersley, ae apex blunt, eee ‘Ficure 4, A-C). Peduncle 2-2.5 cm. long or longer, 2—2.4 mm. in diameter, uncommonly branching beyond 2 orders. Flowers usually , to 5 per inflorescence. Bracteoles scarcely developed and represented by a narrow rim of tissue (Ficure 2, H-I). Mature flower buds yellow, angular in cross section and with a distinct basal shoulder, i.e., abruptly narrowed below, 10-12 mm. long at maturity. Apex of ovary steeply conical and without a distinct style. Plants fertile, with fruits and viviparous seedlings (F1cuRE 2). (Apparently restricted to New Caledonia, Fiji, Samoa, and Tonga (Map 1, D-F); its status in South America is disputed (cf. Breteler, 1977). Rhizophora mangle, apart from its distribution in West Africa and Tropical sae is dis- tinguished by its conspicuous bracteoles and pointed flow shown mi FIGURE: 2.) 22.42. eee ee se Se ns samoensis. Leaf apex initially with an indistinct but usually soon deciduous or re- curved mucro (Ficure 4, H-M). Peduncle 2.5-3.0 cm. long but often recurved apex, X 5 164 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 longer, 2.6-3.2 mm. wide, commonly branched beyond 2 orders. Flow- ers 2 to 9 per inflorescence. Bracteoles distinct, ca. 1 mm. long. Ma- ture flower buds white, neither sharply angular in cross section nor abrupt- ly narrowed to a distinct shoulder at the base, 12-14 mm. long. Apex of ovary extended into a distinct style 1-2 mm. long. Plants sterile, lacking fruits and viviparous seedlings. (Presently known from Fiji and New Caledonia; always in association with its putative parents (Map 1, G- Vee e eed een eee ee ne eee oe ne ob aes Rhizophora X selala. (A probable F, hybrid, R. stylosa & R. samoensis.) G Ficure 4. Leaf morphology of two Rhizophora species and their putative hy- brid. A-C, R. samoensis: A, leaf outline, x 1%; B, C, details of leaf with blunt . D-F, &. stylosa: D, leaf outline, x 1%; E, F, details of i cro, < 5. M, R. X selala: ; ,; H-M, examples of variation in leaf apex (but always with a persistent but recurved mucro, i.e., intermediate between B, C, and E, F), Ore 1978] TOMLINSON, RHIZOPHORA 165 a : R. apiculala @ e of - PASS Rolamarkii wv e -* - ae v Map 3. Distribution of Rhizophora in ee eae es on her- ium specimens seen by the author an nd field observations. A, R. apiculata: sites of its putative parents, R. stylosa (cf. Map 1 1, A-C) and R. ean The characters listed in the key might appear to be elusive, but in the vicinity of Poindimié, New Caledonia, where no less than three species together with two of their hybrids can be found growing within a few hundred meters of each other, there is no problem in naming any indi- 166 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 << vn 7 4 cs r s pals S Qf eed a Co ad i. To x hs oO oO iS _ eae | — || O 4 B a . S a ees = (=) = | oO rca [ = = 2 a ” * e cae ae Me oO Ww U z - ° o | = = f____ 2 . 2a ee - < = eel eee: | ee eee & 8 © 5 $ & B 2 a = ~ 6 Z S = = AODNSNOAYS ADNSANOAYS AONSNOIYS : © = — pa = S Gr) lS 5 | | Aa | | td ons | a. ] }2 L] = a | w L] ry | Wes esi © | : ie . =< 5 By ch) ese | B me 3| 8| & e a oS LI 5 | < ar bes 2 Z | x = = = = _| «| 3) % is ef} os 5G mo (a pe aca = oO ae ee aK <= | a i —| 5 - : —_— \ ¥ Boz WwW ° eeoras ical S D [] o m & spears 2 Sg eg = o a = a , f => —. i ~~ | , ; ‘ S te 3 L 7 7 <4 w g a os _ + [ |e So | 2 a [ a 1 a4 ue N | nN | | i) 2 : Te aaa fe 1 oO , + . _| . = z st | a 2 ee fF = £4 ee eS g . 2-2 ar = : N [aoe = =| = | N =] . — tN = [ N a ies cx os) w clue: sy m it aie \ al — aes - | S) 5 | a + oom oO | | | =) i ase sieebeeasisesss ---} Yo | f Pas < = — aaa nas ‘iB — - en - = a cara ~ a ~ ~ ue | = eee | [ = —- _| nN a & | Ww | | ree Se S e e o 4 ° eS 2 S : g ° ° 2 2 S 5 o AONANOAYS AQNANDIAYA ADN3ANOIYS 5,0. morphology in populations o ora. quantitative variation in certain features of inflorescence 1 me species of Rhizoph e€ Measurem e); R. & lam R. stylosa (below). Parameters represented are theoretical flower number (left), and actual flower number (center). On the right the lines compare average di- 1978] TOMLINSON, RHIZOPHORA 167 vidual once the appropriate diagnostic features have been recognized. This demonstrates that field knowledge is essential for identifying Rhizo- phora taxa. The presence or absence of a persistent mucro will allow pre- liminary segregation of species. Rhizophora apiculata can be recognized at a distance by its larger dark green leaves, an identification confirmed at close quarters by the paired flowers borne below the leafy crown and protected by thick corky bracteoles. Rhizophora % lamarckti has leaves almost equally large, but lighter green. It never has seedlings, and its flowers are borne within the leafy crown in 4’s on a rather lax inflorescence. Any remaining trees with mucronate leaf blades are Rhizophora stylosa; this has medium-sized flowers that are white in bud, and a lax inflorescence usually with several branch orders and more than four flowers. Although Rhizophora mucronata has been recorded for New Caledonia (Guillaumin, 1948) neither field nor herbarium study has revealed its presence. It would be recognizable by its sessile stigma. The two species without a persistent mucro are much more alike, but can be distinguished by the form of the leaf apex, together with the color and shape of the flower bud, and the bracteole structure. Rhizophora X* ae in general, has longer inflorescences with more numerous flowers an R. samoensis. QUANTITATIVE VARIATION Since Rhizophora species are not easily distinguishable, methods were sought to quantify differences between them. Analysis of certain popula- tions by simple biometric methods shows that although variation is large, consistent average differences do exist. Ficures 5 and 6 summarize data from populations of R. apiculata, R. stylosa, and their putative hybrid, R. X lamarckii, mainly in Queensland. Parameters measured were: (1) “Theoretical flower number,” which is the number of nodes in the dichotomously branched inflorescence. Were every axis to ter- minate in a flower, this would indicate the maximum possible flow- er number. (2) “Actual flower number,” which is often less than the theoretical number because of abortion of certain inflorescence axes. (3) “Stamen number,” including larger aborted stamens which are quite numerous in R. X lamarckit. (4) Average length and diameter (at half length) of the peduncle. In these histograms the data for the Queensland Rhizophora stylosa ameter and length of the eae for the three species. Populations measured: Missionary Bay, Hinchinbrook Island, in two localities (Bowen Creek (BC) For R. apiculata one population at Singaua, near Lae, Papua New Gui ; dotted outline) shown for comparative purposes. 6, quantitative ene ‘in stamen number ae oe aborted stamens). Data from same popula- tions as sampled in Ficu 168 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 include two sets from two different but nearby populations (CC and BC), the latter a more vigorous group of trees. To show differences between more distant populations, the data for Rhizophora apiculata include measurements for a population in New Guinea characterized by quite frequent twice-branched (and therefore 3- or 4-flowered) inflorescences and flowers often with less than 12 stamens. It is not known if these differences are genetic. The general result of these histograms is to show that R. & lamarckii is intermediate between its putative parents in certain measurable characters (cf. Tomlinson & Womersley, 1976). An attempt to produce similar histograms for R. * selala was abandoned because the differences between this species and its putative parents were smaller, and insufficient samples were measured. Precise analysis of larger populations would undoubtedly lead to quantifiable expression of differences on a statistical basis, but this morphometric approach is likely to be superseded if other diagnostically useful characters (e.g., chemical) are recognized. CONCLUSIONS In summary, it appears that Rhizophora mucronata has a restricted distribution in the western Pacific, since it is not yet known to range further east than the New Hebrides, but precise assessment of the limits of its eastward range depends on future collecting. Rhizophora stylosa, on the other hand, ranges as far east as Samoa, although it is not known from tropical South America. Rhizophora apiculata does not seem to occur further east than New Caledonia, where it is restricted to the eastern coast. Rhizophora * lamarckii clearly has a wider range in the western Pacific than hitherto appreciated, and it always occurs with its putative parents, kK. apiculata and R. stylosa. Rhizophora samoensis extends as far as New Caledonia, and the result of this overlap between ‘“Indo- Malayan” and ‘‘Atlantic” forms appears to be crossing with R. stylosa and the production of the sterile hybrid, R. > selala, in both Fiji, from where it was first reported, and (as shown here) New Caledonia. These observations are but simple preliminaries to the larger questions about reproductive isolation between Rhizophora species, the breakdown of usual sterility barriers, and the apparent persistence of F, populations without backcrossing. Evidently more precise knowledge of the breeding mechanism of Rhizophora is needed. This should allow hypotheses about the origin and dispersal of Rhizophora in evolutionary time to be put on a less speculative basis. ACKNOWLEDGMENTS Visits to the western Pacific in 1974, 1976, and 1977 have been made possible by grants from the National Geographic Society and a grant (INT 3476-24479) from the National Science Foundation, Office of In- 1978 | TOMLINSON, RHIZOPHORA 169 ternational Programs. Additional support has come from the Atkins Gar- den Fund of Harvard University. Many organizations and individuals have provided support without which such extensive field work would not have been possible. This includes the Directors of the Australian Institute of Marine Science, Townsville; the Director General, O.R.S.T.O.M., and the Director of its center in Nouméa; the Chief, Division of Botany, De- partment of Forests, Lae; the Director of Forests, Port Moresby; the Su- perintendent of the Suva Herbarium; and the Directors of the Arnold Arboretum and the Gray Herbarium, Harvard University. Individuals have included J. S. Bunt, N. C. Duke, D. G. Frodin, M. Galore, Henty, Camilla Huxley, P. Morat, R. B. Primack, Mrs. Thelma Richmond, J.-M. Veillon, Kevin White, and if o. W omersley. Illustrative material is the work of Priscilla Fawcett, Fairchild Trop- ical Garden, and Regula Zimmermann, Harvard Forest. I am indebted to Dr. Elizabeth Shaw, Gray Herbarium, and to Dr. A. C. Smith, Uni- versity of Hawaii, for discussion and clarification of nomenclatural mat- ters. The illustrations of R. stvlosa are reproduced from “Contributions from Herbarium Australiense” by permission of its editor. LITERATURE CITED BRETELER, F, J. 1977. America’s Pacific species of Rhizophora. Acta Bot. Neerl. 26: 225-230. Dinc Hou. 1958. Rhizophoraceae. Pp. 429-493 m — as G. J. VAN STEENIS, ed., Fl. Malesiana. Noordhoff-Kolff N. V., Djak O. A review of the genus Rhizophora, Se a reference to the Pacific species. Blumea 10: 625-634 GUILLAUMIN, A. 1948. Rhizophoraceae. In: Flore analytique et synoptique de la Nouvelle- Calédonie, phanérogames. 369 pp. Office de la Recherche Scientifique Coloniale. Paris. Guppy, H. B. 1906. Observations of a naturalist in the Pacific between 1896 and 1899. Vol. II. Plant-dispersal. 627 pp. Macmillan, London. RICHMOND, T. bE A., & J. M. ACKERMANN. 1975. Flora and fauna of man- grove formations in Viti Levu and Vanua Levu, Fiji. Pp. 179-237 m G. E. WALSH, S. C. SNEDAKER, & H. Teas, eds.. Proceedings, International Symposium on Biology and Management of Mangroves. Gainesville, Florida. SaLvoza, F. M. 1936. Rhizophora. Bull. Nat. Appl. Sci. Univ. Philip. 5: 179- oY) THORNE, R. F. 1969. Floristic ea between New eeow and the Solomon ands Philos. Trans. Roy. Soc. London B. 225: ToMLINSON, P. B., & J. S. WoMERSLEY. 1976. A species of a new to New Guinea ioe Queensland, with notes relevant to the genus. Contr. Herb. Austral. 19: HARVARD FOREST HARVARD UNIVERSITY PETERSHAM, MASSACHUSETTS 01366 170 JOURNAL OF THE ARNOLD ARBORETUM VOL. 59 THE POTAMOGETONACEAE IN THE SOUTHEASTERN UNITED STATES ! RoBert R. HAYNES POTAMOGETONACEAE rete sa Fam. 59, 61. 1829, ““Potamogetoneae,’ . cons (PONDWEED FAMILY ) Glabrous, herbaceous, perennial or rarely annual plants of fresh or brackish waters growing entirely submersed or with both submersed and floating leaves and/or emergent inflorescences. Stems slender, branched or unbranched, often dimorphic, the lower stems rhizomatous and root- bearing, the upper erect and foliaceous, the tips often modified into turions (winter buds). Leaves alternate or subopposite, 2-ranked, entire or toothed, sessile or petiolate, stipulate, 1 to many nerved, the stipules form- ing a tubular sheath around the stem, free from or adnate to the base of the blade, the nerves parallel, connected by perpendicular cross-veins: sub- mersed leaves thin, linear to orbicular; floating leaves often leathery, lan- ceolate, elliptic, or ovate. Inflorescence an axillary or terminal spike or panicle of spikes of perfect flowers. Perianth absent or of 4 separate, rounded, short-clawed, greenish or brownish segments in one series. Sta- mens 2 or 4, the anthers 1- or 2-locular, linear, dehiscing vertically, the filaments aanate to the perianth claw or absent. Gynoecium of 2-16 dis- tinct unilocular carpels, each with a single parietal campylotropous to orthotropous, bitegmic, crassinucellar ovule. Fruit drupaceous, with a membranaceous exocarp, fleshy mesocarp, and stony endocarp. Seed soli- tary, basal cell of the proembryo not dividing but becoming much en- larged, the embryo straight or coiled, with a strongly developed hypocotyl: endosperm helobial in development, porn in mature seed. (Including Ruppiaceae Hutchinson, Fam. Fl. Pl. 2: 48. 1934, nom. cons. Excluding Zannichelliaceae Dumortier, 1829; ae Dumortier, 1829, nomina conservanda.) TyPr GENus: Potamogeton L “Prepared for the Generic Flora of the Southeastern United States, a project of the Arnold Arboretum of Ha — University made possible through the support of the National Science Foundat This treatment, the eighty-fourth in the series, Tennessee, Alabama, Mississippi, Arkansas, and Louisiana. The oe apply primarily to plants of this area, with supplementary information in brackets. Ref- erences that I have not seen are marked by an asterisk. I am indebted to Dr. Carroll E. Wood, Jr., principal investigator on the project, for his advice, suggestions, and help with the literature during the preparation of the manuscript. The illustration of Ruppia was drawn by the late gees H. Marsh under Dr. oe direction largely from material collected by Dr. R. B. Channell in Harrison Coun Wipagrecues that of Potamogeton was ae 2 Karen S. Velmure under my ee n from material that I collected in Alabama. eae No. 2 from the Aquatic Biology Program, The University of Alabama. 1978] HAYNES, POTAMOGETONACEAE ial A family of nearly cosmopolitan distribution with three genera and about 110 species. Two genera, Potamogeton and Ruppia, occur in the southeastern United States. Groenlandia J. Gay (G. densa (L.) Fourr., Potamogeton densa L.) is native to western Europe, North Africa, and southwestern Asia. Members of the Potamogetonaceae have been placed in the Najadaceae by Eyles & Robertson, Fassett, and Gleason (in Gleason & Cronquist) and have been combined with members of Zannichelliaceae and Zosteraceae by den Hartog, Fernald, and N. Taylor to comprise a taxon that has been called respectively Potamogetonaceae, Zosteraceae, and Zannichelliaceae. The Potamogetonaceae, as here interpreted, are separated from the Zannichelliaceae and Zosteraceae by their perfect flow- ers, lack of a spathelike bract, and, in some species, the presence of turions. Thorne’s order Zosterales includes Aponogetonaceae, Juncaginaceae, Po- tamogetonaceae (including Ruppia), Posidoniaceae, Zannichelliaceae, and Zosteraceae. REFERENCES: — A. Water plants. xvi + 436 pp. Cambridge Univ. Press. 1920. [Review the biology of aquatic vascular plants. Potamogetonaceae discussed eas out. | Aston, H. I. Aquatic plants of eae xvi + 368 pp. Melbourne Univ. Press. 1973, | Potamogeton, 275-290; Ruppia, 290-293. | — P, Potamoretonaceae. Nat. Pflanzenfam. II. 1: 194-214. 1889. —_—_— . GRAEBNER. Potamogetonaceae. Pflanzenr. IV. 11(Heft 31): 1-184. 36 re ree BEAL, E. O. A manual of marsh and aquatic vascular plants of North Carolina. N. Carolina ee Exper. Sta. Tech, Bull. 247. iv + 298 pp. 1977. [Pota- mogetonaceae, 39-53, 10 figs. | BOLKHOVSKIKH, Z., V. GRIF, T. Matvejyeva, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. A. A. Feporov, ed. 926 Acad. Sci. USSR. V. L. Komarov Bot. Inst. Leningrad. 1969. [ Potamogetonaceae, 587, 588; includes most chromosome counts through 1964. Gane. M. A. The structure and relationships of the Potamogetonaceae and allied families. Bot. Gaz. 44: 161-188. 3 figs., pls. 14-18. 1907. Coon, C.D. fia B,J; ut, Ee M. Rix, J. SCHNELLER, & M. SEITZ. aes plants of the world. A manual for the identification of the genera of fre water macrophytes. xvii + 561 pp. The Hague. 1974. aa. ceae, 494-496; Ruppiaceae, 507, 508. | Corre, D. S., & H. B. CorreEty. Aquatic and wetland plants of Stee ce United States. xvi + 1777 pp. froutisp. Water as Control r Environ. Protect. Agency. Washington, D. C. 1972. Reissued in e vols. by Stanford Univ. Press. 1975. | Potamogeton, ae 117; Ruppia, 123.) Cronguist, A. The evolution and classification of flowering plants. x + 396 pp. Boston, Massachusetts. 1968. | Potamogetonaceae, 327-330. ] _ A. H. Hotmcren, N. H. Hoimcren, J. L. REVEAL, & P. K. HOLMGREN. Intermountain flora. Vascular plants of the Intermountain West, U.S.A. Vol. 6. 584 pp. Columbia Univ. Press, New York. 1977. [Potamogeton, 24-42; Ruppia, 42, 43. Duncan, W. H. Vascular halophytes of the Atlantic and Gulf coasts of North 12 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 America north of Mexico. Pp. 23-50 in R. J. Rermotp & W. H. QUEEN, Ecology of halophytes. New York. 1974. EAMES, A. J. Morphology of the seer xii + 518 pp. New York, Tor- onto, London. 1961. [Helobiales, 437-440. | EBER, Karpellbau und Plazentationsv eee in der aes der Helobiae. a 127: 273-330. 1934. | Potamogetonaceae, 305- EYLES, °D. E., & J. L. Rosertson, Jr. A guide and key to i aquatic plants of southeastern United States. U. >. Publ. Health Bull. 286. iv + 151 pp. 1 map. |Potamogeton, 69-71; Ruppia, 71.] FasseTT, N. C. A manual of aquatic ane (with revision appendix by OGDEN). iv + oe pp. Madison, Wisconsin, 1957. FP otamogéronaccad. 55-75, figs. 1-81, 99.| FERNALD, M. L. i. manual of botany. ed. 8. a + 1632 pp. New York. 1950. [Zosteraceae tribe Potamogetoneae, 65-80. | GAEVSKAYA, N. S. The role of higher aquatic plants in the nutrition of the als of fresh-water basins. (Transl. from Russian by D. G. MairLanp MULLER, 1969.) 3 vols. Boston Spa, Yorkshire, England. 1966. GLEASON, H. A., & A. Crongurst. Manual of vascular plants of northeastern United States and adjacent Canada. li + 810 pp. Princeton, New Jersey. 1963. [Potamogeton, 33-38; Ruppia, 38, 39.] GRAEBNER, P., & M. FLAHAULT. 6 Familie. Potamogetonaceae. Pp. 394-543 in O. von KIRCHNER ef al., Lebensgeschichte der Blitenpflanzen Mitteleuropas. Band 1, Abt. 1. Stuttgart. 1908. [Six genera; Potamogeton, 400-503; Ruppia, 503-509. } Harapa, I. Cytological studies in Helobiae. I. Chromosome idiograms and a list of chromosome numbers in seven families. ee 21: 306-328. 1956. aye . DEN. The sea-grasses of the world. 275 pp. Amsterdam. 1970. HAyYN R., & W. A. Wentz. Family 3A. ee E. lee IR, & R. W. ScHERY, ee of Panama. Part II. Ann. Missouri Bot. Gard. 62: 1-10. figs. 1-4 HEGNAUER, R. Chemotaxonomie ie: Pasi nzen. Band 2. Monocotyledoneae. 540 pp. Stuttgart. 1963. | Potamogetonaceae, 421-425. | Hetiguist, C. B. Correlation of selected dissolved substances and the distribu- tion of Potamogeton in New England. Ph.D. thesis (unpublished). Uni- versity of New Hampshire, Durham, New Hampshire 12 ee on some eon vascular aquatic planks in New En- gland. Rhodora 79: 445-452. HOEHNE, F. C. Plantas aquaticas. Inst. Bot. Sad Paulo Publ. Ser. D. 168 pp. 1955. [Potamogetonaceae, 41, 45, pls. 20-29. ] HorcuHkiss, N. Underwater and floating-leaved plants of the United States and Canada. U. S. Dept. Interior Fish Wildlife Serv. Bur. Sport Fish. Wildlife Resource Publ. 44. vii + 124 pp. 1967. [Includes all of material of Circ. 187. 1964, on pondweeds. | HutcuHinson, G. E. A treatise on limnology. Vol. 3. Limnological botany. xi + 660 pp. New York. 1975. [Potamogetonaceae discussed throughout. | HuLTEN, E. The amphi-Atlantic plants and their phytogeographical connections. Sv. Vet- akad. Handl. IV. 7(1): 1-340. 279 maps. 1958. | Includes species of Potamogeton. | IrmiscH, T. Ueber einige Arten aus der natiirlichen Pflanzenfamilie der Pota- meen. Abh. Naturw. Ver. Sachsen-Thuringen Halle 2: 1-56. 3 pls. 1858. KADLEC, J. A., & W. A. Wentz. State-of-the-art survey and evaluation : marsh plant establishment techniques: induced and natural. 230 pp. 3 ap- — 1978] HAYNES, POTAMOGETONACEAE 173 pendices. School Nat. Resources, Univ. Michigan, Ann Arbor. 1974. [Lit- erature review of environmental parameters of many genera of aquatic vascular plants; extensive bibliography; list of persons interested in re- search on such plants. Potamogetonaceae discussed throughout. ] Lupnitz, D. Histogenese und Anatomie von Primarwurzeln und Sprossburtigen Wu mala einiger Potamogetonaceae L. (English summary). Beitr. Biol. Pflanzen 46: 247-313. 1969. [Groenlandia, Halodule, Potamogeton, Rup- pia, Zannichellia, Zostera. | Martin, A. C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513-660. 1946. eee e, 550. ] & F. M. UHLER. Food of game ducks in the United States and Canada. U.S. Dept. Agr. Tech. Bull. 634. 156 pp. 153 pls. Marxcrar, F. Bliitenbau und Verwandtschaft bei den einfachsten Helobiae. Ber. Deutsch. Bot. Ges. 54: 191-229. 8 pls. 1936. Mason, H. L. A flora of the marshes of California. vii + 878 pp. Berkeley, California. 1969. [Potamogeton, 49-81; Ruppia, 81-83. ] McATEE, W. L. Waterfowl and their food plants in the sandhill region of Nebraska. U. S. Dept. Agr. Bull. 794: 1-79. 1920. [See also Bulletins 205, 465, 720.] Mrxi, S. New water plants in Asia orientalis I. Bot. Mag. Tokyo 49: 687- 852, 1935. [Potamogeton, 687-690; Ruppia, 687.] MuENSCHER, W. C. Aquatic plants of the United States. x + 374 pp. Ithaca, New York. 1944. [Potamogetonaceae, 27-65, figs. 7-25, maps 12-53.] Ocpen, E. C. Anatomical patterns of some aquatic vascular plants of New York. N. Y. State Mus. Bull. 424. v + 133 pp. 1974. [Potamogetona- ceae, 7-11, maps 23-51, pls. 8-15.] ——., J. K. Dean, C. W. BoyLen, & R. SHELDON. Field guide to the aquatic plants of Lake George, New oe N. Y. State Mus. Bull. 426. iv + 65 pp. 1976. [Potamogetonaceae, 14-18 OciesBy, R. T., A. Voce, J. H. PEverzy, & R. OHNSON. Changes in sub- merged plants at the south end of Cayuga Lake following tropical storm Agnes. pe 48: 251- oe 1976. Raprorp, A. E., H. E. Antes, & C. R. Betz. Manual of the vascular flora of the a, Ixi a 1183 pp. ae Hill, North Carolina. 1968. | Pota- mogetonaceae, 45— RENDLE, A. B. The eee of flowering plants. Vol. I. Gymnosperms and monocotyledons. ed. 2. xvi + 412 pp. Cambridge. 1930. [Potamo- getonaceae, 202-208. ScHENCK, H. Vergleichende Anatomie der submersen Gewachse. Bibliotheca Botanica. Band 1(Heft 1). 67 pp., 10 pls. 1886. SCHUMANN, K. Morphologische Studien. Heft 1. x + 206 pp., 6 pls. Leipzig. 1892. [Potamogetonaceae, 120-154. ScutTHorPE, C. D. The biology of aquatic vascular plants. xvii + 610 pp. London. 1967. [{Potamogetonaceae, 295-297. sane V. Morphological and anatomical studies in Helobiae. I. Vegetative omy of some members of Potamogetonaceae. Proc. Indian Acad. Sci. ee “60: 214-231. 1964; II. Vascular anatomy of the flower of Potamogetona- ceae, Bot. Gaz. 126: 137-144. 1965. SMALL, J. K. Manual of the southeastern wee xxii + ae PP. Chapel Hill, North ea 1933. [Potamogeton, 15-18; Ruppia, 15.] SUBRAMANYAM, K. Aquatic angiosperms. Vili a8 190 pp. aes 1962. [Po- Heer eet 92-98. | 174 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 SwaMy, B. G. L., & N. PARAMESWARAN. The helobial endosperm. Biol. Rev. 38: 1-50. 1963. | Potamogetonaceae, 20, TaAytor, N. Zannichelliaceae. N. Am. Fl. 17: 13- 27. 1909. [Zannichellia, Rup- pia, and Potamogeton. | Tayvor, R. L., & G. A. Mvuttican. Flora of the Queen Charlotte Islands. Part 2. Cytological aspects of the vascular plants. Ottawa. 1968. [Potamoge- tonaceae, 19, 20. | THIERET, J. W. Observations on some aquatic plants in northwestern Minne- sota. Michigan Bot. 10: 117-124. 1971. [Ruppia, Potamogeton, 121-124.] THORNE, R. F. Flowering plants of the waters and shores of the Gulf of Mexico. _S. Fish Wildlife Serv. Fish. Bull, 55(89): 193-202. 1954. A phylogenetic classification of the Angiospermae. Pp. 35-106 in Vol. 9. Evolutionary Biology. M. K. W. C. STEERE, & B. WALLACE, eds. New York. 1976. [See pp. 65, TISCHLER, G. Die iene a der Dene fiir die Verbreitung der Angio- spermen, erlautert an den Arten Schleswig-Holsteins, mit Ausblicken auf andere Fieae ee. Bot. Jahrb. 67: 1-36. 1935. [Potamogetonaceae, 19. Tur, N. M. Potamogetonaceae. Jz: A. L. CABRERA, Flora de la Provincia de Buenos Aires. Parte 1. Coleccion Cientifica del I. N. T. A. Buenos Aires, 1968. [Potamogetonaceae, 279-288. | Unt, N. W. Studies on the floral morphology and anatomy of certain members of the Helobiae. Ph.D. thesis (unpublished). Cornell University, Ithaca, New York. 1947.* Voss, E. G. Michigan flora. Part 1. Gymnosperms and monocots. Cranbrook Inst. Sci. Bull. 55. xv + 488 pp. Bloomfield Hills, Michigan. 1972. [Po- tamogeton, 75-93; Ruppia, 93, 94. Warp, D. B. Checklist of the vascular flora of Florida. Part 1. Univ. Florida Agr. Exper. Sta. Tech. Bull. 726. a pp. 1968. [Potamogetonaceae, 19. ] WENTz, W. A., R. L. Smiru, & J. A. Kaptec. A selected annotated bibliography on aquatic and marsh plants soe their management. 190 pp. + appendix. School of Nat. Resources, Univ. Michigan, Ann Arbor, 1974. ZIEGENSPECK, H. Phylogenie und Physiologie der Potamogetonaceae im Lichte moderner Methoden unter besonderer Beruecksichtigung von Formen aus Uruguay. Revista Sudamer. Bot. 10: 155-164, 197-212. 1953, 1956. KEY TO THE GENERA OF POTAMOGETONACEAE IN THE SOUTHEASTERN UNITED STATES Stamens four; perianth present; stipule free from leaf blade or, if adnate. with the tip free; fruits sessile or very short stalked (see oS BPS pa dp ee eh a ideas eed dan A pect eae ye Eek ee “en Pot oo Stamens two; perianth absent; stipule adnate to leaf blade He oS len (without free tip); fruits long stalked (see FIGURE 2). ...... pa ne 1. Potamogeton Linnaeus, Sp. Pl. 1: 126. 1753: Gen. Pl. ed. 5. 61. 1754 Annual or perennial, glabrous, herbaceous plants growing submersed in fresh (rarely brackish) waters, propagated by seeds, turions, or rhi- zomes. Stems varying in length with water depth, branched or unbranched, terete or compressed, rooting at lower nodes, the nodes occasionally with 1978] HAYNES, POTAMOGETONACEAE 175 oil glands. Leaves submersed or both submersed and floating, alternate to subopposite; submersed leaves pellucid, sessile or petiolate, linear to or- bicular, subulate to obtuse at apex, acute to perfoliate at base, the margins entire or serrate, rarely crimped, the veins 1-35; floating leaves coriaceous, mostly petiolate [rarely subsessile], elliptic to ovate, acute to obtuse at apex, cuneate to rounded or cordate at base, the margins entire, the veins 1—51; stipules tubular, sheathing the stem and young inflorescences, con- nate or convolute, either free or adnate to the base of the submersed leaves, free from base of floating leaves. Turions (winter buds) present or ab- sent, with extremely shortened internodes, divided into inner and outer leaves; the inner few to numerous, rolled into a fusiform structure or un- modified [shortened and oriented at 90° angles with respect to outer leaves]; the outer 1-5 per side, mostly similar to vegetative leaves [cor- rugated near base]. Inflorescence axillary or terminal, a capitate or cylindrical spike or panicle of spikes with 1-20 whorls of flowers, compact or moniliform, with 2—4 flowers in each whorl, submersed or held above surface of water. Perianth of 4, free, rounded, short-clawed segments. Androecium of 4 stamens, the filaments adnate to the perianth claw, the anthers 2-locular, extrorse, the tapetum amoeboid, the pollen spherical to fusiform, sculptured, 3-celled. Gynoecium of 4 carpels, the ovule or- thotropous or campylotropous, the micropyle formed by the inner integu- ment. Fruit dorsally rounded or keeled, beaked. Embryo coiled 1 or more times. Chromosome base number 13 or 14 (Stern). Typr species: P. natans L.; see N. Taylor, N. Am. Fl. 17: 14. 1909. (Name from Greek, potamos, river, and geiton, neighbor.) — PONDWEED. A genus of perhaps 100 species, in two subgenera, POTAMOGETON and CoLEOGETON, both almost cosmopolitan in distribution, and both divided into two sections; represented in the United States by 31 species, some widely distributed, 15 occurring in our area. Subgenus PoTAMOGETON (peduncle mostly rigid and rarely with an en- dodermis, stigmatic papillae small, pollination mostly by wind), sect. ADNATI Hagstrém (stipules adnate for one-half their length to the leaf bases, inflorescence a panicle of spikes), apparently is represented in our area by a single collection of P. Robbinsii Oakes, 2n = 52, from the Mo- bile delta in southwestern Alabama. Although this species occurred in northwestern Georgia at the height of the most recent glaciation (ca. 20,000 years B.P.) (see Watts), its present range is far to the north (from New Brunswick, south only to Delaware, and west to Washington and Oregon). Whether or not a mistake in labeling is involved in the Alabama record is uncertain. Potamogeton Robbinsii is easily recognized by its auriculate leaf bases and its inflorescence, a panicle of spikes. Section AxILLARES Hag- strom (stipules free or adnate for less than half their length, inflorescence a simple spike) consists of about 85 species, mostly of the Eastern Hemi- sphere. Thirteen species occur in our area: P. crispus L. (introduced), P. amplifolius Tuckerman, P. pulcher Tuckerman, P. nodosus Poiret (P. americanus Cham. & Schlecht., P. lonchites Tuckerman, P. fluitans sensu [voL. 59 a, habit of fae eee with ere x 5. c. detail of leaf blade to show carpels visible, & 10; f, flower in vertical section to show relationship of sta- mens to tepals (two stamens cut in half) and two of four carpels, 20; g, tepal with stamen, mn mae . ae h, same, from the side, X 20; i, carpel, side view, at anthesis, }, carpel in vertical section (diagrammatic) to show single parietal campyottopou ovule, < : infructescence, x 6; single drupe, side a x 10; m. same, in vertical section, endocarp hatched, embryo unshaded, X 1 Small), P. natans L., P. illinoensis Morong (P. angustifolius and P. lucens sensu American authors), P. perfoliatus L., P. diversifolius Raf. (P. ca- pillaceus Poiret), P. epihydrus Raf. (P. Claytonii Tuckerman, P. Nuttallit Cham. & Schlecht., P. cavugensis (Wieg.) Hagstrom), P. tennesseensis 1978] HAYNES. POTAMOGETONACEAE EES Fern., P. confervoides Reichenb. (P. Tuckermanti Robbins), P. foltosus Raf. (P. Curtissii Morong, P. niagarensis Tuckerman, P. pauciflorus Pursh), and P. pusillus L. (P. Berchtoldii Fieber, P. panormitanus Biv., P. lacunatus Hagstrém). The exact nature of P. floridanus Small, based on two collections made by A. H. Curtiss in 1886 (Ny) is uncertain (cf. Og- den, 1943, p. 192). Subgenus CoLEOGETON (Reichenb.) Raunk. (peduncles mostly flexible and with an evident endodermis, stigmatic papillae large, pollination most- ly by water), section Convotutr Hagstr. (sclerenchyma abundantly de- veloped, stipule sheaths convolute) is represented in the southeastern United States by Potamogeton pectinatus L. (P. columbianus Suksdorf, P. interruptus Kit.), 2n = 78, which occurs almost throughout the United States. It can be distinguished from other pondweeds by its stipules ad- nate nine-tenths of their length to the leaf blade and by its fruits with an evident beak. Section ConNATI Hagstr. (sclerenchyma poorly developed, stipule sheaths connate) is widespread in the northern United States and Canada but is not represented in our area. The characters important in determining the various taxa of Potamoge- ton are found in the fruiting spikes, fruits, leaves, and turions. Among the most distinctive features of the fruits are the size and shape (whether widest at, above, or below the middle), the absence or presence and shape of keels on the dorsal surface, and the degree of coiling of the embryo. Important characters of the fruiting spike include its position (terminal or axillary), its shape (globose, cylindrical, or moniliform), the number of fruits per spike, and the anatomy of the peduncle. Distinctive features of the leaves include the presence or absence of floating leaves, the degree of adnation of the stipule with the blade of submersed leaves, the number of major veins, and the presence or absence of lacunae bordering the mid- rib of submersed leaves. Evident characteristics of the turions, if present, include the shape of the inner as compared with the outer leaves, and the time of year the turions germinate. Although several vegetative charac- teristics can be used for identifying the taxa of Potamogeton, characters of the fruits and peduncles are the most important ones, and care should be taken to collect only specimens with at least partially mature fruit. The anatomy of Potamogeton has been studied by Ogden (see family references) and Singh (1964). According to Ogden, the stele is composed of fibrovascular bundles, mechanical tissue, and parenchyma cells. The arrangement of the bundles varies with the species. Four patterns can be recognized: 1) proto-type — bundles free, each with one patch of phloem, four median bundles. and a few smaller bundles along the sides; 2) trio- type — three of the median bundles fused to form a single xylem canal, with one patch of phloem on the outer face of the xylem canals and one or two on the inner face; 3) oblong-type— one or two median bundles lacking phloem on each side and with the stele oblong or elliptical in cross section: 4) one-bundled type — only a single bundle, this with four patches of phloem which often are not evident. The stem usually has an endo- dermis in which the cells are either O-cells having the wall of uniform 178 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 thickness, or U-cells in which the walls are thickened on the inner and lateral ne but thin on the outer face. e cortex, which is highly la- cunate, may be without vascular bundles or fibers or with vascular bundles and fibers, or only with fibers at some junctures of the chains of cells separating lacunae. According to Singh, cells of both the upper and lower epidermis have slightly thickened outer tangential walls. On float- ing leaves the upper epidermis has stomata and a distinct cuticle; sub- mersed leaves lack stomata. The flowering structures of Potamogeton have been interpreted various- ly. Miki, Uhl, and Eames have considered each perianth segment to be a sepaloid bract subtending and adnate to the corresponding stamen and have therefore regarded the flower as an inflorescence of four staminate flowers, each with a single stamen and its adnate bract, and four apetalous carpellate flowers, each of a single carpel with one ovule. In this inter- pretation, the spike represents a reduced compound inflorescence. Singh (1965), on the basis of the vascular anatomy and floral morphology of seven species of Potamogeton, chose to consider the flower a normal one with four perianth segments, but Sattler (1965) considered it impossible to classify the structure as either a true flower or an inflorescence, be- cause it displayed characters of both. Sattler questioned the classical use of flower and inflorescence as two separate categories and emphasized the need for more developmental work in the Helobiales (Alismatidae). For the sake of convenience, Singh’s concept is followed in this treatment. Controversy has also existed as to the real nature of the perianth-like structures that are adnate to the stamens. Irmisch (1851), Ascherson (1889), Rendle (1930), and Markgraf (1936) considered the segments to be stamen connectives (outgrowths of the stamens), but both Sattler (1965) and Singh (1965) reached a different conclusion. Sattler, for ex- ample, demonstrated that in Potamogeton Richardsonii these structures are initiated on the floral apex before the stamens and thus cannot be outgrowths of them. He instead considered them to be perianth segments. After inception of the stamen primordia, growth occurs between the peri- anth primordium and that of the stamen, thereby uniting the base of the developing stamen and perianth segment Taxonomic difficulties in the genus are due in part to vegetative plasticity and to hybridization within the genus. Hunt (1962) and Haynes (1974) have reported that certain populations of narrow-leaved pondweeds show seasonal dimorphism. Other dimorphisms occur when populations occupy habitats different from the normal. Fernald separated Potamogeton fo- liosus into two varieties based on size, but Haynes (1974) suggests that these are only different growth forms, plants in flowing waters usually being much larger than those growing in standing waters. In addition. many floating-leaved species produce terrestrial forms that remain after the lake, pond, or pool in which they were growing dries. Reznicek & Bobbette indicate that the terrestrial forms of the P. diverstfolius alliance are often nearly impossible to identify. Taxonomic problems within the genus are compounded by numerous 1978] HAYNES, POTAMOGETONACEAE 179 populations of hybrids intermediate between their two putative parental species. These hybrids, although usually sterile, often form extensive populations by vegetative reproduction from rhizomes or turions. Hag- strom, Ogden (1943), Haynes & Williams, and Hellquist (pers. comm.) have reported probable hybrid populations. e reproductive biology of the genus is largely unstudied. Most species are considered to be wind pollinated, but Thieret (pers. comm.) has noted that dragon flies have been observed to land on Potamogeton nodosus spikes. Although after the insects were collected, Thieret and his students observed pondweed pollen on them, there is no evidence that the insects actually accomplished pollen transfer to another inflorescence. It is not known whether some or all species are self compatible, but some that are considered to be wind pollinated have produced fruit during sea- sons in which the inflorescence never reached the surface of the water (Voss, 1972). Some appear to flower regularly under water and are pre- sumably water pollinated. According to Sculthorpe, certain species of Potamogeton reproduce by rhizome fragments, tubers, or turions. In some, the stem apex is modi- fied by shortening of the internodes into turions that break off an eventually grow to produce new members of the clone. Turions can result in a phenomenal increase in the population, as shown by Yeo, who planted one turion of Potamogeton crispus on 1 April 1963 and found at the end of the growing season that 23,250 turions had been produced! In other species vegetative propagation is by fragmentation of the rhizome or by fragmentation of tubers from the rhizome. All species reproduce to a certain extent by seeds. Muenscher (1936) demonstrated that the germination rate of Potamogeton seeds is greatly reduced by drying. Many species, according to Muenscher, require near- freezing temperatures for a period of one to three months. However, as shown by Sullivan for many of the broad-leaved species, ae re can be induced to germinate without a previous cold treatment if the exocarp and mesocarp are split. The fruits of many pondweeds are among the most common foods of waterfowl (Martin & Uhler), which often digest y the exocarp and mesocarp, the endocarp with the intact seed being ee from the digestive system. oe has shown that a high per- centage of germination may follo The genus is economically onal as a source of food for wildlife. For waterfowl Martin (1951) and Martin & Uhler rank it first in im- portance in the United States and Canada, and Gaevskaya lists 124 species of animals that have been known to feed on the genus. In addition, populations of Potamogeton have often been abundant enough in the southeastern United States to be regarded as weeds warranting the use of various methods of control. REFERENCES: Under family references see especially ASCHERSON, ASCHERSON & GRAEBNER, CHRYSLER, EAMES, IRMISCH, GAEVSKAYA, HARADA, HAYNES & WENTZ, HeEc- 180 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 ER, LUPNITZ, MARKGRAF, MARTIN & UHLER, ee OGDEN (1974), RENDLE, ae i SiNGEE Voss, W ENTz et al., and Un AaLtto, M. Potamogetonaceae fruits. I. Recent and subfossil endocarps of the Fennos candian species. Acta Bot. Fenn. 88: 1-85. 1970. [Potamogeton & Groenlandia. | mogetonaceae fruits. II. Potamogeton Robbinsii, a seldom fruit- ing North American pondweed species. Ann. Bot. Fenn. 11: 29-33. 1974. Aziz, H.,& 5. M. JArer. Potamogetonaceae. /z: Flora of West Pakistan. No. 79. Ll. pps Ore: Bakke, A. L. The foliar transpiring power of the pondweed (Potamogeton na- tans). _Univ. Iowa Stud. Nat. Hist. 12(6): 63-68. 1927. [Stoma illus- BaTE-SMitH, E. C. The phenolic constituents of plants and their taxonomic significance. II. Monocotyledons. Jour. Linn. Soc. Bot. 60: 325-356. 1968. BENNETT, A. Potamogeton mexicanus. Jour. Bot. London 25: 289, 1887. Potamogeton perfoliatus L. var, Richardson Doig, 272 25, [See also P. rufescens, 242-244: P. Zizii, 263-26 Bemerkungen tiber die Arten der Gattung re in Herbarium des K. kK. Naturhistorischen Hofmuseums. Ann. Naturhist. Hofmus. 4(4): 285-294. 1892. . Notes on the Potamogetones of the Herbarium Boissier. Bull. Herb. Boiss. II. 3: 249-260. 1895. . Notes on Potamogeton. Jour. Bot. London 38: 125-130. 1900; 39: 198-201. 1901; 40: 145-149. 1902; 42: 69-77. 1904: 45: 373-377. 1907; Porn oueton. polygontfolius in Newfoundland. Bot. Gaz. 32: 58, 59, 1901. Potamogeton pensylvanicus Cham. et Schlect., introduced to England. Trans. Proc. Bot. Soc. Edinb, 23: 311, 312. 1909, . Notes on Potamogeton. Ibid. 29: 45-53. 1924. Berc, C. O. Limnological relations of insects to plants of the genus Potamo- geton. Trans. Am. Microscop. Soc. 68: 279-291. 1949. Bourn, W. S. Sea-water tolerance of Vallisneria spiralis L. and Potamogeton foliosus Raf. Contr. Boyce Thompson Inst. 6: 303-308. 1934. Boutarp, B., M. Bovitiant, J. CHoprn, & P. LEBRETON. Isolement de | iso- scoparine (C-glucosyl-6 chrysoériol) de Potamogeton natans L. Compt. end, Acad. Sci. Paris 274: 1099-1101. 1972. BRONGNIART, M. A. Nouvelles recherches sur la structure de l'épiderme des végétaux. Ann. Sci. Nat. Bot. II. 1: 65-71. 1834. CARSTENSEN, U. Laichkrautgesellschaften an Kleingewassern Schleswig- an Schr. Naturw. Ver. Schleswig-Holstein 27: 144-179. 1 map. 1 CLARK, W. A. Pondweeds from north Uist (V.-C 110), with a special seas tion of Potamogeton rutilus Wolfg. and a new hybrid. Proc. Univ. Dur- ham Philos. Soc. 10: 368-373. 1946 Cason, E. W., & T. REICHGELT. Potamogeton X crassifolius Fryer in the Netherlands. Acta Bot. Neerl. 3: 398-400. 195 CLos, D. Mode de propagation oe aa au Potamogeton crispus L. Bull. Soc. Bot. France 3: 350-3 56. CLovis, J. F. A new ae ae for West Virginia and a range extension. eee 34: 98. 1969. [| P. Oakesianus. | 1978] HAYNES, POTAMOGETONACEAE 181 Cook, M. T. The development of the embryo-sac and embryo of Potamogeton lucens. Bull. Torrey Bot. Club 35: 209-218. 1908 Coster, B. F. On Potamogeton crispus L. och dess graddknoppar. Bot. Not. 1875: 97-102. 1875. Croizat, L. Observations on the ovary of the Juglandaceae. Southw. Nat. 11: 72-117. 1966. [Includes comments on the flower of Potamogeton. | Bene a E. The genus Potamogeton L. in tropical Africa. Jour. Linn. Soc. 50: 507-540. 1937. a G. Taytor. Studies of British Potamogetons. I. Jour. Bot. London 76: 89-92. 1938; IL. 76: 166-171; III. 76: 239-241; IV. 77: 56-62. 1939; V. 77: 97-101; VI. 77: 161-164; VII. 77: 253-259; VIII. 77: 277-282; 1 97> 30423 WERT 772 342 343 KU. 782 1-1, 1940~, KAIT. 78: 49-66; XIV. 78: 139-147; XV. 79: 97-101. 1941; XVI. 80: 117-120. 1942; XVII-XVIII. 80: 121-124. [See also 80: 21-24. 1942.] & . Two new British hybrid pondweeds. Kew Bull. 12: 332. 1957. Denny, P., & K. A. Lye. The Potamogeton Schweinfurthii complex in Uganda. Kew Bull. 28: 117-120. 1973. EsENBECK, E, Beitrage zur Biologie der Gattungen Potamogeton und Scirpus. Flora 107: 151-212. 59 figs. 1914 FarLtow, W. G. Notes on a fungus parasitic on species of Potamogeton. Ot- awa Nat. 2: 127-129. 1888. FERNALD, M. L. Contributions from the Gray Herbarium of Harvard Uni- versity — No. LXXXVIII. VI. Potamogeton alpinus and P. microstachys. Rhodora 32: 76-83. . . Potamogeton tenuifolius Raf. bid, 33: 209-211. 1931. ———. The linear-leaved North American species of Potamogeton, section Axillares. Mem. Gray Herb. 3: 1-183. pls. 1-40, maps 1-31. Contributions from the Gray Herbarium of Harvard University — No. CXIII. I. A new pondweed from Tennessee. Rhodora 38: 165-169. pl. 412. 1936. . Local plants of the inner coastal plain of southeastern aan Tbid. 39: 321-415, 433-459, 465-491. 1937. [Potamogeton, 380, 381.] Some anemnatophye es of eastern North America. ae 42: 239-276, 281- 302. 1940. [Potamogeton, 2 Francois, L. Recherches sur les plantes aquatiques. Ann. Sci. Nat. Bot. IX. 7: 25-110. 1908. [Potamogeton, 92-102. ] FRANK, P. A. Dormancy in winter buds of American pondweed, Potamogeton nodosus Poir. Jour. Exper. Bot. 17: 546-555. Fryer, A. Notes on pondweeds. 6. On land-forms of Potamogeton. Jour. Bot. London. 25: 306-310. 1887. ——.. Pota ge varians Morong. Notes on pondweeds. /bid. 27: 33-36, 65-67. pls. 286, 287. 1889 & A. aaa The P oiamereions (pondweeds) of the British Isles. 94 pp., 60 pls. London. 1915. Geurs, C. W. Horizontal distribution and abundance of Diaptomus claipes Schacht in relation to Potamogeton foliosus in a pond and under experi- mental conditions. Limnol. Oceanogr. 19: 100-104. 1974. GOPINATH, D. M. An pees study of Potamogeton ee Es (Abstr.) Indian Sci. Cong 5» Bangalore, 1946. Proc. 3: 10 94 GREELEY, J. R. Water ate Re . State Conserv. 5: 14, 15. ene Gricsspy, B. H., & J. SmitH. Application of granular herbicides for the control 182 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 of submerged weeds. (Abstr.) N. Centr. Weed Control Conf. Proc. 15: 40, 41. Hacstro, J. O. Critical researches on the Potamogetons. Sv. Vet.-akad. Hand. VI. 55: 1-281. figs. 1-119. 1916. [A basic paper with many important ob- servations. | Hanna, L. A. The distribution of Potamogeton crispus in North America. Tor- réya 32; 5.1937 HAYNES, R. R. Potamogeton in Louisiana. Proc. Louisiana Acad. Sci. 31: 82- 90. 1968. A revision of North American Potamogeton subsection Pusilli (Pota- mogetonaceae). Rhodora 76: 564-649. 1974. J. L. Reveav. A re-evaluation of Potamogeton fibrillosus Fern. (Po- tamogetonaceae). Rhodora 75: 75-77. 1973. WILLIAMS. Evidence for the hybrid origin of Potamogeton longiligulatus (Potamogetonaceae). Mich. Bot. 14: 94-100. HEGELMAIER, F. Uber die Entwicklung der Bliithentheile von Potamogeton. Bot. Zeit. 28: 281-289, 313-319. 1870. HELpER, R. J. Polar potassium transport and electrical potential oe across the leaf of Potamogeton lucens L. Proc. Nederl. Akad. We . 78: 189-197. 1975.* ERMA. Exchange and polar transport of rae ions across the leaves of Potamogeton. Acta Bot. Neerl. 22: 686-693. HEsLop-Harrison, J. W. Occurrence of the American ere Potamogeton epthydrus Raf., in the Hebrides. Nature 169: 548, 549. 1952 Hi, E. J. Potamogeton Robbinsii. Bot. Gaz. 25: 195, 196. pl. 15. 1898. Hoirerty, G. M. Ovule and embryo of Potamogeton natans. Bot. Gaz. 31: 339-346. pls. 2, 3. : Hunt, G. S. Seasonal aspects of Berchtold’s pondweed. Mich. Bot. 1: 35. 1962. —. . W. Lutz. Seed production by curly-leaved pondweed (Potamoge- ton cris pus) and its significance to waterfowl. Jour. Wildlife Managem. 23: 405-408. 1959. IrmiscH, T. Uber die Inflorescenzen der deutschen Potameen. Flora 34: 81- 93. 1851. . Kurze botanische Mittheilungen. 6. Nymphaea alba und Nuphar lu- teum. 7. Potamogeton densus. 8. Dauer der Ceratophyllum-Arten. Flora 36: 521-528. 1853. Bermerkungen iiber einige Wassergewachse. Bot. Zeit. 17: 353-356. 1859. —. Zur Naturgeschichte des Potamogeton densus L. Flora 42: 129-139, 1 pl. 1859. KAuL, R. B. Anatomical observations on floating leaves. Aquatic Bot. 2: 215- 234. 1976. [Potamogeton, 229.] KLEKOwWSKI, E. J., & E. O. Beat. A study of variation in the Potamogeton capillaceus- diversifolius complex (Potamogetonaceae). Brittonia 17: 175- 181. 65. KULESZANKa, J. Rozw6j ziarn pylku u Potamogeton fluitans. (German sum- mary). ‘Acta Soc. Bot. Polon. 11: 457-462. 3 Lajos, J., & M. Fe.rotpy. Apparent photosynthesis of Potamogeton perfo- tatus L. in different depths of Lake Balaton. Tihany. Biol. Kutatoin- tezetenek. Evk6on. 27: 201-208. 60.* LeBianc, M. Sur les diaphragmes des canaux aeriféres des plantes. Revue Gén. Bot. 24: 233-243. 7 pl. 1912. [P. natans.] 1978 | HAYNES, POTAMOGETONACEAE 183 Liprorss, B. Ueber eee Inhaltskorper bei Potamogeton praelongus Wulf. Bot. Centralbl. 74: 305-313, 372-377. 1898.* LOHAMMAR, A. C. Ma acne inverkan pa Potamogeton-fronas groning. Fauna Fl. Uppsala 1954: 17-32. 1954 Lopinot, A. C. Aquatic weeds. Their identification and control. Illinois Dept. Cons. Div. Fish. Fish. Bull. 4. 47 pp. 1963. [Potamogeton, 24-31.] Love, A, ae evaluation of corresponding taxa. Vegetatio 5, 6: 212-22 54. ———., sa ve, & B. M. Kapoor. Cytotaxonomy of a century of Rocky Mountain orophytes. Arctic Alpine Res. 3: 139-165. 1971. [P. epihydrus, 42.] LowENHAUPT, B. Active cation transport in submerged aquatic plants. I. Ef- fect of light upon the absorption and excretion of calcium by Potamogeton crispus L. leaves. Jour. Cell. Compar. Physiol. 51: 199-208. 1958; II. Ef- fect of aeration upon De eae content of calcium in Potamogeton crispus L. leaves. Ibid. LunpstroM, N. N. Ueber ee Oelplastiden und die biologische Bedeutung der Oeltropfen gewisser Potamogeton-Arten. Bot. Centralbl. 35: 177-181. 1888. Ly Tur Ba & J. L. Guicnorp. Embryogeny of Potamogetonaceae and develop- ment of embryo of Potamogeton lucens L. (In French.) Compt. Rend. Acad. Paris. D. 283: 151. 1976. Martin, A. C. Identifying ai ee seeds eaten by ducks. Jour. Wildlife Managem. 15: 253 ee W.L. Three important wild fee foods U. S. Dept. Agr. Circ. 81. 9 pp. 1918. [Potamogeton, 11-17. ee M. De quelques nouveaux exemples relatifs a l’influence de |’hérédité et du milieu sur la forme et la structure des plantes. Bull. Soc. Bot. France 29: 81-87. 1882. Mix, S. The origin of Najas and Potamogeton. Bot. Mag. Tokyo 51: 472- 480. 1937 Mitter, N. G. Lateglacial plants and plant communities in northwestern New York State. Jour. Arnold Arb. 54: 123-159. 1973. [Potamogeton, 145, fig. 8.] Miser, R. Variation of leaf-form in Potamogeton perfoliatus L. Jour. Indian Bot. Soc. 23: 44-52. 1944. Moore, E. The See in relation to pond culture. U. i. Dept. Int. Fish Wildlife Serv. Fish. Bull. 33: 251-291. pls. 22-39. 1 Moronc, T. L. The Naiadaceae of North America. Mem. aa Bot. Club — 1-65. 1893. [Potamogeton, 11-55.] & E. J. Hitt. How to collect certain plants. Aquatic plants (Naiada- ceae, etc.). Bot. Gaz. 11: 139-141. 1886. MUENSCHER, W. C. a germination of seeds of Potamogeton. Ann. Bot. 50: 805-821. fig. 11. 1936. . Potamogeton pe may grow as an annual. Rhodora 45: 329, 330. 1943. —. Potamogeton latifolius in Texas. Madrono 9: 220-223. 1948. MUuLLIGAN, H. F., A. BARANOWSKI, & R. JoHNson. Nitrogen and phosphorous fertilization of aquatic vascular plants and algae in replicated ponds. I. Initial response to fertilization. Hydrobiologia 48: 109-116. 1976. [P. crispus. | 184 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 OGpEN, E. C. New range-records for linear-leaved species of Potamogeton. Rhodora 44: 406, 407. 1942. . The broad-leaved species of Potamogeton of North America north of Mexico. Jbid. 45: 57-105, 119-163, 171-214. pls. 746-748, figs. 14-16. 1943. Potamogeton tennesseensis new to the Manual range. /bid. 49: 255, 256. 1947. |Gray’s Manual o ny. | . Key to the North eae 2 of Potamogeton. N. Y. State Mus. Cire, 31. 11 pp. 1953 . Potamogetonaceae: Potamogeton. In: C. L. LUNDELL et al., Fl. Tex. 1: 369-382. pls. 48-56. . Potamogeton in New York. N. Y. State Mus. Bull. 423. 20 pp. 1974. Otto, N. E., & T. R. BartLey. Aquatic pests in irrigation systems. U. S. Dept. Int. Water Res. Tech. Publ. 72 pp. [Potamogeton, Postuszny, U., & R. SatrLer. Floral development of Potamogeton densus. Canad. Jour. Bot. 51: 647-656. 6 pls. 1973. [Groenlandia densus. & . Floral development of Potamogeton Richardsonit. ne Jour. Bot. 61: 209-216. 1974. PRINGLE, J. S. Documented plant chromosome numbers 1969: 1. Sida 3: 350, 351. 1969. [P. gramineus. | RAUNKIAER, C. Anatomical Potamogeton-studies and Potamogeton fluitans. Bot. Tidsskr. 25: 253-280. 9 figs. 1903 Reznicek, A. A., & R. S. W. Bospette. The taxonomy of Potamogeton sub- section Hybridi in North America. Rhodora 78: 650-673. 1976. Rosssacu, G. B. Additional phytogeographical notes on plants collected in West Virginia. Castanea 28: 165-169. 1963. [P. zosteriformis.] ST. JOHN, H. A revision of the North American ag of Potamogeton of the section Coleophylli. Rhodora 18: 121-138 Further notes on Potamogeton. Ibid. 20: on 192. 1918. . A critical consideration of Hagstrém’s work on Potamogeton. Bull. Torrey Bot. Club 52: 461-471. fig. 1. 1925, ae R. Perianth ee of Potamogeton Richardsonii. Am. Jour. . 52: 35-41, a. C. Co ae a l'étude du systéme mécanique dans la racine des plantes aquatiques; les Potamogeton. Jour. Bot. Morot 3: 61-72. 9 figs. 1889. Sur les feuilles de quelques monocotylédones aquatiques. Ann. Sci. Nat. Bot. VII. 13: 103-296. 64 figs. 1891. . Notes biologiques sur les Potamogeton. Jour. Bot. Morot 8: 1-9, 21- 43, 45-58, 98-106, 112-123, 140-148, 165-172. 31 figs. SCHILLING, A. i Anatomisch- -biologische Untersuchungen uber die Schleimbildung der Wasserpflanzen. Flora 78: 280-360. 1894. | Potamogeton, 337, 338. SHARMA, A. K., & T. CHATTERJEE. Cytotaxonomy of Helobiae with special ref- erence to the mode of evolution, Cytologia 32: 286-307. 1967. SHINOBU, R. Studies on the stomata of Potamogeton. Bot. Mag. Tokyo 65: 56-60. 3 figs. 1952 Stmes, J. C. Control of the pondweed, Potamogeton crispus, in both flowing and static es with endothal. Northeast. Weed Control Conf. Proc 197,558; 1961. SINHA, A, B. & Y. N. SRIVASTAVA. Unusual formation of winter apices of Po- tamogeton crispus L. during summer season. Curr. Sci. Bangalore 42: 698. 973. 1978] HAYNES, POTAMOGETONACEAE 185 STERN, K. R. Chromosome numbers in nine taxa of Potamogeton. Bull. Tor- rey Bot. Club 88: 411-414. 1961. STUCKEY, R. L. Distributional history of certain eastern North American non- indigenous aquatic and marsh angiosperms. (Abstr.) Am. Jour. Bot. 57 hos; (ots 1970) “Paccrispus. | Swamy, B. G. L., & N. PARAMESWARAN. On the origin of cotyledon and epi- cotyl in Potamogeton indicus. Osterr. Bot. Zeitschr. 109: 344-349. 1962. TAKUSAGAWA, H. Chromosome numbers in Potamogeton. Volumen Jubilare Pro Prof. Sadao Yoshida: 1066, 1067. 1939.* . Cytological studies in the genus Potamogeton of Japan. Bull. Shimane Agr. Coll. 9: 1-34 TEHON, L. R. The present range of Potamogeton crispus L. in North America. Torreya 29: 42-46. 1929. THopay, D., & M. G. Sykes. Preliminary observations on the transpiration current in submerged water-plants. Ann. Bot. 23: 635-637. 1909. [P. lu- cens, THORNE, R. F. Vascular plants previously unreported from Georgia. Castanea 16: 29-48. 1951. [P. illinoensis Morong, P. nodosus Poiret, P. pusillus L., TicHa, I. Some quantitative anatomical characteristics of Potamogeton perfo- ictus L. and P. lucens L. leaves at various insertion heights. (In German.) Biol. Pl. 6: 108-116. 1964. TREVIRANUS, L. C. Vermischte Bemerkungen. 1. Hybernacula des Potamoge- ton crispus. 2. Hybernacula der Hydrocharis Morsus-ranae L. Bot. Zeit. 15: 697-702. 1857 Tur, N. M. Observaciones teratologicas en el género Potamogeton L. Dar- winiana 20: 257-268. 2 pls. 1976. UspenskiJ, E. E. Zur Phylogenie und Ekologie der Gattung Potamogeton. I. Luft-, Schwimm- und Wasserblatter von Potamogeton perfoliatus L. Bull. 13. Watts, W. A. The full-glacial vegetation of northwestern Georgia. Ecology 51: 17-33. 1970. [Potamogeton, 22-26, fig. 7.] WIEGAND, K. M. The development of the microsporangium and microspores in Convallaria and Potamogeton. Bot. Gaz. 28: 328-359. pl. 24. 1899. WILKINSON, R. E. Effects of pelleted herbicides on aquatic weeds. (Abstr.) South. Weed Conf. Proc. 12: 148. 1959. [P. foliosus.] WEBSTER, D. H. Notes on the distribution of epee species and hybrids in Nova Scotia. Proc. Nova Scotian Inst. Sci. 24: 1956 YEO, R. R. Yields of propagules of certain Retice plants. L. Weeds 14: 110- 113. 1966. [P. crispus.] 2. Ruppia Linnaeus, Sp. Pl. 1: 127. 1753; Gen. Pl. ed. 5. 61. 1754. Glabrous, herbaceous, annual or perennial plants growing submersed in brackish or saline waters. Stems slender, terete, branched or unbranched, rooting at the lower nodes. Leaves all submersed, alternate to suboppo- site, sessile, divided into blade and stipular sheath; blade linear or seta- ceous, l-nerved, the margins entire below, minutely serrulate above, the apex acute to truncate; stipular sheath formed by adnation of stipule to blade for entire length of stipule. Turions absent [present with one inter- node, the leaves undifferentiated]. Inflorescence a 2-[1- to few-] flowered 186 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 capitate spike, at first enclosed by sheathing leaf bases; peduncle elongat- ing at anthesis, elevating the inflorescence to or near the water’s surface. often becoming spirally twisted in fruit, then pulling the developing fruit below the surface. Perianth absent. Androecium of 2 sessile anthers, the two 1-loculate halves separated by a broad connective, the dehiscence extrorse, the tapetum amoeboid, the pollen elongate, 4 times as long as broad, arcuate, swollen at the ends and at the center on the convex side, 3- celled, the exine reticulate, exceedingly thin, discontinuous at the swell- ings, the intine rather thick, further thickened in the swollen regions. Gynoecium of 4 or 5 [2-16] distinct, stipitate |sessile] carpels, the gynophore (stipe, podogyne) elongating after anthesis, the placentation parietal, the ovule campylotropous, the micropyle formed by both integu- ments. Fruit asymmetrical, dorsally rounded [or with a prominent trans- versely ridged crest], beaked [or beak absent], long stipitate [or sessile], the exocarp and mesocarp often decaying leaving the operculate endo- carp intact. Embryo straight. TypE species: R. maritima L. (Name dedicated to Heinrich Bernhard Ruppius, 1689-1719, a German botanist.) — DITCH-GRASS. An almost cosmopolitan genus of perhaps 10 species, represented in the United States by Ruppia maritima . _ widgeon-grass, 2” = 20, which oc- curs along the coast from Maine to Texas, then sporadically inland to Michigan, California, and W ee The characters important in determining the various taxa of Ruppia are based on the fruits (whether strongly asymmetrical or only slightly so) ; the carpels (4, 5, or 8-16); the gynophore (slender, stout, or absent) ; the shape of the leaf apex (acute, obtuse, or retuse); the cross sectional structure of the leaf (shape of the lacunae and number of cells separating the lacunae from the epidermis); and the presence or absence of turions. The genus is poorly understood both taxonomically and ecologically, and a thorough revision utilizing data from oe chemistry, and cy- tology, as well as morphology, is needed. (See R According to Ogden (see family references), re oe of the stem and peduncle is very similar to that of the narrow-leaved species of Po- tamogeton. There is a stele with a single vascular bundle, an endodermis of O-cells, and a cortex without interlacunar or subepidermal bundles. The leaf ie a central vein with a lacuna on both sides running the en- tire length of the blade. In their paper on the floral development of Ruppia maritima, Posluszny & Sattler investigated the development t of “the small dorsiventral out- growth on its stamen connective.” (See F1cuRE 2, e, h, i.) Contrary to previous conclusions that these outgrowths are homologous with the rene appendages (tepals) of the Potamogeton flower, they found that . developmental studies reveal a most remarkable difference between Potunoreton and Ruppia. In Potamogeton the appendages are initiated acropetally on the floral apex and develop like typical phyllomes (Pos- luszny and Sattler, 1973, 1974). In contrast . . . in Ruppia maritima the 1978] HAYNES, POTAMOGETONACEAE 187 —=— FIGURE 2. Ruppia. a R. maritima. a, portion of plant with flowers and young fruit, x %; b of shoot to show sheathing leaves (stipules complete- section (diagrammatic) to show ovule, carpel oriented as left carpe 10; h, inflorescence with developing carpels (flower to left), the gynophores elongating, and unfertilized carpels (flower to right), < 5; i, partially mature fruits of two flowers, X 5; j, two mature drupes, X 5; k, endocarp, side view, lide tovlett..< 5.1, endocarp, adaxial side, with lid removed, lid to right, * 5. connective outgrowth develops, unlike either a tepal or a bract, after stamen inception, in fact after the beginning of theca differentiation. It results from growth very slightly abaxial to the tip of the stamen connective. Al- though it becomes dorsiventral, it . . . lacks any vascularization at ma- turity. In general, it is similar to a vertical extension of the stamen con- nective in Potamogeton |Groenlandia| densus (Posluszny and Sattler, 1973), and may be considered homologous with the latter. . . . Carpel development is similar in Ruppia and Potamogeton. . . . Many who have made comparisons of vegetative appendages, mature flowers, and seeds 188 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 (Markgraf 1936; Uhl 1937; Singh 1965; Gamerro 1968) claim that Rup- pia is closely related to Potamogeton. | Floral development] does not con- tradict it, but it certainly does not bring the two genera closer together. . The floral development of R. maritima, though, does fit into a trend of simplification and reduction between Potamogeton and Zannichellia, as noted by Singh (1965). According to this trend, reduction occurs in the number of appendages per flower, flowers per inflorescence, and vege- tative appendages on the renewal shoots before floral induction.” The observations of Gamerro on the floral biology of Ruppia cirrhosa in Argentina are summarized below. Flowering begins by the opening of the leaf sheath which surrounds the developing inflorescence. This usually occurs in the afternoon and is followed by peduncle elongation, which is by cell enlargement alone, since no meristematic tissue is present. During the daylight hours of the first day, the peduncle usually elongates as much as 2 cm., and that night much greater elongation occurs (in most instances), pushing the inflorescence to the surface of the water. Usually very little emergence takes place, for the peduncle bends and lies horizontal just be- low the surface. In any case, there is a film of water surrounding the in- florescence. Occasionally elongation may occur on the second, instead of the first, night. Pollen release occurs during the morning following elongation. Open- ing of the anther locule is indicated by the appearance of bubbles, mostly oxygen, at a line of abscission of each anther half from the connective. The bubbles continue to enlarge and become confluent until abscission is complete. At this time, the bubble, about twice the size of the anther half, carries it to the surface of the water. As the anther half contacts the at- mosphere, it opens violently, spreading the pollen over the surface of the water, with clusters of grains adhering end to end, forming chains. The pollen grains are oriented with the concave cavity upward, and bubbles trapped in the cavity and reticulation of the exine keep the pollen chains afloat. Dehiscence of the anther half occurs by the curving of the walls downward and the edges inward. At the same time, the anther half re- mains afloat by the spreading over the surface of the water of an internal cutinized membrane. Following pollen release, growth of the peduncle resumes, but the in- florescence, which has now been reduced to two apocarpous gynoecia, re- mains just below the aqueous film surrounding the pollen grains. The apex of the inflorescence curves slowly so that the carpels of the lower flower are inclined forward, the gynoecia of both flowers then being in line with the axis of the inflorescence. Gamerro could not positively de- termine the time at which the stigmas are receptive to pollen, but it ap- pears that the flowers are probably protandrous. The straightening of the inflorescence is complete within 24 hours of pollen release. Within two or three days after pollen release, spiraling of the peduncle begins from its base upward. Spiraling is accomplished by the unequal enlargement of cells, the exterior ones enlarging more than the interior. As a result of the spiraling, the carpels are pulled deeper into the water 1978] HAYNES, POTAMOGETONACEAE 189 and are oriented toward the substrate. The spiraling process, which may occur with or without pollination, is completed within three or four days. Gynophore elongation, which occurs shortly after the spiraling process is complete, takes approximately seven to eight days to complete. This elongation, unlike spiraling, is dependent on fertilization. (Cf. FiGuRE 2, h.) The fruits are mature within 15 days after gynophore elongation. The entire process is complete within 30 days after the inflorescence emerges from the leaf sheath. Ortu, studying the influence of temperature and salinity on the ger- mination of seeds of Ruppia maritima, has observed that sodium chloride in concentrations of 52 grams/liter generally inhibits germination; that the dormant period decreases with increase of temperature; and that at 30° C. seeds kept in solutions of 3.5 grams/liter have a higher germina- tion rate, whereas at 20° C. a higher germination rate is observed among those kept in solutions of 6.5, 13, and 26 grams/liter. Economically, the genus is important only as a source of food for water- fowl and other wildlife. Gaevskaya lists 14 species that have been known to feed on Ruppia. REFERENCES: Under family references see especially ASCHERSON, ASCHERSON & GRAEBNER, CHRYSLER, MCATEE, OGDEN (1974), SINGH (1964, 1965), THIERET, and UHL. Aziz, K. ae In: Flora of West Pakistan. No. 80. 3 pp. 1975. BESSEY, C. E. Ruppia maritima L. in Nebraska. Am. Nat. 20: 1052, 1053. 1886, Bourn, W. S. Sea-water tolerance of Ruppia maritima L. Contr. Boyce Thomp- son Inst. - eeee 255.5 feed: 1935. Davis, J. S., B. TOMLINSON. A new species of ete in high salinity in Westen eae Jour. Arnold Arb. 55: 59-66. 1 Davy, J. B. Notes on Ruppia. Erythea 6: 18, 19. DELPINO, F., & P. ASCHERSON. Federico Delpino’s Eintheilung der Pflanzen nach dem Mechanismus der dichogamischen Befruchtung und Bemer- kungen iiber die Befruchtungs-vorginge bei Wasserpflanzen. Bot. Zeit. 29: 443-445, 447-459, 463-467. Essic, E. O. The Ruppia balls of Little Borax Lake. Sci. Mon. 66: 467-471. 1948 FERNALD, M. L., & K. M. WIEGAND. The genus Ruppia in eastern North Ameri- a. Rhodora 16: 119-127. 1914. GAMERRO, J. C. Observaciones sobre la biologia floral y a de la Po- tamogetonacea Ruppia cirrhosa (Petag.) Grande (= R. spiralis L. ex Dum.). (English summary.) Darwiniana 14: 575-608. 4 pls. 1968. Graves, A. H. The morphology of Ruppia maritima. Trans. Conn. Acad. Arts Sci. 14: 59-170. 15 pls. 1908 Hacstrom, J. O. Three species of Ruppia. Bot. Not. fe 137-144. 1911. HarrToo, C. DEN. De Nederlandse Ruppia-soorten. Gorteria 5: 148-153. 1971. HISINGER, E. Recherches sur les tubercules du Ruppia ei et du Zanni- chellia polycarpa, provoquées par le Tetramyxa parasitica. Medd. Soc. Faun. Fl. Fenn. 14: 53-62. 10 pls. 1887. 190 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 HorMEIsTEer, W. Zur Entwickelungsgeschichte der Zostera. Bot. Zeit. 10: 121- 131, 137-149. 7 pl. 1852. [Account of pollen, ovule, and embryo of Zos- tera and of Ruppia. | Lor- HELQUERAS, A. Estudios sobre fanerdgamas marinas en las cercanias de racruz, Ver. 66 pp. Universidad Nacional Autonoma de México. 1968. [Ruppia, 52.] MASON, . The species of Ruppia in New Zealand. New Zealand Jour. Bot. 5: 519-5 967. McCann, a Notes on the genus Ruppia. Jour. Bombay Nat. Hist. Soc. 45: 396— 402. 1945. McKay, E. M. Salt tolerance of Ruppia maritima in lakes of high magnesium sulphate content. Res. Stud. State Coll. Washington 3: 24-26. 1935. McMILLAN, C. Salt tolerance of mangroves and submerged aquatic plants. Pp. 379-390 in R. J. Rermotp & W. H. QUEEN, Ecology of halophytes. New York. 1974. Murseck, S, Uber die ae eele von Ruppia rostellata Koch. Sv. Vet.- akad. Handl. 36: 1-21. Ortu, A. M. Primary data on the germination of seeds of Ruppia maritima L. Novo Giorn. Ital. II. 103: 621. 1969. OSTENFELD, C. H. Ruppia anomala sp. nov., an aberrant phe of the Potamoge- tonaceae. Bull. Torrey Bot. Club. 42: '659- 602, fl. PHILLIPS, R. C. Extension of distribution of Ruppia el var. Be es (Schur.) Aschers. and Graebn. Florida Acad. Sci. Quart. Jour. 21: 185 186. 1958. ———. The ecology of marine plants of Crystal Bay, Florida. /bid. 23: 328- 337. 1960. Ponp, R. H. The morphology of Ruppia maritima —a criticism. Bot. Gaz. 48: 228, 229. 1909. PosLuszNy, U., & R. SAaTTLeR. Floral development of Ruppia maritima var. maritima. Canad, Jour. Bot. 52: 1607-1612. 1974. REESE, G. Systematik und Cytologie der Ruppia maritima. Ber. Deutsch. Bot. Ges. 75: 365. 1962. [R. maritima, 2n = 20 . Zur intragenerischen Taxonomie der Gattung Ruppia L., ein cyto- systematischer Beitrag. ee Bot. 50: 237-264. 1962. Uber die deutschen Ruppia- und Zannichellia-Kategorien und_ ihre Verbreitung in Schleswig-Holstein. Schr. Naturw. Ver. Schlesw.-Holst. VII. 34: 44-70. 1963. Roze, M. E. Recherches sur les Ruppia. Bull. Soc. Bot. France 41: 466-480. 1894. SCHWANITZz, G. Morphogenése des Ruppia pollen. Pollen Spores 9: 9-48. 1967. SERGUEEFF, M. Contribution a la morphologie et la biologie des Aponogetona- cées. Thése. Univ. Geneve. 132 pp. le SETCHELL, W. A. Ruppia and its environmental factors. Proc. Natl. Acad. Sci. U.S.A. 10: 286-288, 1924 . The genus Ruppia L. “Proc. Calif. Acad. Sci. IV. 25: 469-478. 1946. ST, JoHN, H., & F. R. Fosperc. A new variety of Ruppia maritima (Ruppia- ceae) from the tropical Pacific. Bishop Mus. Occ. Pap. 15: 175-178. fig. 7. 1939. STRAWN, K. Factors influencing the zonation of submerged monocotyledons at Cedar Key, Florida. Jour. Wildlife Managem. 25: 178-189. 1961. WALSH, G. E., & T. E. Grow. Composition of Thalassia testudinum and Rup- pia maritima. Florida Acad. Sci. Quart. Jour. 35: 97-108. 1973 [“1972”]. 1978] HAYNES, POTAMOGETONACEAE 191 VERHOEVEN, J. T. A. Ruppia-communities in the Camargue, France. Distri- bution and structure in relation to salinity and salinity fluctuations. Aquatic Bot. 1: 217-241. 1975. Woop, E. J. F. Some east Australian sea-grass communities. Proc. Linn. Soc. New S. Wales 84: 218-226. 1959. Wutrr, H. D. Karyologische Untersuchungen an der Halophytenflora Schleswig- * Holsteins, Jahrb. Wiss. Bot. 84: 812-840. 1937. ECOLOGY AND SYSTEMATICS SECTION UNIVERSITY, ALABAMA 35486 192 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 THE NODAL ANATOMY OF MYROTHAMNUS FLABELLIFOLIUS (MYROTHAMNACEAE): ANOTHER EXAMPLE OF A “SPLIT-LATERAL” CONDITION CHRISTIAN PUFF In 1970, Howard succEsTED the “split-lateral” or “common gap” con- dition as another type of nodal anatomy worthy of recognition. In his review he reported that members of the Asteraceae, Caprifoliaceae, Chloranthaceae, Gentianaceae, Gesneriaceae, Rhizophoraceae, Rubiaceae, and Zygophyllaceae have “split-laterals.” To his list I can add the Myro- thamnaceae, a monotypic family often associated with the Hamameli- dales, but also considered in connection with the Magnoliales. MORPHOLOGICAL DESCRIPTION The branches of the shrubby Myrothamnus flabellifolius show a clear differentiation into long and short shoots and bear decussately arranged leaves which exhibit a number of peculiarities.1. The following characters are essential for an understanding of the vascularization pattern of the plant. The “Unterblatter” (leaf bases) of each pair of opposite leaves are fused (gamophyllous) and form a sheath which remains on the shoo after the “Oberblatt” (leaf blade) has abscised (see Ficure 1, eG the dotted line between ‘“g” and “h” marks the position of the abscission zone). In short shoots, the “Unterblatter” remain rather small (ca. 1.5 mm.), are tightly packed, and overlap considerably —a single section through a short shoot may hit as many as three pairs of leaf bases (also see Puff, 1978, fig. 46). In long shoots (FicureE 1, insert) they appear much longer; a short distance from a node downward, however, they fuse with the stem, making a distinction between leaf and stem impossible (see FiGuRE 1, sections ac). They bear two kinds of appendages: four lateral ridges (F1rcurE 1, insert: R) which run down the whole length of the internode and terminate as “stipules” or stipule-like appendages (F1c- URE 1, insert: S) just below the abscission zone; and adaxially, between leaf base and stem, the “vaginal” or “median” lobes (Ficure 1, insert: V), which often envelop newly developed lateral shoots (F1cuRE 1, insert: L). VASCULARIZATION PATTERN For the sake of greater clarity, consecutive sections cut through a long shoot are presented in Ficures 1 and 2 ‘For a recent compilation of our knowledge of Myrothamnus, see Puff (1978). * The vascularization pattern of short shoots is not different, but is much more dif- ficult to mE eEDYet because of the short internodes and the overlapping, tightly packed “Unterblatter 1978] PUFF, MYROTHAMNUS FLABILLIFOLIUS 193 FIGURE 1. Myrothamnus flabellifolius. a-e: camera lucida drawings of sec- tions through a long shoot. Insert: portion of a long shoot; arrows indicate where sections a—e and f—h (FicurE 2) were cut. Black areas: sclerenchymatic tissue; hatched areas: xylem; dotted areas: phloem. 194 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Sections a—c show the characteristic four ridges that, starting at the abscission zone, run down the whole length of the internode: they are not vascularized. At the level of sections a, b, and c, no differentiation can be made between “leaf” and “stem” in spite of the presence of these ridges. Massive sclerenchyma strands run through the cortex in the me- dian: two somewhat smaller ones (each of them accompanied by a small band on either side) are in the area between two adjacent ridges (sec- tion a). At a somewhat higher level (section b), two lateral traces depart from the vascular cylinder, and slightly above (section c), the two median bundles branch out. At this point, the gaps of the laterals are closed again, and the sclerenchyma has formed a sickle-shaped sheath around them towards the outside, as well as a small strand between the bundles and hess stele. e further approach the node, the vascular cylinders supplying the ie ae (L) can be seen departing from the stele (section d; not cut exactly at a right angle to the stem). It should be mentioned that at this stage the gamophyllous ‘“Unterblatter” and the stem now have be- come two distinct units (sections d and e: dotted line). The leaf bases remain fused, although they also envelop the newly produced lateral shoots. The amount of sclerenchymatic tissue has increased in the stem: now there are two massive lateral strands plus four smaller ones. A little higher (section e), the four traces of the “Unterblatter’” can still be seen ascending in a vertical direction. There is, however, a first indication that the two lateral traces are beginning to split: the bundles which appeared as semicircles in sections c and d are starting to separate into two units. In section f (only fused leaf bases shown), finally, the two lateral bundles fork, with a portion moving toward each of the opposite median traces. Section g, cut slightly below the abscission layer, shows the lateral traces still passing through the leaf base toward the median bundle. They have forked again, and two small traces have entered the sides of the “vaginal lobe” (V). Slightly higher, section h cuts the base of the leaf blade (with the large median and the two cir- cular lateral bundles), and the “vaginal lobe” (with two small lateral traces) and “‘stipules” (no vascularization), neither of which is any longer connected with the blade Within the leaf blade, the traces split up further. Normally, however, only the median bundles fork once or twice, while the lateral traces often remain unbranched (also see Grundell, 1933, fig. 6). DISCUSSION The Myrothamnaceae so far are the ninth family of dicotyledons with “split-laterals.” It, however, appears to be the only family in which “split-laterals” are a family characteristic: preliminary investigations of herbarium material of Myrothamnus moschatus Baillon, the only other species of the genus (or of the family), yielded a vascularization pattern 1978 | PUFF, MYROTHAMNUS FLABILLIFOLIUS 195 g FIGURE 2. Myr Secacial flabellifolius: camera lucida drawings of sections Teen leaf bases. Black areas: sclerenchymatic tissue; hatched areas: xylem dotted areas: phloem. very similar to that of M. flabellifolius. More material, however, is needed to verify this. Myrothamnus flabellifolius, like the other plants known to have split- lateral traces, has opposite (decussate) leaves and “stipules” or stipule- like processes? (for exceptions, see Howard, 1970). Unusual, and so far not known in plants with split-lateral bundles, however, are the gamo- phyllous “Unterblitter” of M. flabellifolius, which completely envelop the stem (and the bases of newly produced lateral shoots). In M. flabel- lifolius, the lateral bundles fork within the leaf (i.e., ““Unterblatt”), and not, as in other plants with “split-laterals,” in tire: stem (cortex): only after ascending vertically in the leaf base for a distance do they fork, pass around about a quarter of the “Unterblatt,” and eventually join the median trace laterally (a “shift” of the “split-lateral” from the stem re- gion to the leaf). As ‘“‘stem” and “leaf,” however, must ph aug aene and morphologically be considered a continuous system (see How 1974), this peculiarity of M. flabellifolius appears to be of little signifi cance. PLANT MATERIAL EXAMINED Southwest Africa (Namibia): Ameib farm near Usakos, on sa to Phillips Cave, Puff 760714-1/2; Gamsberg pass, Puff 760717-1/1 and — South Africa. N. NATAL: Ngotoshe Dist., Ngome, near oot ae 760610- 5/1. TRANSVAAL: Bosmanskop area near Badplaas, Puff 770103-7/3; Wolk- berg area near Haenertsburg, Puff 770108-5/2. Swaziland: Mdzimba Hills near Mbabane, Puff 770119-1/1; Lebombo Mts., Blue Jay Ranch, near Mozambique border, Puff 770123-3/4. *In addition, it has “vaginal lobes.” The true morphological nature of both “stipules” and “vaginal lobes” has yet to be clarified. An ontogenetic study could ove to be useful: according to Weberling (1975), true stipules always develop early and distinctly proleptically, whereas other appendages (“sheath lobes” of Weberling) develop at a later stage. 196 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 All specimens are deposited at the herbarium of the Institute of Botany, University of Vienna (wv). Sections were cut freehand with a razorblade and stained in phloro- glucinol and hydrochloric acid. LITERATURE CITED GRUNDELL, R. 1933. an cee von Myrothamnus flabellifolia Welw. Symb. Bot. Upsal, 1(2): Howarp, R. A. 1970. ise observations on the nodes of woody plants with special reference to the problem of the “‘split-lateral” versus the ‘‘common gap.” Pp. 195-214 in N. K. B. Rogson, D. F. CuTLer, and M. near, eds., New research in plant anatomy. _Jour Linn. Soc. Bot. 63, Suppl. Academic Press, London and New Yo . 1974. The stem-node-leaf ae of the Dicotyledoneae. Jour. Arnold Arb, 55; 125-181. Purr, C. 1978. Zur Biologie von Myrothamnus flabellifolius Welw. (Myro- thamnaceae). Dinteria (in press). WEBERLING, F. 1975. Uber die Beziehungen zwischen Scheidenlappen und Stipeln. Bot. Jahrb. Syst. 96: 471-496. INSTITUTE OF BOTANY PRESENT ADDRESS: UNIVERSITY OF VIENNA Dept. oF BOTA RENNWEG 14 eee OF THE WITWATERSRAND A-1030 VIENNA, AUSTRIA JAN SMU JOHANNESBURG 2001, SOUTH AFRICA Preliminary Announcement Thirteenth International Botanical Congress Sydney, Australia. 21-28th August, 1981 The Programme will consist of 12 sections — molecular, metabolic, cellu- lar and structural, developmental, environmental, community, genetic, systematic and evolutionary, fungal, aquatic, historical, and applied botany. There will be plenary sessions, symposia, and sessions for sub- mitted contributions (papers and posters). Chairman of the Programme Committee:—Dr. L. T. Evans. Field Trips will include visits to arid and semi-arid regions, eucalypt forest, rain forest, heath, coastal vegetation (e.g. Great Barrier Reef, mangroves) etc., and specialist trips. Chairman of the Field Trips Com- mittee:—Prof. L. D. Pryor. First Circular, containing details, will be mailed in August, 1979. Send your name and full address, preferably on a postcard, to ensure your in- clusion on the mailing list. - Enquiries should be sent to the Executive Secretary, Dr. W. J. Cram. Congress address — 13th I.B.C., University of Sydney, N.S.W. 2006, Australia. Sponsored by the Australian Academy of Science. Journal of the Arnold Arboretum April, 1978 CONTENTS OF VOLUME 59, NUMBER 2 Form of the Perforation Plates in the Wide Vessels of Ne a in Palms. Larry H;-Ktorz. . EGG 2 ee eee. eas Generic Limits in the Xeroteae (Liliaceae sensu lato). P. F. STEVENS 129 Rhizophora in Australasia — Some Clarification of Taxonomy and Distribution. POR SIOMEINSOM cota Rn ieee a he The Potamogetonaceae in the Southeastern United States. ROBERT RG WHAYNES: 2. c2- ce fo A ere TO The Nodal Anatomy of Myrothamnus flabellifolius (Myrothamnaceae) : Another Example of a Re Condition. CHRISTIAN PUFF . item nee epee cfd LOD. ain = Sore Se Volume 59, Number 1, including pages 1-104, was issued January 24, 1978. JOURNAL oF the ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 59 NUMBER 3 US ISSN 0004-2625 Journal of the Arnold Arboretum Published quarterly in January, April, July, and October by the Arnold Arboretum, Harvard University. 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 1-45, reprinted, and some back numbers of volumes 46-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, U.S.A. EDITORIAL COMMITTEE B. G. Schubert, Chairman S. A. Spongberg P. F, Stevens C. E. Wood, Jr. ASSISTANT EDITOR E. B. Schmidt Printed at the Harvard University Printing Office, Boston, Massachusetts COVER: The nba representation of fruits of various Leguminosae, form- ing the design of our new cover for Volume 59, is based on material from plants growing in ‘ha Arnold Arboretum of ‘Harvard University. The design was planned and executed by Karen S. Velmure, who has also drawn several of the preceding covers as well as various devices used in the Journal of the Arnold Arboretum and on the offprint covers. The three subfamilies a gs Leguminosae are represented by the fruits shown in the design. The genera are, top row (left to siyaeee Albizia (Mimosoideae), Gleditsia CCassaipinicidens), Caragana (Faboideae), Gymnocladus (Caesal- pinioideae), and lower row, right, Colas (Faboidea e). The family Leguminosae, the second most important family of flowering plants, is to be the subject of an International Legume Conference at the Royal Botanic Gardens, Kew, Richmond, Surrey, England, in July, 1978. This cov- er is our salute to the complete success of that Conference Second-class postage paid at Boston, Massachusetts JOURNAL OF THE ARNOLD ARBORETUM vo.s9)”™~”:C*S NS SS*~*~*~*«(N ER THE GENERA OF CRASSULACEAE IN THE SOUTHEASTERN UNITED STATES STEPHEN A. SPONGBERG CRASSULACEAE A. P. de Candolle in Lamarck & De Candolle, FI. France. ed. 3. 4(1): 382. 1805, nom. cons. (STONECROP or ORPINE FAMILY) Annual, biennial, or [monocarpic or] polycarpic perennials, the plants mostly succulent, often evergreen, herbs, subshrubs [or shrubs, many of treelike habit, or rarely scandent], usually terrestrial from fibrous roots, sometimes developing fleshy or woody caudices or rhizomes [or rarely epi- phytic]. Leaves exstipulate, alternate, opposite, or whorled, the internodes sometimes reduced and the plants rosulate, the blades simple or occa- sionally pinnately compound, usually fleshy and succulent, cylindrical to flattened with entire to crenate or dentate margins, the leaves of vegeta- tive and flowering shoots often dissimilar; vegetative reproduction com- monly by bulbil or plantlet production in the crenations of the leaves or by meristematic activity in detached leaves or stems. Flowers perfect or ‘Prepared for a generic flora of the southeastern United States, a project of the arv series (Jour. Arnold Arb. 39: 296-346. 1958). The area covered includes North and South Carolina, res ia, Florida, Tennessee, Alabama, Mississippi, Arkansas, and numerous problems. I should also like to thank E. B. Schmidt for her careful editing. Figures 1 and 3 were prepared by Karen S. Velmure; Figure 2 was drawn by Rachel A. Wheeler. Living and alcohol-preserved plants used for the illustrations and for er were kindly supplied by R. C. Rollins, R. D. Thomas, R. E. Weaver, Jr., L. Wilbur, K. A. Wilson, and C. E. Wood, Jr. x President and Fellows of Harvard College, 1978. 198 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 imperfect (the plants then dioecious or polygamodioecious), actinomorphic, usually proterandrous, arranged in terminal or axillary, usually bracteate corymbose, helicoid, or scorpioid cymes, panicles, simple to compound dichasia, or solitary [or clustered] in the axils of leaves for rarely in spikes]; insertion of the sepals and petals hypogynous. Sepals (3) 4 or 5 [or up to 30], free or connate and then forming a calyx tube. Petals the same number as the sepals, free or occasionally connate, forming a tubular to urceolate, usually ribbed corolla tube. Stamens as many or twice as many (and then obdiplostemonous) as the sepals, insertion hypogynous or epipetalous, free [or occasionally the filaments basally connate]; anthers basifixed, dehiscing introrsely by longitudinal slits, 4-sporangiate (2-locu- late at anthesis), with a secretory tapetum; pollen 2-celled when shed (as monads), 3-colporate. Gynoecium apocarpous or basally syncarpous, the carpels usually the same number as the sepals, sessile or stipitate, tapering gradually or abruptly into erect or divergent styles, the styles terminated by small, often poorly differentiated apical or subapical stigmas; each carpel subtended by a scalelike [rarely petaloid] nectariferous gland [or rarely, the gland absent]; ovary 1-loculate, superior, with numerous [rarely few to solitary] anatropous, crassinucellar, bitegmic ovules on a parietal or marginal placenta. Fruit a follicle or follicetum of several basally connate follicles, the follicles dehiscent along their adaxial sutures or rarely by abaxial flaplike valves; seeds usually numerous, small, rarely winged; embryo small, straight, embedded in scant, fleshy, cellular endosperm [or endosperm occasionally lacking]; megagametophyte of the Polygonum or (rarely) Allium type; embryogeny of the Caryophyllad type. (Sedaceae Necker, nom. subnud.) Tyre Genus: Crassula L. Upward of 1500 species in as many as thirty-five genera and six generally recognizable subfamilies, distributed throughout the world, primarily in areas of subtropical climate with alternating wet and dry seasons. Great concentrations of species and genera occur in southern Africa, Madagascar, the Canary Islands, southern Asia, Mexico, and the Mediterranean region. The family is poorly represented in Australia (one genus), Oceania (two genera) and South America. Nine or ten species in three genera (repre- senting two subfamilies) are indigenous in the southeastern United States, while eleven or twelve additional species (eight of which represent a fourth genus and a third subfamily) have escaped from cultivation and have be- come naturalized in this area. While the family Crassulaceae has generally been conceded to form a natural, well-defined taxon, agreement with regard to the definition of taxa at essentially all subordinate levels has been fraught with difficulties, and there is no general consensus as to the number of species, genera, or sub- families. Berger (1930), who treated the entire family, recognized six subfamilies that, except for inevitable intermediate taxa, are fairly well defined morphologically, if not biologically. Baldwin (1939) suggested that the relationships within the family are interlinear as well as linear, and Berger (1930, p. 382) cautioned about the use of technical characters 1978 | SPONGBERG, CRASSULACEAE 199 in the definition of taxa. Berger’s sentiments were reiterated by Moran (1942b), who said that “if too much emphasis were placed on technical characters, the numerous exceptions and intergradations would necessitate the combination of genera until but six or only one genus remained.” The characters that have been used to define subfamilies are largely concerned with the number of floral parts and the connation of the petals, as well as the position of the flowering shoots and the arrangement of the leaves. Certainly some of the problems below the level of subfamily are due to the usual succulence of the plants and the poor specimens that result from attempts to press the plants by conventional herbarium tech- niques. It is almost imperative that studies of the Crassulaceae be made with living plants or at least with fluid-preserved materials. Members of the family are generally succulent and usually show a whole series of xeromorphic adaptations (including anatomical and physiological changes), as well as adaptations in the life cycles of some species (e.g., Diamor pha Smailii) for drought evasion. Most species have an abundance of water storage tissue in the leaves (in which a palisade layer is rarely or poorly differentiated) and in the stems (where the parenchymatous or collenchymatous tissues of the cortex and pith are well developed). Vascu- lar tissue is unevenly and irregularly distributed throughout the storage tissues, the phloem is poorly developed, and the xylem of the stems is generally in a continuous cylinder. While trichomes are uncommon in the family, the hairs (as well as the adventitious roots and epidermal cells) of some species are suspected of absorbing water directly from the atmosphere. A waxy coating secreted by the epidermis is characteristic of the leaves, and hydathodes are relatively common, often being distributed over both surfaces of the leaves and visible to the naked eye as small pits. The stomata, which are usually surrounded by three subsidiary cells, are likewise usually well distributed over both surfaces of the leaves. The stomata of some investigated species are open primarily at night and closed during the day, a phenomenon that coincides with carbohydrate decrease, excessive increase in malic acid content, and water-vapor loss at night, and with starch accumulation and reduced acidity during the day- light hours. Carbon fixation occurs in the leaves at night when the stomata are open, and is thus temporally separated from the light reactions in the photosynthetic process. During the day, when the stomata are closed and water loss is thus reduced, the fixed carbon is reduced to carbohydrate. This physiological syndrome, known as crassulacean acid metabolism (CAM), along with the Calvin-Benson cycle (C3 plants) and the Cy, decarboxylic pathway (C, plants), is one of three distinct carbon fixation and reduction mechanisms known in vascular plants. While the CAM pathway was originally described from crassulaceous plants, it is not re- stricted to them, and it may be facultative, depending upon environmental conditions. Kluge (1977) has found that under normal conditions in nature, Sedum acre is a Cz plant, but that under experimental water stress, carbon dioxide exchange shifts from the Calvin cycle to CAM. As 200 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 indicated by Black & Williams (1976), all CAM plants are succulents, but not all succulents are CAM plants. Apparently there is little if any taxo- nomic significance to CAM, since the pathway has been reported for members of four monocotyledonous and thirteen dicotyledonous families, as well as for a fern, Drymoglossum piloselloides. While only the outlines of the process are given above, there has been much research on CAM: for further information, see the papers listed under family references, as well as a few titles included under Kalanchoé. Other features of the Crassulaceae of biological interest are the wide range of chromosome numbers (from 27 = 8 to 2n = ca. 500) and the oc- currence of considerable and widespread intraspecific variation in chromo- some number, including both polyploidy and aneuploidy. Uhl (1970) reports that two species of Graptopetalum Rose have gametic chromosome num- bers of m = 245 + 5 and» = 270 + 5, higher counts than have been reported for any other seed Sint The occurrence of heteroploidy has proved to be confusing, particularly with regard to species delimitation, and especially when few if any morphological characters correlate wit the different chromosome numbers Uhl (1976) has determined that the majority of the Crassulaceae of Mexico (subfams. Sedoideae and Echeverioideae) “. . . belong to a giant biosystematic comparium, since they have been afeiaannee ied directly or indirectly, by artificial hybrids. More than 1100 different hybrids have been produced among at least 165 species of the genera Echeveria, Sedum, Pachyphytum, Graptopetalum, Thompsonella, Villadia, and Lenophvilum ” Moreover, the intersubfamilial hybrid between Echeveria linguifo- ia Tema and Sedum cremnophila Clausen, both with 27 — 66, has nor- mal meiosis and 95 percent normal pollen (Uhl, 1966, 1976). It would appear that continued chromosome studies, together with hybridization programs, in other groups of the family will alter the concept of relation- ships at almost all ranks and will undoubtedly provide the basis for a more acceptable (if not a more utilitarian) taxonomy. n current systems of classification, the Crassulaceae are closely allied with the Saxifragaceae, from which they can be distinguished by their usual succulence and their flowers with the number of petals or sepals and carpels the same, the carpels subtended by nectariferous glandular scales, and the seeds lacking (or with only scant) endosperm. Penthorum L. is the only genus that has created problems in the delimitation of the Crassula- ceae, and Quimby (1971) has concluded that this family is a natural as- semblage only if Penthorum is excluded. Although Penthorum exhibits characters that are in many respects intermediate between the two families (it has occasionally been placed in its own monogeneric family, Pen- thoraceae Rydb. ex Britton), in this flora Penthorum has been included in the Saxifragaceae in subfam. Penthoroideae Engler, for the reasons outlined there (Spongberg, 1972, p. 421). Thorne (1968) and Cronquist (1968) include Crassulaceae and Saxi- fragaceae in the Rosales, while Takhtajan (1969) includes the two families in the Saxifragales. Both Takhtajan and Cronquist note that the Podo- 1978 | SPONGBERG, CRASSULACEAE 201 stemales (Podostemaceae) (included in the Rosales by Thorne) are un- doubtedly derived from the Crassulaceae. Support for a close relationship between the Crassulaceae and the Podostemaceae, seemingly dissociative taxa, comes primarily from embryological similarities noted by Mauritzon (1939), Maheshwari (1945), and Subramanyam (1962) that have been summarized in this flora by Graham & Wood (1975). An evolutionary trend toward adaptation to an aquatic habitat is also evident in the Crassulaceae in some members of Crassula sect. TILLAEOIDEAE. None of the Crassulaceae is of great economic importance, although many of the species are cultivated as ornamentals, both as pot plants and as garden perennials. In our area, species of Sedum and Sempervivum (subfam. Sempervivoideae Berger) are commonly grown in rock gardens and on walls. Some of the species have been used in folk medicine and as greens in salads. REFERENCES: Backer, C. A. Crassulaceae. Fl. Malesiana I. 4(3): 197-202. 1951. [Sedum and Kalanchoé. | BAEHNI, C. Valladia et Altamiranoa. Etude sur la fusion de deux genres de Crassulacées. Candollea 7: 283-286. 1937. BaILey, L. H., E. Z. Bees & STAFF OF THE L. H. Bartey Hortorrum. Hortus third. xiv + 1290 pp. New York. 1976. [Crassulaceae, 329; Crassula, 327-329: Kalanchoé, 620-623: Sedum, 1023-1030. BAILLon, H. Crassulaceae. Hist. PI. 3: 305-324. 1871. [English ed., 3: 304- S92. iy. Batpwin, J. T., Jr. Certain cytophyletic relations of the Crassulaceae. Chron. Bot. 5: 415-417. Cytophyletic analysis of certain annual and biennial Crassulaceae. tearcae. 5: 184-192. 1940. Beevers. H. Crassulacean acid metabolism: an assessment. Pp. 405, 406 i R. H. Burris & C. C. Brack, eds., CO, metabolism and plant productivity. xi + 431 pp. Baltimore. 1976. BentTHAM, G., & J. D. Hooker. Crassulaceae. Gen. Pl. 1: 656-661. 1865. Bercer, A. Crassulaceae. Im: A. ENGLER & K. PrantTL, Nat. Pflanzenfam. ed. 2. 18a: 352-483. 1930. ByOrKMAN, O. Comparative studies on photosynthesis in higher plants. Pp. 1-63 in A. C. Giese, ed., Photophysiology — current topics in Ere eee and photochemistry. Vol. VIII. xviii + 269 pp. New York. Back, C. C., & S. Wittiams. Plants exhibiting ee, common to cras- sulacean acid metabolism. Pp. 407-424 in R. H. Burris & C. C. BLACK, eds., CO. metabolism and plant productivity. xi + 431 pp. Baltimore. 1976. [See also Ann. Rev. Pl. Physiol. 24: 253-286. 1973.] Bortsov, A. G. Crassulaceae. Jn: V. L. Komarov, ed., Fl. USSR. English ed. 9: 8-105. 1971 BRAMWELL, D. Generic delimitation in the Sempervivum group —a numerical approach. Natl. Cactus Succul. Jour. 25: 50, 51. aon: N. L., & J. N. Rose. New or noteworthy American Crassulaceae. Bu IL N. Y. Bot. Gard. 3: 1-45. 1903 ———— . Lenophyllum, a new genus of Crassulaceae. Smithson. Misc. Coll. 47: 159-162. pl. 20. 1904. 202 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 —— & . Crassulaceae. N. Am. Fl. 22: 7-74. 1905. CANDOLLE, A. P. bE. Mémoire sur la famille des Crassulacées. 47 pp. pls. 1-13. fold-out toble. Paris. 1828. [Mémoire II of De Candolle’s Collection de mémoires ... .| Crassulaceae. Prodr. 3: 381-414. 1828. CLAUSEN, R. T. Studies in the Crassulaceae: Villadia, Altamiranoa and Thomp- sonella. Bull. Torrey Bot. Club 67: 195-198. 1940. Nomenclatural changes and innovations in the Crassulaceae. Cactus Succul. Jour. 18: 58-61, 74-77. 1946. ,R. V. Moran, & C. H. Unt. The taxonomy and cytology of Hasseanthus. Desert Pl. Life 17: 69-83. 1945, — As cnet acid metabolism. Natl. Cactus Succul. Jour. 30: Ze os | Summary a oe acid metabolism, including families in ae it is a to o Crete, P. Embryogénie des Ge aneeed Développement de l’embryon chez le Cotyledon umbilicus L. Compt. Rend. Acad. Sci. Paris 222: 1311-1313. 1946. Embryogénie des Crassulacées. Développement de l’albumen et formation des haustoriums chez le ear umbilicus L. Ibid. 1454, 1455, Crews, C. E., S. L. Witiiams, H. M. Vines, & C. C. Back. Changes in the metabolism and physiology of crassa acid metabolism plants dil in controlled environments. Pp. 235-250 in R. H. Burris & C. C. eds., CO, metabolism and plant sna Ae xi + 431 pp. ee 1976. Cronquist, A. The evolution and classification of flowering plants. x + 396 pp. Boston. 1968. Davis, G. L. Systematic embryology of the angiosperms. viii + 528 pp. New York. 1966. [| Crassulaceae, 96, 97.] D’Husert, E. Recherches sur le sac ae des plantes grasses. Ann. Sci. Nat. Bot. VIII. 2: 37-128. pls. 1-3. 1895. [Crassulaceae, including Crassula, Echeveria, Sedum, and Se aa 99-107. Domin, K. Goavseruruani L., subgenus Jovisbarba Koch. Bull. Int. Acad. Sci. Bohéme. 1932: 1-9. pls. 1-9. 1932.* ERptMAN, G. Pollen morphology and plant taxonomy. Angiosperms. xii + 539 pp. frontisp. Stockholm; Waltham, Mass. 1952. [Corrected reprint and new addendum. xiv + 552 pp. New York. 1966. | EvANARI, M. On some types of succulent plants. Chron. Bot. 6: 314, 315. 1941, Evans, H. The physiology of succulent plants. Biol. Rev. Biol. Proc. Cam bridge Phil. Soc. 7: 181-211. 1932. [Includes a review of early ee of crassulacean acid metabolism. | GrauaM, S. A., & C. E. Woop, Jr. The Pais in the southeastern United States. Jour. Arnold Arb. 56: 456-465. Huser, H. Familie Crassulaceae. Jn: G. Hect, Illus. "Fl. Mittel-Europa. ed. 2. 4(2A): 62-125. pls. 140, 141. 1961-1966. Hurcuinson, J. The families of flowering plants. ed. 3. xviii + 968 pp. Oxford. 1973. [Crassulaceae, 565, 566. Jacopsen, H. A handbook of succulent plants: descriptions, ae ae and cultural details for succulents other than Cactaceae. English ed. Vol. I. Abromeitiella to Euphorbia. xiii + pp. 1-489. Vol. Il. Ficus to Zygophyllum. Pp. 490-912. London. 1960. [A third volume is devoted to Mesembryan- themum (Aizoaceae). | 1978] SPONGBERG, CRASSULACEAE 203 _ Das Sukkulentenlexikon. 589 pp. 200 pls. Jena. 1970. James, W. O. Succulent plants. Endeavour 17: 90-95. 1958 JensEN, L. C. W. Primary stem vascular patterns in three subfamilies of the Crassulaceae. Am. Jour. Bot. 55: 553-563. 1968. [ Cotyledonoideae, Kalanchoideae, and Sedoideae. | Kiuce, M. Models of CAM regulation. Pp. 205-216 m R. H. Burris & C. C. BLACK, eds., CO. metabolism and plant productivity. xi + 431 pp. Balti- more. 1976. Is Sedum acre L. a CAM plant? Oecologia 29: 77-83. 1977. Knut, P. Handbook of flower pollination. (English transl. J. R. A. Davis.) Vol. 2. viii + 703 pp. Oxford. 1908. [Crassulaceae, 422— 430. | Kuntze, O. Crassulaceae. Rev. Gen. Pl. 3(2): 82-86. 1898. Lems, K. Botanical notes on the Canary Islands. II. The evolution of plant forms in the islands: Aeonium. Ecology 41: 1-17. 1960. Lone, R. W., & O. Laxera. A flora of tropical Florida. xvii + 962 pp. Coral Gables. 1971. Manesuwart, P. The place of angiosperm embryology in research and teach- ing. Jour. Indian Bot. Soc. 24: 25-41. 1945 Mavrirzon, J. Studien iiber die Embryologie der Familien Crassulaceae und Saxifragaceae. 152 pp. Lund. 1933. - Contributions to the embryology of the orders Rosales and Myrtales. Lunds Univ. Arsskr. II. Sect. 2. 35(2): 1-121. 1939. [Crassulaceae and Podostemaceae, 37, 38; see also Mauritzon’s earlier papers, Bot. Not. 1930: 233-250. 1930, and 1933: 172-180. 1933. Mercatre, C. R., & L. CHatk. Anatomy of the dicotyledons. Vol. I. Ixiv + 724 pp. Oxford. 1950. [Crassulaceae, 578-581. Moran, R. The Crassulaceae of Yosemite National Park. Desert Pl. Life 14: 5-9, 1942a. [Includes discussion of the generic status of Rhodzo la. | _ Delimitation of genera and subfamilies in the Crassulaceae with special reference to the Echeverioideae. /bid. 125-128. 1942b. _ The status of Dudleya and Stylophyllum. Ibid. 149-157. 1942c. _ The generic status of Hasseanthus. Ibid. 164-169. 1942d. __ A revision of Dudleya, subgenus Stylophyllum—I. Ibid. 190-193. 19435 Il. Ibid. 15: 9-14. 1943; III. Ibid. 24-28; IV. Ibid. 40-45; V. Ibid. : ae on Hasseanthus —1. Ibid. 22: 76-82. 1950. —_. _Hasseanthus, a subgenus of Dudleya. Leafl. West. Bot. 72-1107, 1953: Villadia Nelsonii Rose. Cactus Succul. Jour. 43: 84-87. 1971. [Dis- cusses generic relationships. | The genus Meterostachys Nakai (Crassulaceae). Ibid. 44: 262-273. 1972. Morr, A. Saggio monografico sulla struttura istologica delle Crassulacee. Nuovo Giorn. Bot. Ital. 11: 161-187. pls. 5-7. 1879 Newton, L. E. A bibliography of key works for the identification of succulent lants. Cactus Succul. Jour. Great Brit. 29: 67-73. Nisuipa, K. Studies on stomatal movement of crassulacean plants in relation to the acid metabolism. Pl. Physiol. 16: 281-298. 1963. Osmonp, C. B. CO, assimilation and dissimilation in i light and dark in CAM plants. Pp. 217-233 in R. H. Burris & C. C. BLa K, eds., CO, metabolism and plant productivity. xi + 431 pp. Baltimore. ioe 204 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 PoetLNitz, K. von. Zur Kenntnis der Gattung Echeveria DC. Rep. Sp. Nov. 39: 193-270. 1936. PRAEGER, R. L. On some doubtful species . the African section of the Semper- vivuNL group. Proc. Roy. Irish Acad. B. 38: 1-24. yee iva of the Canary rahe area. /bid. 454- 499. 1929, An account of the Sempervivum group. 265 pp. Roy. Hort. Soc. London. on [Treats Sempervivum sensu stricto, Aichryson Webb & Berth., Aeonium Webb & Berth., Greenovia Webb, and Monanthes Haw PRoNER, M. Studja nad idioblastami u Crassulaceae. Wasszawa. 1934 QuiuBy, M. W. The floral morphology of the Crassulaceae. 42 pp. 11 pls. University, Mississippi. 1971. [Published by the author: Ph.D. thesis, Cornell University, 1939 E Raprorp, A. E., H. E. Antes, & C. R. BELL. Manual of the vascular flora of the Carolinas. lxi + 1183 pp. Chapel Hill. 1968. [Crassulaceae, 513-— 516. ] Ranson, 5. L., & M. Tuomas. Crassulacean acid metabolism. Ann. Rey. Pl. Physiol. 11: 81-110. 1960. [Includes extensive bibliography. | RickeTT, H. W. Wild flowers of the United States. The southeastern states. Vol. 2, part 1. x + 322 pp. New York. 1967(2). [Not dated; Cras- sulaceae, 256-260, pls. 91-94. | Rocen, T. Beitrag zur Embryologie der Crassulaceen. Sv. Bot. Tidskr. 22: 368— 376. 1928. [ Penthorum in relation to the Crassulaceae. | Rovnanl, I. Pathways of carbon metabolism in spongy mesophyll cells, isolated rom Sedum telephium leaves, and their relationship to crassulacean acid metabolism plants. 169 pp. Ph.D. thesis, Univ. Georgia. Athens. 1972.* H. M. Vines, & C. C. Brack, Jr. Isolation of mesophyll cells from Sota telephium leaves. Pl. Physiol. 51: 97-103. 1973. [Crassulacean acid metabolism. | Row ey, G. D., & L. E. Newron, eds. Repertorium plantarum succulentarum XXII — 1971. Reg. Veg. 87: 1-24. 1973. [Published under the auspices of the International Organization for Succulent Plant Study. An annual publication since 1950, this is the most recent number, It lists new oo names of succulent plants published during the past year, and the past two numbers have also included bibliographies of the current itera of succulent plants. Beginning with Vol. XXIII, this series is o be published by Abbey Garden Press. | eer ea 5, Crassulaceae. In: A. ENGLER & K. Prant_, Nat. Pflanzenfam. ed. 1. 3(2a): 23-38. & E. G. Baker. Some South African species of Cotyledon. Jour. Bot. 40: 9-23, 80- 94. pls. 431-435. 1902 SCHULZE-MENz, G. K. Rosales. Zn H. MELCHIOR, ed., Engler’s a der Pflanzenfamilien. ed. 12. 2: 193-242. 1964. [Crassulaceae, 199, 200. ] SCHUTTE, K. H., R. Steyn, & M. vAN DER WESTHUIZEN. So acid metabolism in South African succulents: a preliminary investigation into its occurrence in various families. Jour. S. Afr. Bot. 33: 107-110. 1967. SHARMA, A. K., & S. GHosH. Cytotaxonomy of Crassulaceae. Biol. Zentralbl, 86(Suppl.): 313-336. 1967. SHARSMITH, H. K. The genus Sedella. Madrofio 3: 240-248. 1936. [See also, Further notes on the genus Ss i 5: 192-196. 1940. Sedella Britton & Rose = Parvisedum R. T. Clau SMALL, ie K. Manual of the a flora. xxii + 1554 pp. New York. 19 ores 585-589. 1978 | SPONGBERG, CRASSULACEAE 205 SOLEREDER, H. Systematische Anatomie der Dicotyledonen. xi + iv + 984 pp. Stuttgart. 1899. [Crassulaceae, 362-366. | Sources. R. Les relations embryogéniques des Crassulacées, Saxifragacées et Hypericacées. Bull. Soc. Bot. France 83: 317-329. 1936. SponcBERG, S. A. The genera of Saxifragaceae in the southeastern United States. Jour. Arnold Arb. 53: 409-498, 1972 Stoupt. H. N. Gemmipary in Byrnesia weinbergii. Am. Jour. Bot. 21: 562-572. 1934. [Byrnesia Rose = Graptopetalum Rose; comparisons with Kalan- choé pinnata. | SUBRAMANYAM, K. Embryology in relation to systematic botany with particular reference to the Crassulaceae. Pp. 94-112 in Plant embryology: a sym- posium, Council of Scientific and Industrial Research, New Delhi. 1962. TAKHTAJAN, A. Flowering plants: origin and dispersal. (Transl. C. JEFFREY. ) x + 310 pp. Edinburgh. 1969 TuorNE, R. F. Synopsis of a putatively phylogenetic classification of the flower- ing plants. Aliso 6: 57-66. Tinc, I. P. Crassulacean acid metabolism in natural ees in relation to annual CO. uptake patterns and water utilization. Pp. 251-268 in R Burris & C. C. ca ck, eds., CO, metabolism and ee productivity. xi + 431 pp. Baltimore. 1976. Uu1, C. H. Cane studies in the subfamilies Crassuloideae, Kalanchoi- deae and Cotyledonoideae of the Crassulaceae. Am. Jour. Bot. 35: 695-706. 1948. The Crassulaceae and cytotaxonomy. Cactus Succul. Jour. 28: 28-30. » 1956. The chromosomes of the Sempervivoideae (Crassulaceae). Am. Jour. Bot. 48: 114-123. 1961a. _ Some cytotaxonomic problems in the Crassulaceae. Evolution 15: 375- 377. 1961b. _ Chromosomes and phylogeny of the Crassulaceae. Cactus Succul. Jour. 35: 80-84. 1963. _ Chromosomes of artificial hybrids of Mexican Crassulaceae. ( Abstr.) Am. Jour. Bot. 53: 617. 6 ————. The comparium in the Mexican Crassulaceae. (Abstr.) Abstr. 11th Int. Bot. Congr. 223. 1969. [‘‘Artificial hybrids among the Mexican Cras- sulaceae show that more than 100 species from seven genera naa to the same comparium.” Chromosome numbers range from 7 = 12 to = 244.| _ Chromosomes of Graptopetalum and Thompsonella ie. Am. Jour. Bot. 57: 1115-1121. 1970. [Graptopetalum, x = 30-35; Thomp- sonella, x = 26.| Poceisidy in the American Crassulaceae. (Abstr.) Brittonia 24: 129. _ Chromosomes, hybrids and ploidy of Sedum cremnophila and Echeveria linguifolia (Crassulaceae). Am. Jour. Bot. 63: 806-820. 1976 _ Moran. The cytotaxonomy of Dudleya and Hasseanthus. Am. Jour. Bot. 40: 492-502. 1953. & _ The chromosomes of Pachyphytum (Crassulaceae). [bid. 60: 648-656. 1973. Upuor, J. C. T. Dictionary of economic plants. ed. 2. 591 pp. Lehre. 1968. Vickery, H. B. A chemist among plants. Ann. Rev. Pl. Physiol. 23: 1-28. 1972. [Crassulacean acid metabolism, 16-18. 206 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 WALTHER, E. Alwin Berger’s ‘“Crassulaceae.” Cactus Succul. Jour. 2: 383-385, 408, 409. 1931, Cotyledon, Echeveria, or Dudleya? Leafl. West. Bot. 1: 27-29. 1932. [Includes a key to the three genera. | . Phylogeny of Echeveria. Cactus Succul. Jour. 8: 82-88. 1936. Notes on Crassulaceae: new combinations in two genera. Ibid. 10: 22-24, 1938. [Villadia and Altamiranoa; cf. R. T. CLAUSEN, Bull. Torrey Bot. Club 67: 195-198. 1940.] . Echeveria. frontisp. ix + 426 pp. San Francisco. 1972. [A monograph of the genus with numerous drawings and photographs, some in color, Wicus, J. C. A dictionary of the flowering plants and ferns. ed. 7, revised by _K. Ary SHAW. xxii + 1214 pp. + lili. Cambridge. 1966. Wotr, J. Der diurnale Saéurerhythmus. Pp. 809-889 in W. RUHLAND, ed., Handbuch der Pflanzenphysiologie. Vol. 12, part 2. xix + 1421 pp. Berlin. 1960. [Crassulacean acid metabolism, including an historical review and a comprehensive bibliography. | KEY TO THE GENERA OF CRASSULACEAE IN THE SOUTHEASTERN UNITED STATES We Petals (3) 4 or 5, free or only slightly connate at base; stamens free from the petals and inserted hypogynously, B. Stamens 2X the number of sepals and petals (subfam. Sedoideae). C. Gynoecium apocarpous; flowers perfect or imperfect, 4- or 5- merous; mature follicles dehiscing longitudinally; annuals, bien- nials, or perennials... ese eh ae eae eae 1. Sedum. C. Gynoecium syncarpous at base; flowers perfect, 4-merous; mature follicles dehiscing by a tear-shaped, flaplike, abaxial valve; winter PUTAS Stag. Se Ang ay 8 af den be hay wed aee och: 2. Diamor pha. B. Stamens the same number, (3) 4, as the sepals and petals; gynoecium apocarpous; mature follicles dehiscing longitudinally along the adaxial suture; annuals or biennials (subfam. Crassuloideae). .... 3. Crassula. Petals 4, connate; stamens 8 (in 2 whorls of 4), epipetalous; gynoecium basally syncarpous; perennials naturalized in southern Florida (subfam. algnCRGIded®) >. in04 avin kd abv analy ba een queen deca os 4. Kalanchoé. > Subfam. SEDOIDEAE Berger 1. Sedum Linnaeus, Sp. Pl. 1: 430. 1753: Gen. Pl. ed. 5. 197. 1754. Annual, biennial, or perennial, + succulent, often evergreen or decidu- ous herbs [subshrubs or shrubs], cespitose or with mat-forming vegetative shoots and then usually with slender, ascending or erect, terminal or oc- casionally axillary flowering stems. Roots generally fibrous, occasionally thick and tuberous, sometimes from a rhizome or a fleshy or woody caudex. Leaves glabrous, often glaucous [or rarely glandular-pubescent], mostly alternate, occasionally opposite or whorled, often imbricate and sometimes subrosulate, sessile or short-petiolate, often spurred at base, the blades flattened to cylindrical, ovoid, or subglobose, when flattened. linear to lanceolate, oblanceolate, ovate, or suborbicular in outline with entire to dentate margins [rarely spine-tipped]; leaves of vegetative and flowering 1978] SPONGBERG, CRASSULACEAE 207 shoots often dissimilar. Flowers perfect or imperfect (and the plants then dioecious or polygamodioecious), arranged in axillary or terminal, often bracteate, diffuse to compact corymbose, helicoid or scorpioid cymes (often with several helicoid or scorpioid cymes terminating a single flower- ing stem), in simple or compound dichasia [or rarely solitary or in spikes] ; insertion of the floral parts hypogynous. Sepals 4 or 5 [or rarely up to 9], imbricate or valvate in bud, free or connate basally, often fleshy and bract- like or leaflike. Petals 4 or 5 [or as many as the sepals, or sometimes absent in carpellate flowers of dioecious species], free or slightly united basally, erect or patent, yellow, greenish, white, pinkish, purplish [or very rarely blue], usually narrowly elliptic or subulate in outline, sometimes keeled. Stamens 8 or 10 [or twice the number of sepals] in 1 or 2 whorls, free [or those opposite the petals occasionally epipetalous]; filaments subulate to filiform, tapering to the yellowish or reddish anthers. Gynoecium of 4 or 5 [or as many as the number of sepals] erect or divergent carpels, the carpels free [or only slightly coherent basally], tapering into short, slender styles above; ovaries unilocular, with numerous for rarely 1 or few] ovules on 2-lobed parietal placentae along the adaxial sutures. Fruits erect or divergent follicles, often with stylar beaks, with many for rarely 1 or few] seeds, dehiscent along the adaxial sutures; seeds usually small, sometimes winged, with fleshy endosperm surrounding the straight embryo; mega- gametophyte of the Polygonum or Allium type. Base chromosome num- bers 4, 5, 6, 7, 8, 9, 10, 11. (Including Anacampseros Haw., Hylotelephium H. Ohba, Rhodiola L., Tetrorum Rose.) LECTOTYPE SPECIES: 5.0076 ‘Ls; see Hitchcock, Prop. Brit. Bot. 156. 1929. (Name from Latin, sedere, to sit or hold fast, in reference to the manner in which numerous species grow on rock outcrops and on walls; said by some to be derived from Latin for to assuage or soothe, apparently in reference to their reputed healing qual- ities.) —-STONECROP, LIVE-FOREVER, ORPINE. A large genus of some 300 to 600 species (Willis, 1966) widely dis- tributed in the Temperate Zone of the Northern Hemisphere, extending into the Arctic (ca. four species), and into the Southern Hemisphere of the Old World in the equatorial mountains of East Africa and in Madagascar. In the Western Hemisphere species of the genus range into South America in the Andes of Bolivia and Peru. Great concentrations of species are found in western North America; in Mexico, where twenty-eight species are endemic to the trans-Mexican volcanic belt alone (Clausen, 1959); in Europe and the Mediterranean region; and in eastern Asia. In the south- eastern United States seven or eight species are native, while three or four additional taxa have been reported as escapes from cultivation. Clausen (1975) recognized twenty-two additional species as native to North Ameri- ca north of the Mexican Plateau, and he also recorded thirteen (both native North American and exotic) taxa as naturalized, while he included fifty-three species in his annotated list of cultivated species. Largely in response to interest in the species as garden plants, the first modern treatment of Sedum was prepared by Praeger (1921), who, in 208 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 accounting primarily for taxa known in cultivation, laid the basic taxo- nomic framework for the genus. Praeger distributed the species among ten sections and numerous groups, while Berger (1930), treating the entire genus, recognized twenty-two sections, as well as subordinate series and groups. Froderstrom (1930, 1931, 1932, 1935), who prepared the most exhaustive world-wide treatment, recognized two sections (RHODIOLA and TELEPHIUM) but treated the remaining species on a geographic basis, dividing taxa within each major geographic area into two classes (on thie basis of erect versus divergent carpels) and numerous subordinate groups. In Praeger’s and Berger’s classifications, which are generally more utilitari- an, are currently in wider use, and also coincide at the sectional rank for our species, the native and adventive southeastern species belong to four ections. In treating the North American species, Clausen (1975) has pees three of these four subgeneric taxa but has elevated each of them to the rank of subgenus while including the species of the fourth, sect. EPETEIUM Boiss., in subg. SEpuM. He recognizes five sections within subg. SepuM and two within subg. Ruoproia.2 Section Ruoprora (L.) Scopoli (Rhodiola L., pro gen., Sedum subg. Rhodiola (L.) H. Ohba) (perennials with deeply rooted, fleshy or woody caudices terminated by clasping, well-developed or reduced, scalelike leaves with annual, leafy flowering stems developed in their axils, the mature follicles erect, with stout stylar beaks) is primarily a Eurasian group that is represented in North America by only two (or perhaps three) species. Sedum Rosea (L.) Scopoli (Rhodiola Rosea L., Sedum Rhodiola DC Os roanensis Britton, S. Rosea var. roanensis (Britton) Berger, Rhodiola roanensis (Britton) Britton), 2n = 18, 22, 32, 33, 36, is a wide- ranging circumpolar species, with a discontinous distribution at higher altitudes in mountainous regions, that occurs throughout much of Asia, northern Eu- rope, and arctic and boreal North America. In our area, it has been recorded from over 6,000 feet on Grandfather Mountain, North Carolina, and on Roan Mountain, on the North Carolina- i eapeees boundary. These southern localities, where S. Rosea grows in moist soil in crevices and on rock ledges, usually on north-facing cliffs, are some 500 miles south of the nearest known populations in Bucks County, Pennsylvania. Farther north it occurs at high elevations in New York, Vermont, and northward along the coast of Maine into Canada. In western North America, S. Rosea extends southward in the Cordillera into California, Colorado, and New Mexico, and in the Rocky Mountains (where it is sometimes segregated as S. integrifolium (Raf.) A. Nelson (Rhodiola integrifolia Raf., S. Rhodiola Torrey, not DC.) or treated as S. Rosea subsp. integrifolium (Raf.) Hultén), it is often sympatric with S. rhodanthum A. Gra y (Clementsia rhodantha (A. Gray) Rose), 2n = 14, the second ane ae of the section in North America. * Although Clausen (1975, p. 474) ee a subg. ee the change in status had been proposed by H. Ohba (1975, p. 285) some months before the publication of Clausen’s treatment. Despite this, Ohba (1977) credits i combination to Clausen 1978 | SPONGBERG, CRASSULACEAE 209 Characterized by usually imperfect, four-merous flowers in dense, ter- minal corymbose cymes and by fleshy or woody caudices that reputedly emit an odor of roses when dried, plants of Sedum Rosea are dioecious or occasionally polygamodioecious; individuals and entire populations with perfect flowers occur in some areas. The petals are greenish or yellowish, often tinged red at the apex, or deep red, and the numerous, closely spaced, glaucous, alternate leaves of the erect stems are oblanceolate, with acute apices and entire or irregularly dentate margins. Extremely polymorphic over its wide geographic range, Sedum Rosea consists of numerous races that differ from one another primarily in flower color, leaf-margin dentation, and plant size and habit. Numerous segregate taxa have been proposed, particularly from Asia. Britton & Rose (1903, 1905) considered the southern Appalachian populations to comprise a distinct species (S. roanensis, as a species of Rhodiola); they also recog- nized four segregate species in western North America. Hultén (1945), however, contended that the majority of segregate taxa should be considered informally as local races, while others should, at most, be treated at the subspecific or varietal rank. Sedum Rosea is also variable cytologically, and Uhl (1952) has shown that two chromosome races, one with 2” = 22, the other with 2n = 36, are widespread in North America. Wiens & Halleck (1962), however, have reported that Rocky Mountain plants of S. Rosea have 2n = 32, anda count of 2n = 18 has been reported by Sokolovskaya & Strelkova (1960). Other than inconsistent differences in flower color, plant sexuality, and habit, the cytological differences appear to lack morphological correlations, yet the 22-chromosome race occurs in glaciated northeastern North America (and in Eurasia), while the 36-chromosome race is apparently confined to unglaciated areas of the continent. Plants from our area remain uninvesti- gated cytologically, and, although it has been questioned whether these southern populations are still extant (Wherry, 1934; Uhl, 1952), Clausen (1975) reported seeing two plants on Roan Mountain in 1972. Section TeLEPHiuM S. F. Gray (Sedum subg. Telephium (S. F. Gray) R. T. Clausen, Hylotelephium H. Ohba) (perennials from caudices lacking scalelike leaves, with slender to thick fusiform roots, producing erect, rarely decumbent, usually annual, leafy flowering stems; flowers perfect, 5-merous, with erect, stipitate carpels, the mature follicles erect, with slender stylar beaks) is also widely distributed in temperate regions of Eurasia and is represented in our flora by Sedum telephioides Michaux (Anacampseros telephioides (Michaux) Haw., Hylotelephium telephioides (Michaux) H. Ohba), 2x = 24. The only species of the section indigenous to North America, S. telephioides grows on rock ledges and in crevices on exposed and shaded cliffs, and on stony slopes in rich woods in the moun- tains and adjacent upper Piedmont from southern Pennsylvania south- ward into Georgia; it is also known from limestone bluffs along the Ohio River in southern Indiana (Harrison County) and from southern Illinois.? * House (Annot. List Pl. N. Y. 376. 1924) pointed out that records for S. tele- 210 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Blooming in late summer and fall, Sedum telephioides is rather variable vegetatively but is easily recognized among our native sedums by its erect stems with flattened, mostly alternate or sometimes opposite, occasionally glaucous, broadly ovate leaves with subentire to coarsely toothed margins, and its numerous flowers arranged in corymbiform cymose inflorescences 2.5-8(—14) cm. broad. The carpels are pinkish, subtended by nectariferous scales about as broad as long, and the sepals and stamens are shorter than the pink to whitish petals. Closely allied to Sedum Telephium L., a wide-ranging Eurasian species, and the numerous (up to 25) additional Eurasian taxa included within this complex (several of which have been variously either segregated from or included within S. Telephium), S. telephioides has been retained as a dis- tinct species by American botanists despite the suggestions of both Praeger and Berger that it is probably only a variety of S. Telephium. Like the S. Rosea group, the S. Telephium complex is exceedingly variable both morphologically and cytologically, and its classification and nomenclature are both confusing and in need of careful study. As a group, the Eurasian taxa can reputedly be distinguished from S. telephioides by their fusiform roots, their flowers with the sepals scarcely one-third as long as the petals, their stamens about equaling or exceeding the petals, and their nectariferous scales longer than broad. Of the taxa included within the Sedum Telephium group, S. tele phioides appears to be most closely allied to S. Telephium subsp. alboroseum (Baker Frod. (S. erythrostictum Miquel, S. alboroseum Baker, including S. specta- bile Boreau, fide Fréderstro6m, Hylotelephium erythrostictum (Miq.) H. Ohba), 2n = 48, 50, 51, of China, Manchuria, and Japan, which has been reported as an escape from gardens i in northeastern North America, south- ward into Virginia and North Carolina. An herbaceous perennial with spreading to procumbent stems to 60 cm. tall, with alternate, opposite, or whorled leaves and pinkish white flowers in flat corymbose cymes, subsp. alboroseum flowers in late summer and fall and is also closely linked to S. Telephium subsp. Telephium var. purpureum L. (S. Telephium subsp. purpurascens (Koch) Areschoug, S. purpureum (L.) Schultes, S. pur- purascens W. D. J. Koch), 2n = 36, 48.4 Also a common garden plant, . Telephium var. pur pureum hae been reported as an escape and persisting outside of cultivation near waste heaps in Warren County, North Carolina. Native to Europe, plants of S. Telephium var. pur pureum are stout, glabrous herbs from fusiform roots, with alternate, broadly oblong or elliptic leaves with coarsely and irregularly toothed margins and flowers with purplish red to rosy petals in dense, terminal and lateral, subglobose cymes. Unlike those of S. telephioides, the leaves of S. Tile ii var. ees from western and central New York state are generally erase to S. Rosea. One substantiated report from Dutchess County is probably based a plant ae from cultivation (cf. Wherry, 1936). “Material determined to be 2n = 48 tentatively assigned to subsp. Telephium var. ei (as subsp. rere (Koch) Areschoug) by Jalas & Rénkk6 (1960). 1978] SPONGBERG, CRASSULACEAE 211 purpureum are reduced in size up the flowering stems. Sedum Telephium has also been listed as persisting outside of cultivation in Knox and Sevier counties, Tennessee, but it is not known to which subspecies these plants should be assigned. Allard (1932, 1940), moreover, contended that true S. Telephium is a long-day plant that rarely blooms in gardens south of New England. It is obvious that the reports of taxa of this complex in our area should be examined critically during the course of an overall study of the group. The majority of species of Sedum have been placed in sect. SepuM (Seda Genuina W. D. J. Koch, sect. Eusedum Boiss., subg. Sedum fide R. T. Clausen) (annual or usually perennial herbs or subshrubs with creeping rhizomes, cespitose sterile stems, and erect or ascending flowering stems; flowers perfect, the mature follicles divergent). This section is represented in the southeastern states by four native species and two adventives that have become naturalized. Sedum pulchellum Michaux (including S. vigilmontis Small), 2n = 22, 44, 66, occurs primarily on limestone, but also grows on granitic and sili- ceous outcrops in open and wooded habitats from western Virginia through Kentucky and Tennessee to southern Illinois, Missouri, and Kansas, and southward into northwestern Georgia, northern Alabama, Arkansas, Okla- homa, and Texas. Sedum pulchellum is usually a winter annual or occa- sionally a biennial or perennial due to persistence of vegetative shoots. Its seeds germinate in the fall and produce small, lax rosettes of flat, greenish gray, oblong-spatulate leaves (FicurE 1, a). Early in spring, decidedly different erect flowering stems develop with narrowly linear, cylindrical, terete leaves with auriculate or sagittate bases (FicuRE 1, b). These leafy stems, 10-30 cm. high, are terminated by three (rarely to seven) widely divergent, recurved, secund scorpioid cymes 3-4 cm. long comprising many four-merous flowers with whitish to deep pink or purplish petals that are about twice as long as the small sepals. In their vegetative, over-wintering stage, plants of Sedum pulchellum have been confused with S. Nevii A. Gray (S. Beyrichianum Masters (?), S. Nevii var. Beyrichianum (Masters) Praeger (?)), 2” = 12, a narrow endemic known with certainty from only five localities (three in Alabama, one in southeastern Tennessee, and one in extreme west-central Georgia near the Alabama line). A slender, only slightly glaucous perennial with loose, sterile rosettes of leaves 1-5 mm. wide, flowering stems with 12-40 narrowly oblanceolate or linear-oblanceolate, usually subterete leaves 3-19 mm. long, 1-3 mm. wide, white, four-merous flowers in two or three slender helicoid cymes terminating flowering stems, and seeds 0.5—0.6 mm. long, S. Nevii grows on exposed and/or shaded cliffs and rock ledges. Until Baldwin’s (1942) cytological investigations revealed that the Alabama plants are diploids with 2n = 12, it was generally assumed that S. Nevii had a wider range in the eastern United States. Virginia plants referred to S. Nevii, however, proved to be diploids with 2n = 28, and, as a result of this chromosomal difference and very subtle morphological dissimilarities, Clausen (1946) segregated the Virginia populations as S. glaucophyllum. 712 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 59 4 é 2 aide: = a aly, 8 Ny & uy OP 4 Ay Me NAY 2 Figure 1. Sedum. a-d. S. pulchellum: a, over-wintering rosette, X 1%: b, flowering shoot, X 1; c, flower, X 5; d, cross section through four carpels of gynoecium, 20, e-}, S. pusillum: e, habit of mature plant, < 1; f, longitudinal section through immature follicles (mote nectaries, solid black, at base of car- pels), X 6; g, immature follicles, * 6; h, mature, dehisced follicles, X 6; i, e ; J, embryo, & 25. k, S. glaucophyllum, leafy shoot, X 1. 1, S. ternatunt, leafy shoot, * !2. m, S. telephioides, outline of leaf, « 1. ~ Moreover, western records for S. Nevii in southern Illinois and Missouri have resulted from confusion with S. pudchellum (cf. Wherry, 1934, 1936). Further cytological investigations by Uhl (1970) have revealed chromo- some numbers of 27 = 28, 44, and 56 within Sedum glaucophyllum, which is now known to range from southern Maryland, southward through the mountains and adjacent Piedmont of West Virginia, Virginia, and North Carolina, where it is known from Rockingham and Surrey counties. A 1978 | SPONGBERG, CRASSULACEAE rales) glaucous perennial with dense, sterile rosettes of leaves 2-7 mm. wide, flowering stems with 30-50 flat, oblanceolate leaves 5—25 mm. long and —~6 mm. wide, white, four-merous flowers in three or sometimes four slender helicoid cymes terminating the flowering stems, and seeds 0.5-0.9 mm. long, S. glaucophyllum grows in moist rock crevices on both sunny and shaded cliffs. Acceptance of the distinct species status of Sedum glaucophyllum has been tentative since morphologically it is practically indistinguishable from S. Nevii, and the characters used to separate it are largely qualitative. Clausen himself (1946) cautioned that “. .. if the morphological and cytological differences are not correlated . . . the classification again must be revised and S. glaucophyllum, though a genetical species, would need to be handled taxonomically as a synonym under S. Nevii, since it would not be a practical unit capable of recognition by ordinary observation.” However, since S. Nevii sensu stricto is a narrowly restricted endemic, the use of the name S. glaucophyllum for plants throughout the distinct, more northeastern range of that taxon is facilitated on a geographic basis and seems justified considering the state of our biological knowledge of both axa. Another species of sect. SEpuM, Sedum ternatum Michaux (Anacamp- seros ternata (Michaux) Haw.), 2n = 16, 24, 32, 48, is widely distributed throughout the eastern United States, from New Jersey and Pennsylvania, west through Ohio and Illinois to Missouri, and southward to the Carolinas, Georgia, and Arkansas. Easily recognized by its flat, elliptic to spatulate or rounded leaves usually in whorls of three on sterile shoots, and its obovate to spatulate leaves on the erect flowering stems, S. ternatum is a procumbent, mat-forming perennial with white, four-merous flowers with crimson or purplish anthers. The flowers are usually arranged in three terminal scorpioid cymes. The plant often occupies rich, moist sites in woodlands and frequently grows on rocky slopes and alone stream banks where it sometimes forms extensive, carpet-like stands. Baldwin (1942) discovered the existence of four chromosome races: diploids (2m = 16), triploids (2n = 24), tetraploids (2n = 32), and hexaploids (2n = 48). While these races lack morphological correlations, the diploid race (West Virginia, Kentucky, and Virginia) and the hexaploid (one locality near Tuscaloosa, Alabama) have limited geographic ranges in contrast to the widespread tetraploid and the sporadic occurrence of triploid plants. m sarmentosum Bunge, a native of northern China and Japan, is also a prostrate, mat-forming perennial with long, creeping sterile shoots with leaves in whorls of three and erect flowering stems. Its flowers, how- ever, are arranged in terminal compound cymes and are bright yellow. Occasionally cultivated as an ornamental, S. sarmentosum has been re- ported as naturalized in a few dry, open, rocky areas in the mountains of North Carolina. Small (1933, p. 588) probably confused this species with S. mexicanum Britton (S. sarmentosum Masters, not Bunge) when he said that it (S. mexicanum) “. . . is often cultivated and occasionally escapes from gardens” in the Southeastern States. 214 JOURNAL OF THE ARNOLD ARBORETUM [VvoOL. 59 Another adventive, Sedum acre L., 2n = 16, 40, 48, 60, 80, 100, 120, is native to a wide area of the Old World in Europe, northern Asia and Asia Minor, and northern Africa. Plants of this species are small, creep- ing, mat-forming, evergreen perennials, with bright yellow flowers pro- duced in two or three terminal cymes. They can be recognized by the overlapping, alternate, flattish, deltoid leaves that are broad and slightly spurred at the base. One of the most widely cultivated sedums in rockeries and on walls, S. acre has apparently become naturalized in Avery and Mitchell counties, North Carolina Section Eperrirum Boiss. (sect. Procrassula (Griseb.) Schonl., pro parte: Procrassula Griseb.) (annual or biennial herbs with slender, poorly de- veloped root systems, the flowering plants lacking sterile shoots, with alternate, flattened or cylindrical leaves and branched cymose, corymbi- form, or dichasial inflorescences comprising several to numerous 4— 9-, mostly 5-merous flowers with white, yellow, red, or blue petals) is some- times included within sect. Sepum and is represented in our area by two species. Sedum pusillum Michaux (Tetrorum pusillum (Michaux) Rose), 2m = 8, is a diminutive winter annual endemic to the Piedmont of the Carolinas and Georgia. In this region S. pusillum is totally restricted to granitic flat-rock communities, and it is sometimes confused with Dia- mor pha Smallii Britton (q.v.). Sedum pusillum is typically found growing in the partial shade of pines or Juniperus virginiana L. at the margins of the rock outcrops and in sheltered depression pits on outcrop surfaces, often near vernal pools. Diamorpha Smallii, on the other hand, grows in exposed depression pits with very shallow soils. A small, early-spring- flowering, usually branched herb 4-12 cm. high with bluish green (or sometimes reddish) cylindrical leaves and four- or rarely five-merous white or pinkish flowers with red anthers arranged in small helicoid cymes or simple or compound dichasia, S. pusillum differs in several significant anatomical and morphological features from Diamorpha. These dif- ferences, several of which are diagnostic in the field, are discussed under Diamorpha. The diploid chromosome complement of 27 — 8 known for S. pusillum is the lowest diploid number recorded in the Crassulaceae (Baldwin, 1940) Sedum Nuttallianum Raf. (S. Torreyi G. Don, S. sparsiflorum Nutt. in Torrey & Gray), 2x = 20, also a diminutive winter annual, ranges from Arkansas and Missouri, southward into Oklahoma and Taras. Flowering from April to July, S. Nuttallianum grows in the shallow soils of chert and sandstone glades and ledges and in the clayey soils of pasturelands. Individuals of S. Nuttallianum are small plants 5-15 cm. high, usually branched from the base, with short, pale, silvery green, cylindrical, alternate leaves and sessile or short-pedicellate, five-merous flowers with lanceolate yellow petals arranged in two to five terminal cymes. Although the individuals are small, the plants commonly grow in large, dense populations and give an overall effect in the field of a low, mat-forming perennial. Despite continued investigations of the genus, Sedum remains extremely 1978] SPONGBERG, CRASSULACEAE 21s complex taxonomically and confusing nomenclaturally. While the sub- generic classification is not without problems, particularly some concerning the limits and validity or potential generic status of certain sections, the greatest confusion appears to center around the definition of species and subspecific taxa. Certainly some of the confusion is the result of the fact that “dried specimens can be the source of erroneous impressions concern- ing the dimensions and shapes of structures, as well as the habits of plants . . .” (Clausen, 1959, p. 337). In addition, many species are narrow endemics, while others are extremely wide-ranging and exhibit a polymorphism that appears to be proportional to the extent of the geo- graphic range. Furthermore, individuals of a given taxon may appear quite different from one another depending upon their habitats and the time of year the plants are observed or collected. The color of the stems and foliage, for example, is prone to change, usually becoming reddish under dry conditions, with age, or when grown in intense light. But, as pointed out by Uhl (1961), “the most conspicuous problem is the frequency with which heteroploidy occurs within what many authors have regarded as a single species... .” While this phenomenon is characteristic of the Crassulaceae in general, it is particularly evident in Sedum, where both polyploidy and aneuploidy have been documented for numerous species, several of which are native to our area. Furthermore, every gametic chromosome number from m = 4 tom = 38 is known in the genus. However, S. spathulifolium W. J. Hooker, 2n = 30, is cited as an exception to the rule, inasmuch as it is a wide-ranging species (Vancouver Island, south to southern California) that apparently has the same chro- mosome number over its entire area (Uhl, 1961b). As a possible explanation of the usual lack of morphological correlation with the heteroploid cytological condition found in numerous species, Uhl (1961b) suggests that these taxa are presently undergoing rapid evolution and that “each karyotype has not yet evolved its own distinctive genotype and phenotype. . . .”. This hypothesis gains support from the fact that the usual habitats of species of Sedum are cliffs and rock outcrops, habitats that are discontinuous and unstable in terms of geologic time and, as a result, are probably conducive to rapid evolutionary rates. Likewise, Ceca (1959) suggests that the morphological and physiological differences between the closely related species endemic to the trans-Mexican volcanic belt are largely due to gene mutations, while changes involving chromosome number have led primarily to genetic isolation between sympatric taxa. Investigations employing electrophoretic techniques might give support to these interpretations. While the chromosomes of Sedum are generally small, Uhl (1961) and Uhl & Moran (1972) point out that they seem to possess localized kinetochores, ruling out the possibility that the heteroploid condition is the result of diffuse kinetochores and chromosome breakage like that documented for Carex (Cyperaceae) (Davies, 1956). Levan (1933) reported heteromorphic sex chromosomes for a single plant of S. Rosea 216 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 from a dioecious population, but in his 1952 report, Uhl could not dis- tinguish specific sex chromosomes. In their cytological survey of Japanese and South Korean Crassulaceae, Uhl & Moran (1972) calculated the approximate nuclear volume of cells of taxa within two complexes of Sedum. A nearly proportional increase of nuclear volume was found in sect. A1zoon, a complex comprising diploids (where » = 16) to apparent heptaploids, suggesting that the change in chromosome number within this group is the result of “multiplication of the more or less complete genome.” This situation is in contrast to that encountered in S. polytrichoides Hemsley ex Hemsley & Forbes, sensu lato, where no clear relationship between chromosome number and _ nuclear volume was evident. Uhl & Moran speculate that in this highly dysploid species, where plants are known to vary from x = 11 ton = 35, “very extensive rearrangement and repackaging of essentially the same amount and kind of chromosomal material, probably chiefly by translocations, have accompanied and made possible the many changes in chromosome number in this complex.” Interspecific hybridization, except within sect. TELEPHIUM, is thought not to occur in nature, and there have been no attempts to produce artificial hybrids utilizing cultivated materials (Evans, 1971). Even with- in taxa exhibiting polyploidy it has generally been concluded that the polyploid condition has been achieved through autopolyploidy and not through allopolyploidy. Certain hypotheses concerning the origins of some taxa, however, have been advanced that are based on the assumption that interspecific hybridization has occurred. Notable among these hypotheses is Baldwin’s (1940) suggestion that Diamorpha Smallii, 2n = 18, is the amphidiploid product of “fusions between the 4- and 5 chromosome tendencies” within Sedum as represented by S. pusillum and S. Nuttal- lianum, respectively. The scant data concerning floral biology of Sedum indicate that there is a range from proterogyny and homogamy to marked proterandry. In proterandrous flowers the anthers of the outer whorl of stamens dehisce first, followed by the slightly delayed dehiscence of the anthers of the inner whorl. The movement of the stamens to an erect position over the carpels just before dehiscence and their subsequent return to the periphery of the flower afterward are suggestive of similar movements observed in flowers of several genera of subfam. Saxifragoideae of the Saxifragaceae. Although the maturation and receptiveness of the stigmas and the de- hiscence of the anthers are usually not synchronous, self-pollination may occasionally occur; however, no studies have been located that deal with compatibility systems within the genus. Nectar is secreted by the small, scalelike glands that subtend the carpels (the glands considered to be dorsal outgrowths of the carpels), and the few recorded observations suggest that visitors attracted to the flowers are usually dipterans, hymenopterans, and lepidopterans. Dispersal of seeds from the mature, dehisced follicles is probably by wind action that sets up short-term vibrations of the follicles. The seeds 1978] SPONGBERG, CRASSULACEAE 217 are small, and those of Sedum Rosea are narrowly winged at each end. The low, often creeping, vegetative shoots of numerous species root at the nodes, and propagation and dispersal undoubtedly occur frequently by vegetative means. Both pieces of stem and individual detached leaves of most species develop adventitious roots and shoots, and Yarbrough (1936) found that in leaves the mode of origin of these structures is through the activity of a secondary meristem that develops from vacuolated, differen- tiated parenchyma. This “regeneration” is in contrast to the formation of the plantlets produced along the leaf margins or at leaf or petiole bases in species of Kalanchoé, which are the result of the activity of a residual primary meristem. The vascular anatomy of the flowers of a relatively large number of species was investigated by Quimby (1971), who placed these taxa in four groups based on the number of whorls of traces supplying the flowers. The majority of species (37) have four whorls of traces (Group III), while others have six (Group I, the presumed underived condition found in species of sect. TELEPHIUM), five (Group II), or three (Group IV) whorls. Vertical compression of the receptacle has apparently led to the reduction in the number of whorls. Additional study of the vascular- ization of flowers of Sedum pusillum (Sherwin & Wilbur, 1971), con- ducted to help clarify the generic lines between Sedum and Diamorpha, showed that S. pusillum does not fit into either Group III or Group IV. Sherwin & Wilbur suggest that the vascular pattern found in flowers of S. pusillum, which Quimby placed in Group III, is perhaps derived from the basic Group III pattern. Furthermore, their work indicates that more detailed investigations of individual species may show that floral vascular- ization in Sedum is more complex and variable than outlined by Quimby. The carpels of Sedum, each of which is supplied by five vascular bundles, have been considered primitive since they are open during on- togeny, closing only during later development when the margins along the adaxial sutures are only slightly fused, and the marginal epidermal layers remain distinct (Eames, 1931). Henslow (1891) reported that the ventral carpel traces in Sedum Telephium are not inverted (a condition encountered in Silene and Dianthus [Caryophyllaceae] and in some species of Rhodo- dendron subg. Anthodendron |Ericaceae]), but Subramanyam (1955) de- tected only inverted traces in species he examined. Relatively few investigations of the general vegetative anatomy of Sedum have been undertaken; see the works of Solereder (1899) and Metcalfe & Chalk (1950) for some details and references to most studies, as well as the more recent papers by Piaget (1966) and Jensen (1968). Aspects of megasporogenesis and embryology of Sedum coincide with the general pattern given for the family with only a few significant ex- ceptions. In S. Telephium subsp. fabaria (W. J. D. Koch) Kirschleger and S. populifolia Pallas (Hylotelephium populifolium (Pallas) H. Ohba), a subshrub from eastern Asia, an Allium-type megagametophyte has been reported to develop from the chalazal dyad cell, while in other investigated species a monosporic Polygonum-type is produced. Endosperm formation 218 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 is ab initio Cellular and a chalazal haustorium is developed, but Sub- ramanyam (1963) found that in S. ternatum the haustorium remains 1-nucleate. Sedum ternatum is also distinguished by the presence of apical “caps” on the hooked synergids, structures that have been reported else- where only for Helianthemum vulgare (Cistaceae) and species of Ribes (Saxifragaceae or Grossulariaceae). See Subramanyam’s articles (1955, 1963) for additional details and information about microsporogenesis. Pollen of some European species of Sedum has recently been examined by ‘t Hart (1974), who found the grains to be 3- or rarely 2- or 4-zono- colporate. The sexine a, a tectum, is ornamented with short striae in a rugulate pattern, and, although considerable variation in morphology was evident, *t Hart concluded that the range of variation encountered within a single taxon, and occasionally within a single sample, equaled that found during the survey. All of the species examined (including one miscellaneous American species) belong to one variable pollen type and form a continuous morphological series that can only arbitrarily be divided into subtypes. However, ’t Hart ventures to assert that within the Rupestria series species with low base chromosome numbers have relatively underived pollen subtypes, while those species with higher base numbers exhibit derived pollen subtypes. This correlation was not evident in the Sedum acre group, suggesting less phylogenetic unity within that taxon. Determination of the generic relationships of Sedum is dependent, in large measure, on the delimitation of the genus and the several groups that have variously been either segregated from or included within the genus. At one extreme, as Moran (1942) points out, Kuntze (1898) concluded ‘*. . . that all the species of the Crassulaceae with free carpels, i.e., practically the entire family, were best placed in the one genus Sedum.” Conversely, there has been a continued tendency on the part of authors of floristic works (culminating in the treatments of Britton & Rose, 1903, 1905) to accord generic status to small groups or single species that within a local area appear distinct. In other instances, species or groups of species once included within Sedum have been removed and merged with other genera as warranted by careful analyses of relationships (e.g.. Hasseanthus Rose, a segregate of Sedum, is now treated as a subgenus of Dudieya Britton & Rose, a member of subfam. Echeverioideae Berger). In particular, sect. RuoproLa is often recognized as a distinct genus, especially when the species of that section with which floristic workers are concerned have four-merous, imperfect flowers. Sedum rhodanthum of the Rocky Mountains (its flowers five-merous and perfect), like several of the Asiatic species, clearly belongs to the section as emended by Praeger, and it is transitional to species in other sections of the genus. However, despite the obvious similarities, S. rhodanthum was removed to the mono- typic genus Clementsia by Rose (in Britton & Rose, 1903) in an attempt to create greater unity within sect. Ruopio_a, which Britton and Rose also treated as a genus. Similarly, species of sect. TELEPHIUM, some of which are transitional to sect. SepuM, have been transferred to the segregate genus Telephium Hill, not L., or to Anacampseros P. Miller, and Ohba 1978 | SPONGBERG, CRASSULACEAE 219 has recently placed these species in the new genus {ylotelephium. In our area, in addition to recognizing Diamorpha, Britton and Rose also segre- gated S. pusillum, placing it in the monotypic genus Tetrorum Rose because of its annual duration and its four-merous flowers. Certainly some of the problems of delimiting acceptable genera are due to the fact that affinities within the entire family appear to be inter- linear as well as linear (Baldwin, 1939). Nonetheless, it is apparent that in the New World the closest affinities of Sedum are largely with genera of subfams. Sedoideae and Echeverioideae. Most of the closely allied genera of the Sedoideae share the characteristic terminal flowering stems and flowers with free, spreading petals, while genera of subfam. Eche- verioldeae are csually characterized by lateral flowering stems and by flowers with basally connate, erect or spreading petals. Sedastrum Rose (Sedum sect. Sedastrum (Rose) Berger) is distinguished from Sedum by its persistent, rosulate basal leaves and its flowers with enlarged nectaries and white petals, while Parvisedum R. T. Clausen (Sedella Britton & Rose, not Fourreau; Sedum sect. Sedella (Britton & Rose) Berger), a small genus of four annual species of the California mountains, is characterized by its carpels with solitary ovules. Diamorpha differs from Sedum in several significant respects discussed elsewhere, but the contrasts between Villadia Rose (including Altamiranoa Rose) are less concrete. Consisting of about 45 species of Mexico, Central America, and the Andes of South America, species of Villadia produce terminal in- florescences, and in the last analysis the genus is separated from Sedum only on the basis of the degree of basal connation of the petals (Moran, 1971). Lenophyllum Rose, a small genus of about five species restricted to the Gulf Coastal Plain of southern Texas and Mexico, is characterized by opposite leaves and flowers arranged in terminal racemose or spicate cymes. In habit the species of Lenophylium resemble species of Echeveria DC. (ca. 143 species ranging from Texas to the Andes of South America), but Uhl & Moran (1953) suggest that Lenophyllum, which was once placed in subfam. Echeverioideae, is best removed to subfam. Sedoideae. Clausen (1975) contends that the genus was probably derived from the same phyletic line that gave rise to Villadia, Echeveria, and Pachyphytum. While the relationships of the several genera of subfam. Echeverioideae are taxonomically complex and cytologically diverse and complicated (Uhl & Moran, 1953), it would appear that Echeveria is closely allied to Sedum. Walther (1936, 1972) speculates that Echeveria was derived from species not unlike certain subshrubby species of Sedum belonging to sects. PacHysEpUM Berger and DENpRosEDUM Berger, which are endemic to Mexico (the center of diversity of Echeveria) and share the characteristic lateral flowering stems of subfam. Echeverioideae.® Although hybrids between species of Echeveria and Sedum sects. PACHYSEDUM and * Walther employs the name Sedum sect. Bergerosedum Walther to refer to those species of the two sections mentioned aes some of these species were also treated sectional name, however, was SReRC Ieee superfluous when published. In ad- 220 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 DENDROSEDUM have been produced (Walther, 1953; Uhl, 1966, 1967, 1970; one placed in the hybrid genus « Sedeveria Walther), this evidence of close relationship loses some of its potential significance due to the numerous intergeneric hybrids produced between Sedum and other genera of subfam. Echeverioideae (Uhl, 1966; Uhl & Moran, 1973; Knobloch, 1972) Uhl & Moran (1973), considering the cytological evidence, indicate that many Mexican species of Sedum (where x = 29-36) belong to a comparium that includes, in addition to Echeveria (where x = 27-34), spe- cies of Pachyphytum Link, Klotzsch, & Otto, x = 31-33 (12 species of Mexico distinguished by pairs of scales on the adaxial surfaces of the petals), Thompsonella Britton & Rose (Echeveria sect. Thompsonella (Britton & Rose) Berger) (two species of south-central Mexico with the flowers arranged in a panicle or thyrse and with red lines on the adaxial surfaces of the petals), and Graptopetalum Rose (Byrnesia Rose: Sedum sect. Graptopetalum (Rose) Berger; Echeveria sect. Graptopetalum (Rose) Kearny & Peebles), « = 30-35 (11 species of Arizona and Mexico with red dots, often in transverse bands, on the adaxial surfaces of the petals). Uhl & Moran (1973, p. 655) state that the chromosomes of these taxa “. . . apparently aes ee homology for one another, as shown - their ability to pair, often more or less normally, in intergeneric hybrids. Very likely it was from an ancestral plexus with such [| base chromosome] numbers, in the more primitive genus Sedum, that the other four genera were derived.” Plants of Sedum Telephium and other species are occasionally used as greens in salads, and Uphof (1968) reports that the Eskimos of Alaska eat the leaves of S. Rosea fresh, soured, or in oil. Sedum acre has ap- parently been used as a laxative, while the juice from plants of the Mexican and Guatemalan S. dendroideum Mog. & Sessé is astringent and has been used to harden gums and to treat chilblains and dysentery. The juice of this species has also been reported by Palarea (1954, as S. praealtum A. DC.) to cure opacities of the lens and cornea in humans. In addition, various medicinal uses have been reported for S. Rosea, and numerous studies of the chemical and physiological effects of this species have been conducted, primarily by Russian investigators: see Clausen (1975) for an extensive list of Russian references. REFERENCES: Under family references, see, in particular, BACKER: BAILEY et al.; BAILLON; BaLpwin, 1939, 1940; BERGER; Borisov; BRAMWELL; BrITTon & Rose, 1903, N; KNUTH; MET- CALFE & CHALK; Moran, 1942a, b, & d, 1971, 1972; Mort; Quimpy; RICKETT; ROUHANT; RouHan 1 et al.; SCHONLAND, 1890; SHARSMITH: SMALL; SOLEREDER; UHL, 1961. 1966, 1970, 1976; Unt & Moran, 1953, 1973: Upuor: WALTHER, 1936, 1972; and WILLIs. dition, Walther (1972, p. 39) apparently treats the name as a generic one, but I have been mele to locate any reference to its publication at that rank. 1978] SPONGBERG, CRASSULACEAE 271 Attarp, H. A. Length of day in relation to the natural and artificial distribu- tion of plants. Ecology 13: 221-234. 1932. [Considers Sedum Telephium and S. spectabile. | _ Sedum Telephium L. in the Bull Run Mountain area, Virginia. Castanea 5: 17-19, 1940. [Plants of Sedwm Telephium, escaped along roadsides and in abandoned dooryards, fail to bloom. & W. W. Garner. Further observations on the response of various species of plants to length of day. U. S. Dept. Agr. Tech. Bull. 727. 64 pp. 1 pl. 1940. Astruc, L. Phyllotaxie et point végétatif du Sedum maximum. Revue Gen. Bot. 56: 141-171. 1949. [Sedum maximum often treated as a subsp. of S. Telephium. | Atwoop, H. Sedum glaucophyllum. Cactus Succul. Jour. 25: 148, 149. 1953. [Includes notes on culture. BALDWIN, J. T., JR. Somatic chromosome numbers in the genus Sedum. Bot. Gaz. 96: 558-564. 1935. Polyploidy in Sedum ternatum. Jour. Hered. 27: 241-248. 1936. The cyto-taxonomy of the Telephium section of Sedum. Am. Jour. Bot. 24: 126-132. 1937. . Cytological basis for specific segregation in the Sedum Nevi complex. Rhodora 44: 11-14. : . Polyploidy in Sedum ternatum Michx. II. Cytogeography. Am. Jour. Bot. 29: 283-286. 1942. Polyploidy in Sedum pulchellum — I. Cytogeography. Bull. Torrey Bot. Club 70: 26-33. 1943. _ Affinities of Sedum Nevii. Rhodora 46: 450, 451. 1944. Banacu-Pocan, E. Cytological studies in three species of the genus Sedum L. Acta Biol. Cracoviensia Bot. 1: 91-101. pls. 10, 11. 1958. [Sedum Rosea, SSF. Os. | Baskin, J. M., & C. C. BASKIN. Germination ecology of Sedum pulchellum Michx. (Crassulaceae). Am. Jour. Bot. 64: 1242-1247. 1977. Bitter. G. Beitrage zur Kenntnis der Gattung Sedum I. Notizbl. Bot. Gart. Berlin 8: 281-284. 1923 Brwe., M. Sur la présence d’un glucoside 4 essence dans les tiges foliées et les racines du Sedum Telephium L. Jour. Pharm. Chim. VII. 26: 289-298. 1922.* [The results of this investigation also appear in Bull. Soc. Chim. Biol. 4: 242-250. 1922,* and in shortened form in Compt. Rend. Acad. Sci, Paris 174: 186-188. 1922.] Caup.e, C., & J. M. Baskin. The germination pattern of three winter annuals. Bull. Torrey Bot. Club 95: 331-335. 1968. [Considers germination of Sedum pulchellum. | CHITTENDEN, F. J. Sedum. In: F. J. CHITTENDEN, ed., Roy. Hort. Soc. Dict. Gard. ed. 2. 4: 1919-1927. 1956. [See also p. 508 in P. M. SYNGE, ed., Supplement, ed. 2. 1969; the treatment by Chittenden includes a key to cultivated species. | CLauseN, R. T. Studies in the Crassulaceae — III. Sedum, subgenus Gormania, section Eugormania. Bull. Torrey Bot. Club 69: 27-40. 1942. The section Sedastrum of Sedum. Ibid. 70: 289-296. 1943 Nomenclatural changes and innovations in the Crassulaceae, Cactus Succul. Jour. 18: 58-61; 74-77. 1946. [Segregation of Sedum glauco- phyllum and comparison with S. Nevii. | 222 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 The distribution and variation of Sedum Nevii. Ibid. 21: 180-185. 1949, Sedum of the trans-Mexican volcanic belt: an exposition of taxonomic methods. 380 pp. Ithaca, New Yor 9. Sedum of North America north of the Mexican Plateau. 742 pp. Ithaca, New York. 1975. [A monograph of North American Sedum, in- cluding numerous photographs and line drawings by Elfriede Abbe. | . Biennial species of Sedum of the Sierra Madre Occidental and Mexican Plateau. Bull. Torrey Bot. Club. 104; 209-217, 1977. & C. UnL. Studies in the Crassulaceae —V. Revision of Sedum Cockerellit and related species. Brittonia 5: 33-46. 1943. The taxonomy and cytology of the subgenus Gormania of Sedum. Madore 7: 161-180. pl. 22. 1944, Davies, E. W. Cytology, evolution and origin of the aneuploid series in the genus Carex. Hereditas 42: 349-365. 1956. Descuatres, R. Recherches sur la phyllotaxie du genre Sedum. Revue Gén. Bot. 61: 501-570. 1954. Eames, A. J. The vascular anatomy of the flower with refutation of the theory of or polymorphism. Am. Jour. Bot. 18: 147-188. 1931. [| Sedum, 56. p. EVANS, R L. A gardener’s guide to sedums, 44 pp. Alpine Gard. Soc. Guide. nd, 1971) on M. L. Sedum Rosea, not S. roseum. Rhodora 49: 79-81, 1947. [ The specific epithet Rosea is an old generic name FREEMAN, O. M. Notes on the flora of Polk k County, North Carolina. Castanea > 37-57. 1955. [Records Sedum sarmentosum as established on stone outcrops in Green River Cove FRODERSTROM, H. Etude des crassulacées de Chine eae provenant des récoltes du Pere Licent. Bull. Mus. Hist. Nat. Paris II. 1: 440- 443. 1929 . The genus Sedum L., a systematic essay. Part I. Acta Horti Gothob. 5(appendix): 1-75. pls. 1-28. 1930: Part II. Jbid. 6(appendix): 1-111. ls, 1-65. 1931; Part III. Jbid. 7 (appendix): 1-126. pls. 1-68. 1932; a IV. Lbid. eee 1-262. pls. 1-115. 1935. . Bot. Tidskr, 30: 216. 1936.] — Lay a Observation on ace flowering photoperiodicity, Rec. Trav. Bot. Neer! 40: 392-412. (1943-1945) 1946. [Sedum, 399-401. 1.| GUNTHaRT, A. Beitrase zur Bluthenbiologie der Cruciferen, Crassulaceen und der Gattung Saxifraga. Bibliot. Bot. 11(Heft 58): ix i 97 pp. pls. 1-11. 1902. [Contains information concerning Sedum and Sempervivum. | Ham_et, R. Revision des Sedum du Caucase. Trav. Jard. Bot. Tiflis 8: 1-37, 1903. [| Correction. . Contribution 4 l’étude phytographique du genre Sedum. Candollea 4: 1-52. 1929-1931, Hansinc, E. D., L. E. Lonctey, L. SANDO, & R. B. Harvey. Hardiness tests of some perennial sedums. Proc. Am. Soc. Hort. Sci. 33: 686-689. 1936. Hara, H. Contributions to the study of variations in the plants closely related to those of Europe or North Am Fac, Sci. Univ Tokyo Bot. 6: 29-96. 1952. [Information concerning the Sedum Telephium group and 77llaea, 62-64. | Hart, H. ’r. Cytological and morphological variation in Sedum acre L. in western Europe. Acta Bot. Neerl. 20: 282-290. 1971. [Sedum acre, 2n = 40, 60, 80, 100, and 120.] 1978 | SPONGBERG, CR ASSULACEAE 223 __ Chromosome numbers in the series Rupestria Berger of the genus Sedum L. Ibid, 21: 428-435. 1972 _ The pollen morphology of 24 European species of the genus Sedum L. Pollen Spores 16: 373-387. & T. C. Wrictey, eds. Sedums. Some observations on the genus. 58 pp. Morden. 1971.* Henstow, G. On the vascular systems of floral organs, and their importance in the morphology of flowers. Jour. Linn. Soc. Bot. 28: 151-197. pls. 23-32. 1891. [Floral anatomy of Sedum Telephium, 180. | HoLiiNncsHEAD, L. Chromosome studies in Sedum, subgenus Gormania, section Eugormania. Bull. Torrey Bot. Club 69: 41-43. 1942. Hooxer, J. D. Sedum pulchellum. Bot. Mag. 1022 .:62232 A516: & T. Tuomson. Praecursores ad floram Indicam: being sketches of the natural families of Indian plants, with remarks on their distribution, struc- ture, and affinities. Jour. Linn. Soc. Bot. 2: 1-103. 1858. [Crassulaceae, 89-103; includes notes on Sedum, particularly sect. Rhodiola. | Huser, J. A. Zur Systematik der Gattung Sedum L. 118 pp. + map. Landshut. 1929 _ Die Gattung Sedum L. in ihren Beziehungen zu den ubrigen Gattungen der Crassulaceae. Sukkulentenkunde 6: 40-42. 1957. Hutren, E. Flora of Alaska and Yukon. V. Acta Univ. Lund. II. Afd. 2. 41(1): 797-978. 1945. [Sedum, 895-899, maps 691-693, p. 974, including a discussion of S. Rosea. | The amphi-Atlantic plants and their phytogeographic connections. Sv. Vet.-akad. Handl. IV. 7(1): 1-340. fold-out map. 1958. [Sedum Rosea, including var. roanensis, 52, maps 33, 34; S. acre, 106, map 87. | _ Flora of Alaska and neighboring territories: a manual of the vascular plants. xxii + 1008 pp. 8 pls. Stanford. 1968. [Includes distribution map for Sedum Rosea, 561, as well as photograph in color. | Humpurtgs, E. C., & J. M. THurston. Bulbil formation in Sedum Telephium. Nature 183: 1343, 1344. 1959. HyLanper, N. Nomenklatorische und systematische Studien uber nordische Gefasspflanzen. Uppsala Univ. Arsskr. 1945(7): 1-337. 1945. [Discussion of Sedum Telephium, 190-192. ] Javas, J. Populationsstudien an Sedum Telephium in Finnland. Ann. Bot. Soc. Fenn. ‘Vanamo’ 26(3): 1-47. 1954. _ & M. T. Ronkkoé. A contribution to the cytotaxonomy of the Sedum Telephium group. Arch. Soc. Zool. Bot. Fenn. ‘Vanamo’ 14: 112-116. 1960. Knosiocu, I. W. Intergeneric hybridization in flowering plants. Taxon 21: 97- 103. 1972. [Lists hybridization of Sedum with Echeveria, Graptopetalum, and Villadia. | Leptcur, A. Embryologie des Crassulacées. Développement de l’embryon chez le Sedum dasyphyllum L. Compt. Rend. Acad. Sci. Paris 246: 1265-1268. Levan, A. Uber das Geschlechtschromosom in Sedum Rhodiola DC. Bot. Not. 1933: 195-197. 1933. [2m = 22; staminate plants reported to have hetero- tén), 2n = 36; S. Rosea; and the status of the Asiatic R. atropurpurea (Turez.) Trautv. & Mey., p. 150. | 224 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Marton, L. The alkaloids of Sedum acre L. Canad. Jour. Res. B. 23: 165, 166. 1945.* Masters, M. T. Hardy stonecrops: sedums. Gard. Chron. 1878: 266-268, 302, 303, 336, 337, 376, 463, 590, 591, 626, 658, 684, 685, TLO5 fi 130) 1o 1s 784 (index of names). 1878. McCormick, J. F., & W. N. Ruswinc. Differential radiation sensitivities of races of Sedum pulchellum Michx.; a useful method of plant identification. Radiation Bot. 4: 247-251. oe L. CrostHwaitE, C. Damon, W. N, J. Frnpiey, V. A. Hicks, R. LEONARD, & W. 4H. Suances SE. he aerate between two closely related species. (Abstr.) ASB Bull. 14: 33. 1967. [S. pusillum and Diamor pha Smallii.] Moran, R. Sedum in Baja California. Cactus Succul. Jour. 41: 20-25. 1969. [Records Sedum alamosanum and S. niveum from this region. | Murpy, W. H. Plant speciation associated with granite outcrop communities the southeastern Piedmont. Rhodora 70: 394-407. 1968. [Sedum pusillum and Diamorpha Smallit (as S. Smaillii), 399-401; includes distribu- tion maps. | Oupa, H. On the genus Sedum in Burma. Jour. Jap. Bot. 50: 353-361, 1975. A revision of the eastern Himalayan species of the subgenus Rhodiola of the genus Sedum per ones ar Pp. 283-362, pls. 10, 11 in H. Onasut, Fl. East. Himalaya, 3rd Rep. Univ. Tokyo Mus. Bull. No. 8. xv + 458 pp. pls. 1-33. 1975, . The Philippine species of Sedum (Crassulaceae). Jour. Jap. Bot. 52: 321-325. 1977, . The taxonomic status of Sedum Telephium and its allied species (Crassulaceae). Bot. Mag. Tokyo 90: 41-56. 1977. [Hylotelephium, gen. nov., includes Sedum telephioides as H. telephioides (Michaux) H. Ohba.] PALAREA, E. R. nonsurgical treatment of opacities of the lens and of the cornea. Ohio Jour. Sci. 54: 114-116. 1954. [Use of Sedum praealtum as a curative for cataracts and nebulae. | PraceT, J. Observations sur le mode d’insertion et la chute des rameaux sec- ondaires chez les Sedum. Candollea 21: 137-239. 1966. Pragcer, R. L. On the affinities of Sedum Praegerianum W. W. Sm., with a tentative classification : a Section Rhodiola. Trans. Bot. Soc. Edinburgh 17: 107-119. pls. 2-4. Notes on Sedum. Fann Bot. 55: 211-215. 1917. [Discussion of S. Nevii and S. Beyrichianum.] . An account of the genus Sedum as found in cultivation. Jour. Royal Hort. Soc. 46: 1-314. 1921. [Includes many helpful line drawin Proner, M. Sur la présence de l’heptose dans quelques espéces d’orpin (Shaun L.). Bull. Sci. Pharm. 43: 7-14, RAputescu, D. CercetXri palinologice referitoare la speciile de Crassulaceae spontane in R. P. R. (In Rumanian; French sum mary.) Acta Bot. Horti Bucuresti 1961-1962(1): 423-433. 1963. [Pollen morphology of Rumanian species of Sedum and Sempervivum. | Rypperc, P. A. Sedum ternatum: American white stonecrop. Addisonia 13: 21, 22. pl. 427. 1928. SARATIKOV, A. S., E. A, KRASNOV, L. A. Kunyxina, & L. M. Dusipzon. Isola- Rosea and R. quadrifida. (In Russian.) Izv. Sibersk. Otd. Akad. Nauk SSSR. Ser. Biol.-Med. Nauk 5: 54-60. 1967.* 1978] SPONGBERG, CRASSULACEAE 225 SHarp, A. J., R. E. SHanxs, H. L. SHERMAN, & D. H. Norris. A preliminary checklist of the dicots of Tennessee. 114 pp. Nashville. 1960. [Records Sedum Telephium as persisting after cultivation in Knox and Sevier counties. SHerwin, P. A., & R. L. Wirsur. The contributions of floral anatomy to the generic placement of Diamorpha Smallii and Sedum pusillum. Jour. Elisha Mitchell Sci. Soc. 87: 103-114. 1971. SmitH, H. E. Polyploidy in Sedum pulchellum — II. Stomatal size and fre- quency. Bull. Torrey Bot. Club 70: 261-264. 1943 Sedum pulchellum: a physiological and morphological comparison of diploid, tetraploid, and hexaploid races. Ibid. 73: 495-541. 1946. Sorpa, T. A cytological study on the genus Sedum, with remarks on the chromosome numbers of some related plants. Jour. Fac. Sci. Hokkaido Univ. Bot. 5: 221-231. 1944. SoxoLovsKAyA, A. P., & O. S. STRELKOVA. Geographical distribution of the polyploid species of plants in the Eurasiatic Arctic. (In Russian; English summary.) Bot. Zhur. 45: 369-381. 1960. [Sedum Rosea, 2n = 18. Sources, R. Développement de l’embryon chez le Sedum acre L. Bull. Soc. Bot. France 74: 234-251. 1927. [See also Compt. Rend. Acad. Sci. Paris 181: 521, 522. 1925.] _ Modifications au tableau récapitulatif des lois de développement chez Sedum acre L. Le type embryonomique de cette espece chez les autres Crassulacées. /bid. 83: 13-18. 1936. SpracuE, T. A. Sedum Rosea. Jour. Bot. 77: 126. 1939. [Notes on the specific epithet Rosea. | STEYERMARK, J. A. Rediscovering Sedum pulchellum at its northeastern limit. Rhodora 44: 73-76. 1942. [Distribution in Missouri and elsewhere in eastern North America. | STRASBURGER, E. Ein Beitrag zur Entwicklungsgeschichte der Spaltoffnungen. Jahrb. Wiss. Bot. 5: 297-342. pls. 35-42. 1866-1867. SUBRAMANYAM, K. Morphological studies of some species of Sedum. I. Floral anatomy. Am. Jour. Bot. 42: 850-855. 1955. _ Embryology of Sedum ternatum Michx. Jour. Indian Bot. Soc. 42A (Maheshwari Commemoration Volume): 259-275. 1963. Torpert, N. E., & L. P. Zixv. Isolation of carbon-14-labeled sedoheptulose and other products from Sedum spectabile. PI. Physiol. 29: 288-292. 1954. Unt, C. H. Heteroploidy in Sedum Rosea (L.) Scop. Evolution 6: 81-86. P52. —— Chromosomes of Sedum in the western United States. (Abstr.) Am. Jour. Bot. 49: 664. 1962. _ Chromosomes of Sedum cremnophila, Echeveria linguaefolia, and their hybrids (Crassulaceae) (Abstr.) Am. Jour. Bot. 54: 641. 1967. Heteroploidy in Sedum glaucophyllum. Rhodora 72: 460-479. 1970 [1971]. _ Intraspecific variation in chromosomes of Sedum in the southwestern United States. Rhodora 74: 301-320. 1972. - Chromosomes of Mexican Sedum. (Abstr.) Am. Jour. Bot. 61(5, suppl.): 51. 1974. _ Cytogeography of Sedum lanceolatum and its relatives. Rhodora 79: 95-114. 1977. 226 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 & R. Moran. Chromosomes of Crassulaceae from Japan and South Korea. Cytologia 37: 59-81. 1972. [Includes chromosome data for Sedum, Meterostachys, and Orostachys. | WattHER, E. Notes on Sedum. Cactus Succul. Jour. 2: 455-457. 1931. [Estab- lishment of Sedum sect. Bergerosedum. | ———. Sedeveria, a new bigeneric hybrid. /bid. 25: 20, 21. 1953. Wess, D. A. What is the type of Sedum Telephium L.? Feddes Repert. Sp. Nov. 64: 18, 19. 1961. [The combination proposed by Webb, viz. Sedum Telephium subsp. Telephium var. purpurascens (K. Koch) D. A. Webb, is superfluous, since the purple-flowered variety of the type subspecies of S. Telephium is var. purpureum L.] ———. Sedum and Rhodiola. In: T. G. Tutin, V. H. Heywoon, et al., eds., Fl. Europea 1: 356-364. 1964. Wuerry, E. T. The sedums of the eastern United States. Gard. Chron. Am. 38: 264-266. 1934. . The ranges of our eastern parnassias and sedums. Bartonia 17: 17-20. 1936. Wiens, D., & D. K. Hatteck. Chromosome numbers in Rocky Mountain plants. I. Bot. Not. 115: 455-464. 1962. [Sedum Rosea, n = 16: S. rhodanthum, i Woopwarp, F. I. The climatic control of the altitudinal distribution of Sedum Rosea (L.) Scop. and S. Telephium L., 2. The analysis of plant growth in controlled environments. New Phytol. 74: 335-348. 1975, & C. D. Picorr. The climatic control of altitudinal distribution of Sedum Rosea (L.) Scop. and S. Telephium L., 1. Field observations. Jbid. 323-334, YarBROUGH, J. A. Regeneration in the foliage leaf of Sedum. Am. Jour. Bot. 23: 303-307. 1936. ZARDINI, E. M. Las especies del género Sedum (Crassulaceae) espontaneas en la Republica Argentina. Bol. Soc. Argent. Bot. 14: 95-106. 1971. ZHUKOVA, P. G. The caryological characterization of some plants of the Chukotsk Peninsula. (In Russian.) Bot. Zhur. 50: 1001-1004. 1965. [Reports 27 = 18 for Rhodiola atropurpureum (Turcz.) Trautv. & Mey. (= Sedum Rosea subsp. integrifolium sensu Hultén).] 2. Diamorpha Nuttall, Gen. N. Am. Pl. 1: 293. 1818. Diminutive, characteristically red-pigmented winter annuals from slen- der primary roots and well-developed, fibrous secondary roots; plants developing ill-defined, over-wintering basal rosettes in fall, the erect or ascending, branched or unbranched stems developing in spring. Leaves alternate, sessile, cylindric-ovoid or cylindric-oblong, succulent, with entire margins and rounded apices (often appearing acute when dried). Flowers perfect, on short, distally thickened pedicels, arranged in leafy, compound cymose inflorescences or variously reduced to a simple dichasium or a single terminal flower. Sepals 4, very small, deltoid or deltoid-linear, per- sistent in fruit. Petals 4, alternate with the sepals, white or pinkish, strongly cucullate, the 4 opposite anthers held in small distal pockets until late in anthesis. Stamens 8, obdiplostemonous; filaments tapering gradually to the basifixed, reddish anthers. Gynoecium of 4 (rarely 5) carpels; carpels 1978] SPONGBERG, CRASSULACEAE ya} united basally for ca. 1/3 their length, free above, each constricted into a short, beaklike style, the styles terminated by small capitate stigmas, ap- parently shriveling after pollination; ovary superior, 4- (rarely 5-) locular in the syncarpous region, each carpel 1-locular in the apocarpous region; ovules numerous on axile placentae below, pendulous on marginal placentae above. Carpels divergent, becoming obliquely oriented in fruit, producing a 4- (or 5-) beaked follicetum, each follicle dehiscing extrorsely by a tear- shaped, flaplike valve. Seeds several in each follicle, small, pyriform; seed coat reddish brown, finely granulate-striate; embryo large, with rounded cotyledons. Base chromosome number 9. TyPr specigs: D. pusilla sensu Nutt. non Michaux = D. Smailii Britton (cf. Wilbur, 1977). (Name from Greek, di, two, and morphe, form or shape, possibly in reference to the initial confusion of the type species with Til/aea or, more likely, Sedum pusillum; Baldwin (1940) interpreted the name as signifying “deformed or contrary formed.”) — ELF-ORPINE. An arresting monotypic genus endemic to the southeastern United States in North and South Carolina, Georgia, Alabama, and southeastern Ten- nessee. Growing in dense, usually pure populations in the shallow, humus- free soils of depression pits on exposed rock surfaces, Diamorpha Smallii Britton (Tillaea ? cymosa Nutt., Diamorpha cymosa (Nutt.) Britton ex Small, D. pusilla (Michaux) Nutt., Sedum cymosum (Nutt.) Frod., S. cymosum var. Smallii (Britton) Fréd., S. Smallii (Britton) Ahles), 2” = 18, is a conspicuous and consistent member of the vernal flora of granitic flat-rock communities of the upper and lower Piedmont of the Carolinas, Georgia, and eastern Alabama (McVaugh, 1943). Not totally restricted to granitic rock, D. Smailii also occurs in similar situations on sandstone outcrops and on sandy flats in Georgia, on the Cumberland Plateau of southeastern Tennessee, and in the mountainous regions of northeastern and central Alabama. When Diamorpha Smallii was first discovered, apparently at Flat Rock, north of Camden, South Carolina, the type locality of Sedum pusillum Michaux, Nuttall confused plants of the two taxa, and, although the initial confusion was later clarified by Asa Gray (1876), the nomenclature of Diamorpha has been troublesome and further complicated by controversy over the acceptance of Diamorpha as a genus distinct from Sedum. As- pects of nomenclature have been clarified only recently by Wilbur (1964, 1977), who has applied the correct specific epithet, Salli, to the single species. The name Diamorpha Smallii was originally proposed by Britton for a reputed second species (treated as a variety by Fréderstrém, 1935) based on immature plants that do not differ from other plants of Diamor- pha of an equal developmental stage. See Wilbur’s papers for the rationale behind the adoption of Britton’s epithet, as well as for historical and other nomenclatural details. The controversy over the distinctness of Diamorpha as a genus has undoubtedly stemmed from the overall morphological resemblance of D. Smallii and Sedum pusillum, as well as from their almost identical distri- 228 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 bution and close ecological association. Moreover, McCormick (in McCormick & Platt, 1964) reported natural and artificial hybrids between Diamorpha and Sedum (presumably S. pusillum), but such hybrids have not been documented (cf. Sherwin & Wilbur, 1971). Diamorpha has several distinctive morphological features, anatomical differences, and a chromosome number uncommon in Sedum, all of which substantiate generic status. Diamorpha is unique in the dehiscence of its follicles by flaplike abaxial valves (F1GuRE 2, e), a type not found else- where in the Crassulaceae. The absence, except during developmental stages, of an adaxial suture line on the carpels and the coherence of the carpels at the base for about one-third their length further separate Diamor- pha from Sedum in general, while its cucullate petals, its floral vasculari- zation pattern (Sherwin & Wilbur, 1971), its usually deeper red pigmenta- tion, and its different ecological niche separate D. Smallii from S. pusillum in particular. As a winter annual rarely more than 10 cm. tall, Diamorpha Smallii has attracted the attention of several investigators (McCranie, McVaugh, Wiggs & Platt, McCormick & Platt, McCormick ef al., Sharitz, Baskin & Baskin, and Sharitz & McCormick) who have studied its life cycle and aspects of its ecology under both field and laboratory conditions to de- termine its specific adaptations to its unique habitat. As a result, Diamor- pha is one of the biologically better-known taxa in the southeastern flora. Seeds of Diamorpha germinate during late October and early November, a period of autumn rains and warm daytime temperatures (ca. 20° C.). FicurE 2. Diamorpha. a-g, D. Smallii: a, habit of flowering plant, x 114; b, flower, X 8; c, longitudinal section through immature follicles (note nectaries at base of staminal filaments) showing pendulous ovules, X 10; d, cross section through four carpels of gynoecium in basal, syncarpous region, 10; e, dehisced follicles, X 6; f, seed, X 30; g, embryo, x 25. 1978 | SPONGBERG, CRASSULACEAE 229 Although germination occurs readily under these conditions, seedling es- tablishment is dependent on extensive secondary root development that, in turn, requires a substrate pH in the narrow range of 4.5 to 5. During establishment, plants can tolerate three or four weeks of total inundation or two to three weeks of desiccation before dying. If established, the plants assume a compact, rosette-like form, develop their characteristic deep red pigmentation, and become semidormant, with little further vegetative growth. Resistant to frost, the plants resume growth and initiate flower- bud primordia during late February. Branching and height of plants are determined at this time and are correlated with soil factors and competition, the latter of which is in turn related to population density. The inflo- rescences expand during March, and, depending upon locality, anthesis begins during late March and early April, reaching its peak toward the middle of April. Details of the pollination ecology and breeding system have not been studied in great detail, but it has been noted that the four anthers opposite the sepals dehisce at the onset of anthesis, while the remaining four, which are held by the four petals, have a retarded dehiscence, shedding pollen toward the end of flowering. Bees and other unidentified insects have been observed visiting the flowers, and cross-pollination is to be expected. Death of plants within a population begins by the middle of April, and by the middle of May most of them are dead. According to Wilbur, seed release from the mature follicles occurs by late spring and early summer, an observation contrary to Wiggs & Platt’s finding that follicle dehiscence does not occur until late summer or fall, “until the continued action of mois- ture . . .” causes dehiscence, “a process requiring 2-5 months.” Accord- ing to Wiggs & Platt, retention of the seeds in the follicles above the out- crop surface over summer, during which time the seeds after-ripen, is an adaptive mechanism that protects the seeds from exposure to the high summer temperatures of the outcrop surfaces. Observations of a series of populations of Diamorpha both in the field and in transplant studies on a simulated outcrop led McCormick & Platt to conclude that “Diamorpha is undergoing ecotypic variation,” inasmuch as clinal phenological variation on a northeast to southwest gradient was detected and physiological differences between populations, related pri- marily to moisture gradients, were found. They summarized their findings and those of other investigators, saying that Diamorpha “has become adapted to the rigorous outcrop environment, both through drought evasion and drought tolerance. All stages of the life-cycle express unusual toler- ances to extremely low or high moisture levels, and the entire life-cycle is adapted to make maximum use of favorable intensities and durations of moisture and to avoid periods of low moisture and high temperature.” The relationships of Diamorpha would appear at first glance to be with annual species of Sedum, yet certain evidence does not support this hypoth- esis. Fréderstrém (1935) allied Diamorpha (as Sedum cymosum) with S, pusillum, 2n = 8, and S. Nuttallianum Raf., 2n = 10, in his group 8, EPETEIUM AMERICANUM, and suggested affinities of this group with annual 230 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 European and Asiatic species. Baldwin (1940), like Berger (1930), main- tained the generic status of Diamorpha and proposed that the genus might represent “amphidiploid results of fusion between representatives of the 4- and 5-chromosome tendencies” found within annual species of Sedum. However, Baldwin also suggested the possibilities of affinities with Tillaea of subfam. Crassuloideae. The anatomical studies of floral vascularization conducted by Sherwin & Wilbur (1971) have shown, however, that Froéderstrém’s group 8 is “highly heterogeneous.’ The differences between the taxa were seen by Britton & Rose to be sufficient to place the three species in different genera, viz., Diamorpha, Tetrorum (to include Sedum pusillum), and Sedum. The rather enigmatic and apparently isolated position of Diamor pha within the Crassulaceae was recognized by De Candolle (1828), who established the tribe Crassulaceae Anomalae to include Diamorpha and Penthorum L Torrey & Gray (1840) also grouped these two genera together in the crassulaceous tribe Diamorphae, a tribe maintained by Small (1933) for Diamorpha alone. It is probable that Diamorpha and the taxa to which it has been allied share superficial resemblances as a result of similar selection pressures and represent convergent groups within the Crassulaceae. In a family notorious for indistinct generic limits, the distinctiveness of Dia- mor pha is probably indicative of relatively great age. Although no local uses have been recorded for Diamorpha, the plant has been important as a subject of continued biologic investigation. With the current knowledge of its life cycle, Diamorpha Smallii will undoubtedly continue to be valuable as an easily manipulated experimental plant. REFERENCES: Under family references see BALDWrn, 1940; BERGER; BRITTON & Rose, 1905; DE CANDOLLE; QuimBY; RICKETT; SCHONLAND, 1890; and SMALL. Baskin, J. M., & C. C. Basxry. Germination characteristics of Diamor pha cymosa seeds and an ecological interpretation. Oecologia 10: 17-28. 1972. GATTINGER, A. The flora of Tennessee and a philosophy of botany. 296 pp. Nashville. 1901. [Records Diamorpha, p. 90, as abundant near Sewanee and on Lookout Mountain, covering the surface in “Rock City.’ Gray, A. Miscellaneous botanical contributions. Proc. Am. Acad. Arts Sci. 11: 71-104. 1876. Hay, J. D., & H. L. Racspate. Estimation of successional trends in Diamor pha systems on granite outcrops. (Abstr.) ASB Bull. 20: 58. 19 : , & D. J. Suure. Dynamics of Diamorpha systems on granite outcrops. (Abstr.) ASB Bull. 19: 73. 1972. Lammers, W. T. Photosynthetic studies of three granite outcrop endemics: Amphianthus pusillus Torr., Isoétes melanospora Engelm., and Diamorpha cymosa (Nutt.) Britton. (Abstr.) ASB Bull. 5: 12, 1958. . A study of certain environmental and physiological factors influencing the adaptation of three granite outcrop endemics: Amphianthus pusillus Torr., /soétes melanospora Engelm., and Diamorpha cymosa (Nutt.) Brit- ton. Diss. Abstr. 19(12): 3096. 1959. 1978 | SPONGBERG, CRASSULACEAE 231 ae eo J. F., L. Crostuwarte, C. Damon, W. Ecan, J. FInpiey, V. A. Hicks, R. ae & W. H. RitrenHouse. Niche differentiation between two ae related species. (Abstr.) ASB Bull. 14: 33. 1967. [Diamorpha Smallii and Sedum pusillum. | & R. B. Ecotypic differentiation in Diamorpha cymosa. Bot. Gaz. 125: 271-279. 1964. [See abstract, ASB Bull. 10: 33. 1963.] McCranig, E. J. Studies in the life cycle of Diamorpha cymosa. 32 pp. M. S. thesis, Emory Univ., Atlanta. 1938.* McVaucu, R. The vegetation of the granitic flat-rocks of the southeastern United States. Ecol. Monogr. 13: 119-166. Morpy, W. H. Plant speciation associated with eanite outcrop communities of the southeastern Piedmont. Rhodora 70: 394-407. 1968. [Diamorpha Smallii and Sedum pusillum, including distribution maps, 399-401. | O’ConnELL, J. E. Cytology, morphology, and taxonomy of Diamorpha cymosa. Jour. Elisha Mitchell Sci. Soc. 65: 194. 1949. Suaritz, R. R. Population dynamics of two plant species in granite outcrop communities. (Abstr.) ASB Bull. 18: 55. 1971. ; McCormick. Population dynamics of two competing annual plant species. aes y 54: 723-740. 1973. [Diamorpha and ues Ga uniflor SHERWIN, P. A., & R. L. Wisur. The contributions of floral anatomy to the generic placement of Diamorpha Smallii and Sedum pusillum. Jour. Elisha Mitchell Sci. Soc. 87: 103-114. 1971. Torrey, J., & A. Gray. A flora of North America. Vol. I. xiv + 711 pp. New York. 1838-1840. [Crassulaceae, including tribe Diamorphae, 556-562, published in 1840. | Wiccs, D. N., & R. B. Piatt. Ecology of Diamorpha cymosd. Ecology 43: 654- 670. 1962. [See also Wiccs, Diss. Abstr. 19: 2443. Wiipur, R. L. Notes on the genus Diamorpha (Crassulaceae). Rhodora 66: 87-92. 1964. Nomenclatural notes on the genus Diamorpha (Crassulaceae). Taxon 26: 137-139. 1977 Subfam. CRASSULOIDEAE 3. Crassula Linnaeus, Sp. Pl. 1: 282. 1753; Gen. Pl. ed. 5. 136. 1754. A diverse, polymorphic genus of diminutive [to large], prostrate to erect, aquatic or semi-aquatic [to xeromorphic], herbaceous [to sub- shrubby], branched [or unbranched], mostly succulent annuals, biennials [or monocarpic or polycarpic perennials], usually from fibrous roots [or occasionally from tuberous rhizomes], the plants of mosslike aspect, often rooting at the nodes. Leaves opposite, small and inconspicuous [or usually of moderate size, or rarely lacking and the internodes swollen], sessile [to distinctly petiolate], the internodes elongate [or often reduced in length and the leaves clustered and subrosulate, variously imbricate, or decussate and 4-ranked, the arrangement and shape of the leaves often giving the plants a Lycopodium- -like or bizarrely geometric appearance], the leaf bases often connate and sheathing, the blades usually glabrous [or variously pubescent, sometimes with structural ‘ ‘windows” and sub- 2a2 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, 59 cuticular air bladders], with entire [or crenate or serrulate] margins. Flowers usually small, 3-, 4-, [or 5-, rarely 6-9-] merous, solitary [or some- times clustered] in the axils of leaves [or terminal, usually arranged in terminal or axillary corymbose, thyrsoid, or subumbellate inflorescences, occasionally with bulbils or plantlets replacing the flowers]. Sepals small, greenish, shorter than the petals, + erect [or flared], often connate at base. Petals small, [erect to] spreading, of various shapes, white [to pinkish or sometimes yellowish], greenish [or bluish], usually connate at base, [often with a small subapical soe: Stamens the same number as the petals, insertion hypogynous [or epipetalous on the shallow corolla tube], opposite the sepals, the oe subulate to flattened or filiform, with ovate to oblong, basifixed anthers [sometimes with conspicuous con- nectives|. Gynoecium of 3-5 or rarely more, erect [or divergent], free or basally connate carpels, the carpels tapering [gradually] or abruptly to short and thick [or long and slender] styles, the stigmas small, terminal [or sometimes subapical] ; ovaries unilocular, with 1, 2, or several to many ovules on placentae along the adaxial sutures. Fruits erect [or divergent] follicles, often with short stylar beaks, dehiscent along the adaxial sutures: seeds 1 to many, small, the embryo surrounded by endosperm. Base chrom- osome numbers 7, 8. (Including Bulliarda DC., Hydrophila House, Tillaea L., Tillaeastrum Britton.) Lectotype species: C. perfoliata L.: see Hitch- cock, Prop. Brit. Bot. 143. 1929. (Name from Latin, crassus, thickish, in reference to the thick, succulent leaves and stems of many of the species.) — CRASSULA. A large, polymorphic, and ecologically diverse genus of nearly 300 species restricted in distribution, with the exception of members of one of the seven sections, to southern Arabia and to sub-Saharan Africa. The genus has its center of diversity, both in number of species and in morpho- logical variation, in southern Africa. Unlike the species of sections re- stricted to Africa, members of sect. Trrtarotprar DC. emend. Schonl., to which the following discussion is largely limited, are cosmopolitan in distribution, occurring on all continents except Antarctica. They are, moreover, the only representatives of the Crassulaceae in southern South America, New Zealand, and Australia. In the southeastern United States sect. TILLAEOIEAE is represented by Crassula aquatica (L.) Schonl. (Tillaea aquatica L., T. simplex Nutt., Tillacastrum aquaticum (L.) Britton, Hydrophila aquatica House, Bulliar- da aquatica (L.) DC., Elatine tetrandra Maxim.; including Crassula Drum- mondi (Torrey & Gray) Fedde), water pigmy, 2” = 42, which has a documented distribution only in Louisiana. It may also occur in North Carolina,® and Hultén (1958) maps a locality for the species in the pan- handle of Florida. Elsewhere in North America it occurs along the eastern "One specimen of Crassula aquatica in the Gray Herbarium has been considered an M. A. Curtis collection from North Carolina, but the collection data do not appear to date from Curtis’s time. Radford, Ahles, & Bell (1968) do not report the species for the Carolinas. 1978 | SPONGBERG, CRASSULACEAE VE) seaboard from Quebec and the Canadian Maritime Provinces, southward into New England, New York, Delaware, and Maryland, and along the Pacific coast in Alaska, Washington, Oregon, and California. Fassett (1928, p. 106) suggests that in northeastern North America C. aquatica is probably a preglacial relic “. . . which has found estuarine conditions favorable for its existence.” Inland, it has been collected from scattered, disjunct localities in the Canadian Northwest Territories (Mackenzie District), and in Minnesota, Nevada, Utah, Colorado, Wyoming, and New Mexico. Beyond North America C. aquatica has a wide, disjunct distribution in Asia and northern Europe. A minute, glabrous annual or biennial with prostrate and nodally root- ing to ascending or erect, usually branched stems, Crassula aquatica is tolerant of both fresh and brackish water. In the Northeast, it is most often found growing in estuarine situations or in the mud or wet sand of receding pond margins where it often forms mosslike mounds. In northern Louisiana, where the species has been found to be relatively common (Thomas, 1971), it is usually found growing in the moist depressions of animal or vehicle tracks in pastures and old fields. Due to its insignificant stature, it is probably often overlooked by collectors, and Thomas has classified the species as a “‘belly-plant.” The small, usually four-merous, greenish or white flowers are solitary in the axils of the connate-sheathing leaf blades and are borne on short pedicels that usually elongate in fruit. The variable pedicel length of flowering plants of Crassula aquatica and the tendency for the pedicels to elongate after anthesis have led both to confusion between species and to the recognition of a segregate species. Plants from Louisiana to Mexico and from California and Washington with the pedicels longer than the subtending leaves in fruiting specimens have been segregated as C. Drummondii (Torrey & Gray) Fedde (Tillaea Drummondii Torrey & Gray, Tillaeastrum Drummondii (Torrey & Gray) Britton, Tillaea aquatica var. Drummondii (Torrey & Gray) Jepson), but such segregation is artificial when the range of pedical length within C. aquatica is considered, and most authors now include C. Drummondi within C. aquatica. Other plants from Prince Edward Island with the flowers distinctly pedicellate were originally confused with C. Vaillantit (Willd.) Roth, a species of Central Europe and northern Africa. While there has been confusion about the other species of Crassula sect. TILLAEOIDEAE that occur in North America, it appears that only two additional ones have been found in the United States. Britton & Rose (1905) included two others as occurring in Mexico. Crassula erecta (Hooker & Arnott) Berger (Tillaca erecta Hooker & Arnott, C. minima Miers), sand pigmy, 2” = ca. 16, ca. 20-25, a species of dry, open loca- tions, is native from southern Oregon southward, through California and Arizona, into Baja California; it also occurs disjunctly in Chile. Crassula muscosa (L.) Roth (Tillaea muscosa L.), a European species, has been reported from Amador and Calaveras counties, California. It has ap- parently become naturalized in the Great Valley of California and in an area in the southern North Coast Ranges (Clausen, 1975, p. 608). 234 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Ficure 3. Crassula. a~l, C. aquatica: a, b, two plants, x 3; c, tip of shoot with axillary flower, < 6; d. flower, & 25; e, flower with petals and one stamen removed (note nectaries at base of carpels), « 25: f, longitudinal section through flower, two carpels removed, * 25; g, flower after anthesis, X 25; h, cross section through flower after anthesis, x 25; i, immature rollaches, 30-2574. dehisced follicles, & 20; k, seed, X 50; 1, embryo, X 50. The union of Tillaea L. with Crassula has not been universally accepted. Authors of many floristic works have retained T illaea, and some have divided Tillaea itself into small genera. Britton (1903), for example, established Tillaeastrum for species (C. aquatica) with solitary flowers and several-ovuled carpels, retaining Tillaea, sensu stricto, for species with clustered flowers and with carpels containing one or two ovules (e.g., C. erecta). However, as defined by Schonland, sect. TILLAEOIDEAE, com- prising several groups or series,” combines these and other species that had otherwise been treated as segregate genera or sections of Crassula. Segregate taxa have been based primarily on the number of floral parts, ovule number, and nectar-scale shape. But, as Schonland (1916, p. 42) notes, these characters are “useless for generic distinction, as they may "Schonland gave no rank to his subdivisions of sect. TILLAEOIDEAE; Jacobsen (1970) has designated them as series. 1978] SPONGBERG, CRASSULACEAE 235 separate species which are otherwise closely allied and moreover they are sometimes not constant in one and the same species.” As a group, species of sect. TILLAEOIDEAE consist of small, morpholog- ically variable, annual, biennial, or perennial plants adapted to wet soils (some are true aquatics), while others are decided xerophytes. The petals, which are usually spreading at anthesis, never form an urceolate corolla (a condition Schénland considers to be a xerophytic adaptation), and they lack the subapical mucro characteristic of Crassula sensu stricto. Despite these deviations, Schénland states that ‘no sharp line can be drawn be- tween the Tillaeoideae and other sections of Crassula.” He suggested further that species of sect. TILLAEOIDEAE are the least derived species of Crassula and that a great age for the section could be deduced from the widespread distribution of its species. Froderstrom (1930, 1931) went so far as to suggest that Tillaea, which he maintained as a genus, was ancestral to the entire Crassulaceae. But Berger (1930) cautioned that species of the TILLAEOIDEAE are probably highly derived, reduced forms, an opinion shared by Uhl (1948), who suggested that the wide distribu- tion of the section (and of individual species) is attributable to very efficient seed dispersal. Quimby (1971), who studied the floral anatomy of C. aquatica, also suggests that Tillaea, as a genus, could be considered a reduced form of Crassula. Among other genera of the Crassulaceae, Schonland favored a possible origin of Crassula from the usually obdiplostemonous genus Sedum. Some species of Sedum sect. EmpETEIUM Boiss. (Sedum sect. Procrassula (Griseb.) Schénl., pro parte, Procrassula Griseb.) are haplostemonous, and, except for their alternate leaves, are suggestive of Crassula, The idea of an origin of Crassula from Sedum is given some support by Mauritzon (1933), who, on the basis of embryological data, concluded that Crassula (and Tillaea) is evolutionarily younger than Sedum. Quimby (1971) states that floral anatomical evidence, i.e., the single whorl of stamens and the lack of or small size of certain carpel traces, also supports this hypothesis. Interspecific hybrids in Crassuda appear to be very rare. However, the presence of an extensive polyploid series suggests reticulate relationships among the species. That sect. TILLAEOIDEAE has a base chromosome num- ber of x — 8, as was suggested by Baldwin (1936), has been confirmed by the recent investigations of Merxmiiller e¢ al. (1971) and Friedrich (1973); yet the one discordant count for the section is that of 2n = 42 for C. aquatica. Otherwise, somatic numbers range from 16 to ca. 128, all multiples of eight. Most other sections of Crassula have a base chromo- some number of 7 (from the diploid to the decaploid level), but some sections share both base numbers. As noted by Uhl (1956), “the sig- nificance of the different basic chromosome numbers with respect to taxonomy is not yet known,” but Friedrich asserts that ‘polymorphic aggregates, e.g. the Crassula lycopodioides complex, are extremely hetero- geneous with regard to their caryological conditions. It appears that different species have been united often unjustifiedly in such cases.” 236 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 The ease by which plants of Crassula species are propagated vegetatively has enhanced the use of these succulents as house plants. Crassula lyco- podioides Lam., 2n = 16, 32, 48, 64, 96 (sect. TILLAEOIDEAE series Lyco- podioides Schonl.) and C. pyramidalis Thunberg, 2n = 14, 28 (sect. PyRAMIDELLA Harvey), both of South Africa, are but two ei the many species of Crassula available at florist shops. Perhaps the best known species is C. argentea Thunberg (C. portulacea Lam.), 2n = 42, 56 (sect. STELLATAE Schonl.), also of South Africa and commonly known as the jade tree or Chinese rubber tree. Widely grown as a house plant, it is sometimes grown out-of-doors in warmer temperate regions as a perennial ornamental. The nomenclature and taxonomy of these and other species, however, is often confused and open to dispute, a situation due in large measure to the difficulties of preparing adequate, identifiable herbarium specimens of succulent plants. REFERENCES: Under family references see Battery ef al.: BERGER; Borisov; Brirron & Rose, 1903, 1905; Dre CaANpboLLE; D’Hus BERT: HUBER; JACOBSEN; KNUTH; Mavritzon, 1933; Mort: QurmBy; Raprorp, AHLES, & BELL; SCHONLAND, 1890; and UHL, 1948, 1956. Under Sedum see CLAUSEN, a p. 608; FRODER- STROM, 1930, 1931; HULTEN, 1958, p. 266, and 1968, p. 5 BaLpwin, J. T.. Jk. Chromosome numbers in Crassula. _ Genet. 33: 455- 463. 1936. Copy, W. J. A history of Tillaea aquatica (Crassulaceae) in Canada and Alaska. Rhodora 56: 96-101. 1954, Cor, F. W. Some species of Crassula known as the jade tree. Am. Hort. Mag. 46: 50-57. 1967. Fassett, N. C. The vegetation of the estuaries of northeastern North America. Proc. Boston Soc. Nat. Hist. 39: 73-130. pls. 6-15. 1928. [C. aquatica, as Tillaea, ee DIS: de To, FrrepRicH, H. C. Zur Crlotaxonomi der Gattung Crassula. Garcia de Orta Bot. Lisboa 1: ns 66, 107. . Vorarbeiten zu einer Moaseueite der Gattung Crassula L. II. Weitere neue Sippen aus dem westlichen Siidafrika. Mitt. Bot. Staats. Miinchen 11: 323-348. 1974. [See also 3: 585. 1960, 6: 623 . 1967.] GANKIN, R. A new California record for Tillaea muscosa (Crassulaceae). Leafl. West, Bat. 9: 256; 1962, Hara, H. Contributions to the study of variations in the Japanese plants closely related to those of Europe or North America. Part I. Jour, Fac. Sci. Univ Tokyo Bot. 6: 29-96. 1952. [C. (Tillaea) aquatica, 64, Hiccrns, V. Notes on the genus Crassula. Jour. S. Afr. Bot. 24: 107-118. 1958. [See also 25: 83. 1959. ] ———. Crassulas in cultivation. 78 pp. + 8 pls. London. 1964. Love, A., & D. Love. Cytotaxonomical conspectus of the regia flora. Acta Horti Gothob. 20: 65-291. 1956. [C. aquatica, 2n = Maas, E. Die Entwicklung des Bae gia einiger eae Arten. Beih. Bot. Centralbl, 48: 118— 154. pl. 2. 193 MERXMULLER, H., H. C. Frrepricu, & J. n RAU. Cytotaxonomische Untersuch- ungen zur Gattunessteuktie von Crassula. Ann. Naturhist. Mus. Wien 75: 111-119, 1971, 1978 | SPONGBERG, CRASSULACEAE ed] Owen, M. L. Tillaea simplex. Bot. Gaz. 20: 80, 81, 1895. [Crassula (Tillaea) aquatica on Nantucket. | laea in Nantucket. Rhodora 14: 201-204. 1912. Reurovus. L. Sur la formation de périderme chez Crassula falcata. Bull. Soc. Bot. Geneve II. 14: 63-69. 1922. Rost. T. L.. & K. Paterson. The developmental anatomy of adventive plantlets from leaves and leaf segments of Crassula argentea (Crassulaceae). Bot. Gaz. 137: 203-210. 1976. ScHONLAND, S. On the South African species of Crassuda, Linn., sect. Tillaeoi- deae. Schoénl. Ann. Bolus Herb. 2: 41-80. pls. 2-7. 1916. Materials for a critical revision of Crassulaceae. (The South African species of the genus Crassula L. (emend. Schonl.).) Trans. Roy. Soc. S. Afr. 17: 151-293. 1929. Sporer. H. Die Blattanatomie der siidafrikanischen Crassula pyramidalis Thunberg. Ein Beitrag zur Anatomie der Xerophyten. Osterr. Bot. Zeitschr. 65: 81-101. pls. 1, 2. 1915. Sropp, K. Aberrante Dehiszensformen bei Frichten einiger Crassula-Arten. Beitr. Biol. Pflanzen 34: 165-175. 1957. Tuomas, R. D. Collecting vascular plants in the habitat near the ground — or, locating and collecting “belly plants.” Castanea 36: 148, 149. 1971. TOLKEN. H. R. The Linnaean species of Crassula. Jour. S. Afr. Bot. 38: 67-80. 1972. WaLTHER. E. Crassula. Cactus Succul. Jour. 1: 131-135. 1930. Subfam. KALANCHOIDEAE Berger 4. Kalanchoé Adanson, Fam. Pl. 2: 248. 1763. Glabrous or pubescent, [monocarpic or] polycarpic perennial herbs, subshrubs, or shrubs, often of twining or rambling habit, the majority terrestrial from fibrous roots [or some epiphytic], many gemmiparous and viviparous. Leaves opposite and decussate or in whorls of 3, simple or occasionally pinnately compound, the simple leaves sessile to petiolate, the bases of the blades or the petioles often amplexicaul; blades + suc- culent, flattened to + cylindrical with entire, crenate, dentate, or pin- natisect margins, often with adventitious buds in the crenations of the leaf margins producing plantlets. Flowers perfect, usually numerous, erect to pendent in terminal or sometimes axillary, bracteate, paniculate or cymose inflorescences, sometimes with adventitious buds in the axils of the bracts; insertion of the calyx and corolla hypogynous. Sepals 4, weakly to strongly connate, forming a cylindrical, often inflated calyx tube, the 4 free lobes shorter to longer than the tube. Petals 4, violet through red and pink to white, yellow, or greenish, connate for most of their length, forming a cylindrical to + campanulate or + urceolate, usually 4-angled corolla tube, the tube sometimes basally constricted over the gynoecium and the 4 free lobes spreading to recurved. Stamens 8 [or rarely 4], epipetalous in 2 whorls of 4, usually exserted; filaments slender, tapering to the basifixed, yellowish anthers. Gynoecium 4-carpellate, the erect or divergent carpels connate at least basally; carpels constricted above the ovaries into slender, usually included styles, the stigmas small, 238 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 often poorly or not differentiated; ovaries unilocular with numerous ana- tropous ovules on placentae along the adaxial sutures. Fruits erect or divergent, papery to coriaceous follicles dehiscent along the adaxial su- tures; seeds numerous, very small, usually + oblong and with rugose or wrinkled seed coats; embryo straight, small, embedded in endosperm, the megagametophyte of the Polygonum type. Base chromosome numbers 17, 18, 20. (Vereia Andrews, including Bryophyllum Salisbury, Kitchingia Baker.) Typr species: K. laciniata (L.) A. P. de Candolle. (Name apparently derived from the Chinese name for one of the species; cf. W. H. Harvey in Harvey & Sonder, Fl. Capensis 2: 327-380. 1894.) — KALANCHOE, BRYOPHYLLUM. Approximately 125 species in three sections, the large majority endemic to Madagascar, others indigenous to Africa and Socotra, one or two widely distributed in both the Old and New World tropics. In the southeastern United States eight species have been reported as naturalized, all in southern Florida. Lakela & Craighead (1965) included three species in their checklist, while Long & Lakela (1971) considered six. Recent col- lections by Brumbach on Captiva and Sanibel islands, Lee County, in- dicate that one or two additional species have become established, pri- marily on the latter island. Six of the eight species occurring in our area belong to sect. Bryo- PHYLLUM (Salisb.) Boiteau & Mannoni (BryopAyllum Salisb.), a section characterized by opposite or usually whorled leaves, pendent flowers with tubular, often inflated calyces with the lobes shorter to longer than the tube, the corollas often constricted basally with the stamens inserted near the base, and the carpels erect. Plants of many species of this section produce plantlets in the crenations of the leaves. The remaining four naturalized species are members of sect. KALANCHO# and are character- ized by opposite leaves, erect flowers with the calyx of free or basally united sepals, unconstricted corollas with the stamens inserted at or near the middle, and erect and contiguous carpels. In species of sect. KALAN- CHOE, plantlet production is apparently restricted to cut or injured sur- faces. The third section, Krtcutncta (Baker) Baillon (Kitchingia Baker), composed of plants with opposite leaves, pendent to nodding flowers with the calyx lobes about as long as the calyx tube, basally constricted corollas with the stamens inserted near or above the middle, and divergent carpels, is not known to be represented in the southeastern United States. Of the species belonging to sect. BRYOPHYLLUM, Kalanchoé pinnata (Lam.) Persoon (Cotyledon pinnata Lam., Bryophylium calycinum Salisb., B. pinnatum (Lam.) Kurz, Sedum madagascaricum Clus.), 2n = 36, 38(?), 40, is undoubtedly the most widely naturalized species in our area. Often forming dense colonies in waste and disturbed areas, it has been reported from Palm Beach, Hendry, Lee, Collier, Monroe, and Dade counties. Plants of K. pinnata, which can reach two meters in height, are easily recognized by their hollow stems and _ their large paniculate in- florescences of flowers with papery, inflated, greenish white to pinkish or 1978 | SPONGBERG, CRASSULACEAE 239 reddish calyx tubes 2.5-4 cm. long. Initially producing simple leaves with crenate margins, mature plants develop pinnately compound leaves with three or five leaflets. Both types of leaf produce plantlets in the crena- tions of the margins, and it is by this vegetative means that extensive colonies of plants have developed. Widely distributed throughout the tropics and subtropics of both hemispheres, it is probable that much of the wide range of K. pinnata, the nativity of which is uncertain, is due to the activities of man. Flowers of Kalanchoé Gastonis-Bonnieri Hamet & Perrier, 2n = 34, are similar to those of K. pinnata but are smaller, and the calyces are less inflated. The corollas, which extend beyond the calyx tubes, are either yellowish green or reddish, but apparently only the latter color form is represented in our flora. The plants produce large, consistently simple, lanceolate to spatulate, whitish to purplish leaves, sinuate to crenate at the margins. Native to Madagascar, K. Gastonis-Bonnieri has become well established on Sanibel Island, Lee County, Florida. Kalanchoé tubiflora (Harvey) R. Hamet (Bryophyllum tubiflorum Harvey, K. delagoénse Eckl. & Zeyh., B. delagoénse (Eckl. & Zeyh.) H. Schinz: including K. verticillata Scott-Elliot, B. verticillatum (Scott-Elliot ) Berger), 2n = 34, 40, 68, has also been reported as naturalized in several areas of Florida in Martin, Palm Beach, Lee, and Collier counties, where it has been noted growing in waste places and on shell mounds. Widely grown as a pot plant, K. tubiflora is characterized by its distinctive violet- brown-spotted, more-or-less cylindrical leaves that are shallowly grooved on the adaxial surface and notched at the apex where bulbiferous spurs produce an abundance of plantlets. The flowers, which are notable for their relatively short calyces but long, pinkish to red corollas with spreading lobes, are produced in cymose corymbose inflorescences. Native to Mada- gascar, K. tubiflora is probably indigenous to areas of South Africa as well. Also well known as a pot plant, Kalanchoé Daigremontiana Hamet & Perrier (Bryophyllum Daigremontianum (Hamet & Perrier) Berger), 2n — 34, has become established in Palm Beach County and on Sanibel Island, Lee County. It has also been collected at Key West and may be naturalized elsewhere in the Florida Keys. The erect, glabrous, robust plants of this species, which is native to Madagascar, attain one meter in height and produce decussate, petiolate leaves with large, succulent blades that are spotted or blotched reddish brown beneath and are often + auriculate at base. Abundant plantlets are produced by adventitious buds in the notches along the crenate to serrate margins. The pendent, lavender flowers are similar to those of K. tubiflora. Kalanchoé laxiflora Baker (Kitchingia laxiflora Baker, Bryophyllum crenatum Baker, Kalanchoé Tieghemi Hamet, K. crenata Hamet, not Haw.), 2x = 34, also native to Madagascar, has been reported as natural- ized in waste places in Collier, Monroe, and Dade counties. A glabrous perennial to 5 dm., K. laxiflora is easily distinguished by its decussate, petiolate leaves with variable ovate to pandurate blades with crenate margins and + auriculate bases. Like the other species of sect. Bryo- 240 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 59 PHYLLUM, plants of K. laxiflora produce plantlets in the crenations of the leaves. The flowers, with short, + inflated calyces and relatively long, red to rosy- or yellowish-orange corollas, are produced in branched, paniculate inflorescences. Unlike other naturalized species of sect. BRYOPHYLLUM, plants of Kalanchoé Fedtschenkoi Hamet & Perrier, 2n = 34, develop into low, dense, many-branched shrubs with procumbent stems that eventually turn upward. Long, stiff adventitious roots are produced along the procumbent parts of the stems, and the short-petiolate, decussate leaves are closely spaced. The bluish green blades are ovate with rounded apices, and the crenate margins are brownish tinged in the crenations. The flowers. pro- duced in terminal dichasial inflorescences, are brownish pink. Another native of Madagascar, K. Fedtschenkoi is known to be established in our area only on Sanibel Island, Lee County, Florida. Of the two species of sect. KALANCHo# that are naturalized in our area, K. marmorata Baker (K. grandiflora A. Rich., not Wight & Am., K. somatiensis Baker, K. macrantha Baker), 2n — 34, is perhaps the most distinctive. A glabrous shrub usually branched from the base and with erect or sometimes procumbent stems, plants of K. marmorata are easily recognized by their sessile to shortly petiolate leaves with undulating to crenate margins that are green to plum colored and often brownish spotted on both surfaces. The erect flowers, borne in corymbose cymes, are characterized by long, white corolla tubes, 8-10 cm. long, terminated by short, deltoid lobes. Native to Ethiopia and Somalia, it is reported by Long & Lakela (1971) to be established in hammocks and disturbed sites in southern Florida, where it often forms extensive colonies. Kalanchoé crenata (Andr.) Haw. (Veria crenata Andr., K. integra (Medic.) O. Kuntze var. crenata (Andr.) Cufodontis), a shrub with long- petiolate, elongate-lanceolate leaves with doubly dentate margins and yellow flowers produced in glandular-pubescent inflorescences, is the second species of sect. KALANCHO# naturalized in Florida. Long & Lakela (1971) list it as occurring in disturbed sites and hammocks in South Florida, and I have seen a specimen collected in 1973 from Sanibel Island, Lee County. Considerable taxonomic and nomenclatural con- fusion apparently surround K. crenata, and plants of this and closely allied taxa have been given varying taxonomic recognition (cf. Bailey et al., 1976; Jacobsen, 1960, 1970; Baldwin, 1938: and Long & Lakela, 1971). As a result, various names, including K. laciniata (L.) DC., have been applied to our plant, apparently a South African native, and taxonomic and nomenclatural aspects need clarification. Baldwin (1938) found 2n = 34 and 2n = 68 for plants of this complex supposedly corresponding to K. daciniata, sensu lato, and 2n — 102 for plants determined as K. crenata. Much of the taxonomic attention given Kalanchoé has centered around its generic boundaries and the validity and status of Bryophyllum and Kitchingia. Many botanists have followed Baillon (1885), who included Kitchingia as a section of Kalanchoé, and Hamet (1907, 1908), who in- 1978] SPONGBERG, CRASSULACEAE 241 corporated both Bryophyllum and Kitchingia within Kalanchoé as the sole genus of subfam. Kalanchoideae. Hamet placed the species in 14 groups of undesignated rank, while Boiteau & Mannoni (1947-1949) arranged the species in three sections (and numerous subsections) that reflect the originally proposed generic boundaries. Those who have maintained or supported the maintenance of three genera have included Berger (1930), and Tillson (1940), who offered anatomical evidence in support of her position. Tillson found that in all three taxa “. . . the four petal traces are adnate to the corresponding antepetalous stamen trace as it leaves the floral axis.” Despite the level of stamen insertion, however, consistent differences in the level of di- vergence of the two traces were noted. In Bryophyllum the traces separate below the level of corolla insertion, while in both Kalanchoé, sensu stricto, and Kitchingia separation occurs at some level in the corolla. However, the intermediate morphology of several species originally indicated that the three genera should be treated as one, and this position has been supported by interpretations given both embryological (Maurit- zon, 1933) and cytological (Baldwin, 1938) data. More recently, Fried- mann (1971), who has surveyed the chromosome numbers of numerous Madagascan kalanchoés, has cautioned that additional cytological data together with geographical evidence may indicate that two genera, Bryo- phyllum (including Kitchingia) and Kalanchoé, are justified. It is obvious that additional evidence will be required to validate either the continued use of one genus or of two or three genera, and to gain unanimity among botanists. The choice made here to consider subfam. Kalanchoideae as consisting only of Kalanchoé, itself comprising three sections, is pro- visional — a choice of convenience that preserves the identity of the three taxa, yet reflects what are very close and undoubtedly reticulate relation- ships between them. Chromosomes of Kalanchoé are small, yet chromosomal morphology is generally discernible and may be variable within a species complement; heterochromatic fragments have also been observed in several species. Chromosome numbers have been reported to range between 2” = 34 and 2n — ca. 500, but most species are either diploid with 2” = 34 or 36, or tetraploid with 2n = 68 or 72. Kalanchoé crenata is probably a hexaploid with 2n = 102: K. Grandidieri Baillon, 2n = ca. 140, is likely an octo- ploid; and K. Faustii Font Quer is a decaploid with 27 = 170. Multi- valent or secondary associations during meiosis apparently are uncommon. While the majority of species of sect. KaLANCHOE have chromosome numbers based on 18, and most of those of sects. BkyopHyLLUM and KitT- CHINGIA are based on 17, there is incomplete coincidence between base chromosome number and sectional lines. Baldwin (1938) proposed that, of the possible base numbers for the genus, viz. 17, 18, 20, the primary number is 17. His suggestion was based on the hypothesis that Kalanchoé had an origin involving members of subfams. Crassuloideae and Cotyle- donoideae Berger where base chromosome numbers of 7 or 8 and 9, 242 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 respectively, were suspected. Both subfamilies are well represented in the geographical area where Kalanchoé is indigenous. Although the prevalent diploid number found in Kalanchoé is 34, a number based on 17, Uhl (1948) proposed that the Kalanchoideae may have had an origin completely within the Cotyledonoideae, and that, as a result, the primary base chromosome number for Kalanchoé is 18. Uhl’s proposal is strengthened if members of Crassula sect. TILLAEOIDEAE are considered to be derived members of that genus. This interpretation re- quires that, of the two possible base numbers for the Crassuloideae. 7 or 8, seven be taken as the primary number, thereby excluding the great ma- jority of taxa of the Crassuloideae except species of sect. TILLAEOIDEAE as possible ancestors of the Kalanchoideae. Based on morphology alone, this hypothesis seems doubtful, while the separation between the Kalan- choideae and Cotyledonoideae is based solely on four-merous flowers in the former and five-merous flowers in the latter subfamily. The production of plantlets along the margins of leaves by species of Kalanchoé sect. BrvopHYLLUM has attracted the attention of numerous investigators. Stoudt summarized the studies prior to 1938 and pointed out that for all the Crassulaceae investigated, a residual meristem or a secondary “cicatrice’”’ meristem formed from parenchyma cells produces root primordia endogenously and shoot primordia exogenously. Differences in plantlet formation consist of the presence or absence of residual meristems, the location of the meristematic activity, the conditions under which meristematic activity and/or growth of primordia are initiated, and, if residual meristems are present, the degree to which primordia are dc veloped on mature plants. In species of sect. BRYOPHYLLUM, meristematic activity and subsequent growth of plantlets can occur on either attached or detached leaves. In species of sect. KaLancHoé, growth of a new plantlet is apparently induced only when leaves are detached. Moreover, the residual meristematic region in species of sect. KALANCHO# is at the base of the leaf or petiole, not in the crenations of leaf margins. Numerous studies have attempted to determine the factors that trigger or inhibit the activity of the meristems and the growth and development of the primordia, which in species of Kalanchoé are well differentiated when the leaf on which they occur is still immature. The literature con- cerning this aspect of plantlet production is large, and no attempt has been made to summarize it here. It appears, however, that there are con- flicts and contradictions in the literature, and it would be helpful if a plant physiologist could summarize this ieee: to this end, numerous references are included below. Because of the ease of propagation of plantlets and the resulting avail- ability of clonal populations, plants of Kalanchoé have proven to be apt subjects in experimental studies. Much information concerning the genus has accumulated and is scattered in the physiological literature. For ex- ample, it has been determined that Kalanchoé Blossfeldiana Poelln. is a typical short-day plant, while K. crenata, K. tubiflora, and K. Daigre- montiana are long-short-day plants. Van de Pol (1972) has summarized 1978] SPONGBERG, CRASSULACEAE 243 available information concerning floral induction and flowering in Kal- anchoé, and his paper should be consulted for numerous pertinent refer- ences of a physiological nature not cited here. In another study, Groner (1974) has reported that parent plants of Kalanchoé Daigremontiana produce a water-soluble allelopathic agent that greatly retards growth (as well as eliciting a syndrome of related characteristics) in daughter plantlets that become rooted within the radius of the parent plant’s root system. Groner suggests that this influence is an adaptation for the control of both population size and density. In Madagascar, where the species is native, plantlet production is increased during the rainy season, when it is probable that the plantlets are detached and washed to new areas for establishment. An allelopathic effect was also observed on germination and growth of plants of several other species of monocots and dicots, while others were unaffected, and Groner (1974, 1975) suggests that the allelopathic compound is a unique glycoside, bryo- phyllosid, recently characterized by Karsten (1965). No information concerning the breeding system or pollination ecology of Kalanchoe has been located, except for the suggestion that flowers of K. pinnata are adapted for hummingbird pollination (Knuth, 1908). However, Craft (1942) has described the occurrence of extrafloral nectaries on the uppermost leaves and lowermost floral bracts in K. pinnata. These “nectaries” (glandular tissue lying in pitted areas on the lamina) appear and produce droplets high in glucose content only when the plants are in flower. Several hybrids have been reported in Kalanchoé, but all references located refer to hybrids that have arisen either in cultivation or through intentional crosses. A triploid hybrid, 21 = 51, between K. Daigremonti- ana and K. tubiflora, both of sect. BRYopHyLLUM, has been reported by Baldwin (1949) and Warden (1958). Displaying a morphology inter- mediate between the parental species, the hybrid is apparently well estab- lished in American horticulture and is sometimes referred to as K. & Ay- brida Hort. Another hybrid has involved K. Daigremontiana as the pollen parent and K. pinnata as the seed parent, while the intersectional hybrid, K. Daigremontiana « K. Blossfeldiana, and its reciprocal, have been produced by Resende (1956). Maintaining two genera for the parental species, Resende gave the hybrid group the name * Bryo- kalanchoé lisbonensis, apparently an invalid one lacking a Latin descrip- tion. Resende’s ahseryations on these sterile hybrids indicate that “the capacity to form pseudobulbs is recessive and that SD [short-day] is dominant over LSD [long-short-day |.” Aside from their usefulness in experimental studies, several species of Kalanchoé are of economic importance as pot plants. Within recent years numerous cultivars of K. Blossfeldiana have been selected and marketed as winter-blooming house plants, and most florist shops offer this species for sale during the Christmas season and into the spring months. Many other species are widely grown by fanciers of succulent plants and may occasionally be found for sale in florist shops. 244 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 REFERENCES: Under family references see BACKER; BAILEY et al.; BERGER; BRITTON & Rose, 1905; Dr CANDOLLE; JACOBSEN; JENSEN; KNUTH; Lonc & LAKELA; Mavritzon, 1933; Morr; QUIMBY; SCHONLAND, 1890; SHARMA & GHOSH; STovuptT; and UHL, 1948. Baitton, M. H. Liste des plantes de Madagascar. Bull. Mens. Soc. Linn. Paris 1: 465-472. 1885. |Crassulaceae, 467-470; inclusion of Avtchingia as a a4 of Kalanchoe. | eS a AR le the genus and its chromosomes. Am. Jour. Bo ro 38. [Includes key to species common in American ee . Hybrid of Aulanchoé Daigremontiana and K, verticillata. Bull. Torrey Bot. Club 76: 343-345. 1949. | Includes drawing of the hybrid plant. | Baneryt, L. & S. SEN. The somatic and meiotic chromosomes of Kalanchoé pinnata ar Persoon (syn. Bryophvilum calycinum Salisbury). Caryologia 8: 195-204. 1956. Baron, C. M., R. J. D. Grauam, & L. B. STEWART. Vegetative propagation. Kalanchoé verticillata, Trans. Bot. Soc. Edinb. 30: 70 BERGE, A. Beitrage zur Entwicklungsgeschichte von #r vopnallan calycinum. [11 pp. 8 pls. Zurich. 1877 BoITEAU, P. & - MANNONI. Les plantes grasses de i Les i. Cactus 12: 5-10. 1947; 13: 7-10. 1948; 14: 23-28. 1948; 15-16: 37-42. 1948; 17-18: 57, 58. 1948; 19: 9-14. 1949; fs 43-46. 1949; 21: er 1949; 22: 113, 114. 1949 Braun, E. L. Regeneration of Bryophyllum calycinum. Bot. Gaz. 65: 191-193. 1918. Btnsow, R. C. The circadian rhythm of photoperiodic responsiveness in Kalnchot eae clocks. Cold Spring Harbor Symp. Quantit. Biol. 3 260. 19 CHILD, . M. & A. - Bettamy. Physiological isolation by low temperature in Bryophvilum. Bot. Gaz. 70: 249-267. 1920 Cook, O. F. Vegetative propagation of Reveille. a review. Jour. Hered. 15: 463-466. 1924. Crart, J. H. Extra-floral nectaries in Bryophyllum calycinum. Proc. Iowa Acad. Sci. 49: 113-115. [1942] 1943. . Preliminary studies of the physiology and morphology of the germinat- ing foliar embryos of Bryophyilum calycinum. Ibid. 50: 171-179. [1943 1944, DAUPHINE, A. De levolution de l'appareil conducteur dans le genre Kalanchoé. Ann, Sci. Nat. Bot. IX. 15: 153-163. 1912, & R. Hamer. Contribution a l'étude anatomique du genre Kalanchoé. Ibid. 14: 195-219, 1911. Desraux, G., & M. Asti. Influence de la photopériode sur la biologie et la eae des Br ee (Crassulacées). Compt. Rend. Biol. Paris Bee 195 —2069. DostTAL, . Cytokinin ie in integrating Bryophyllum shoots. Beitr. Biol. Pflanzen 47: 259-276. 1971 ENGELMANN, W. Endogene Rhythmik und photoperiodische Bliihinduktion bei Kalanchoé. Planta 55: 496-511. 1960. 1978] SPONGBERG, CRASSULACEAE 245 Ferre, J.. & A. Lesecue. Embryogénie des Crassulacees. Développement de lembryon chez le Kalanchoé verticillata Jacq. Compt. Rend. Acad. Sci. Paris 255: 1462-1464. 1962. FrepEerico, H. The change of the apex of Kalanchoé Blossfeldiana towards in- florescence formation. Biol. Jaarb. 28: 76-87. 1960. FRIEDMANN, F. Sur de nouveaux nombres chromosomiques dans le genre Kalan- choe (Crassulacées) a Madagascar. Candollea 26: 103-107. Formes de croissance et multiplication végétative des Kalanchoé malgaches. Candollea 30: 177-188. 1975. Garnp, K. N., & R. L. Gupta. Alkanes, alkanols, triterpenes and sterols of Kalanchoé pinnata. Phytochem. 11: 1500-1502. 1972. Grecory, F. G., I. Spear, & K. V. THrmann. The interrelation between CO, metabolism and photoperiodism in Kalanchoé. Pl. Physiol. 29: 220-229. Groner, M. G. Developmental variations in asexual plantlets from leaves of Kalanchoé Daigremontiana. (Abstr.) Am. Jour. Bot. 55: 715. 1968. _ Intraspecific allelopathy in Kalanchoé Daigremontiana. Bot. Gaz. 135: 37-39. 1974. Allelopathic influence of Kalanchoé Daigremontiana on other species of plants. /bid. 136: 207-211. 1975. Hamet, R. Monographie du genre Kalanchoé. Bull. Herb. Boiss. II. 7: 869-900. 1907. Ibid. 8: 17-48. 1908. _ Kalanchoé luciae sp. nov. Ibid. 8: 254-257. 1908. [Includes key to the groups of Kalanchoé, p. 257. Observations sur le Kalanchoé tubiflora nom. nov. Beih. Bot. Centralbl. 29: 41-44, 1912 ———.. Sur muelaues Kalanchoé de la flore malgache. Ann. Mus. Col. Marseille III. 3: 123-164. 1915 Sur un groupe dé transition reliant le genre Kalanchoé au genre Cotyledon. Rev. Gén. Bot. 28: 80-84. 1916 ——— & J. Marnier-LaposToLie. Le genre Kalanchoé au jardin botanique “Les Cédres.” Arch. Mus. Hist. Nat. Paris VIII. 8: 1-110. pls. 1-37. 1964. _ PERRIER DE LA BATHIE. Contribution a l'étude des Crassulacées malgaches. Ann. fea Nat. Bot. IX. 16: 361-377. 1912 & Nouvelle contribution a l’étude des Grastulactes malgaches. ae Mus. Col. Marseille III. 2: 113-207. 1914. _ Troisiéme contribution a l’étude des Crassulacées malgaches. i 3: 63-121. 1915. Harper, R. Vegetative and reproductive development of Kalanchoé ae ana a3 influenced by photoperiodism. Soc. Exper. Biol. Symp. 2: 117-138 1948. JADIN, & A. Jurttet. Recherches anatomiques sur trois espéces de Kalan- choé de Madagascar donnant des résines parfumées dans leurs écores. Ann. Mus. Colon. Marseille II. 10: 137-156. pls. 7-10. age P. H. Flowering kalanchoés, House Pl. Porch Gana 2(8): 52-59. 977. [Horticultural aspects of commonly available kalanchoés, particularly i: Blossfeldiana. | Jones, M. B., & T. A. Mansrietp. A circadian rhythm in the level of carbon dioxide éempensation in Bryophyllum Fedtschenkoi with zero values during the transient. Planta 103: 134-146. 1972 246 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Jurrsi¢, J. Beitrag zur Morphologie und Biologie von Bryophyllum Daigre- Se a oe et Perrier de la Bathie) Berger. Anzeiger Akad. Wiss Wien 73: -127. 1936. KAKESITA, Kk. nes on regeneration in Bryophyllum calycinum (preliminary te). Jap. Jour. Bot. 4: 27-35. Experimental studies on regeneration in Bryophyllum calycinum. Ibid. 5: 219-252. 1930. KarsTEN, U. Bryophyllosid, ein neues Kampferolglykosid aus Bryophyllum Daigremontianum (R. Hamet et Perr.) Berg. Naturwissenschaften 52: 84, 85. 1965. Komata, Z. Chromosomes in vegetative hybrids of . a genus. (In Polish, English summary.) Folia Biol. 4: 51- LAKELA, O., & F. C. CRAIGHEAD. ener pres - - a plants of Collier, Dade, and Monroe counties, Florida. viii + 95 pp. Coral Gables. 1965. [Kalanchoé, 41.] Lauzac-MarcuaL, M. Réhabilitation du genre oe Salisb. (Crass- ulacées Kalanchoidées). Compt. Rend. Acad. Sci. Paris 278: 2505-2508. 1974. |Arguments for the re-establishment of aie at generic ank, Lorn, ‘y Rules and mechanism of inhibition and correlation in the regeneration of Bryophyllum calycinum. Bot. Gaz. 60: 249-276. 1915. [See aso Ibid. 62: 293-302. 1916; Ibid. 63: 24-50. 1917; Ibid. 65: 150-174. 1918; Jbid. 66: 69. 1918.] ———. The stimulation of growth. Science 41: 704-715. 1915. [See also Jbid. 44: 210, 211. 1916; Ibid. 45: 436-439. 1917; Ibid. 46: 547-551. 1917; Ibid. 54: 521, 522. 1921.] The law controlling the quantity of regeneration in the stem. Jour. Cen. Physiol. 1: 81-96. 1918. [See also bid. 337-362. 1919; Ibid. 687- 715; Lbid. 2: 297-307. 1920; Ibid. 373-386; Ibid. 651-657; Ibid. 4: 447-461. 1922. | Maton, J. Photoperiodic flowering response of chlorophyll-free Kalanchoé plants. Biol. Jaarb. 29: 84-88. Meue icu, F. P. Factors affecting ornare ny the foliar meristems of Bryo- phyllum calycinum. Bot. Gaz. 92: 113-140. 1931. nee , G. V. gt el crenatum, Madagascar sprouting-leaf. Addisonia 0c. 132, 19 a DE LA Cre . Crassulacées malgaches nouvelles. Bull. Mus. Hist. Nat. Paris 29: 452-455. 1923. . Observations nouvelles sur le genre Kalanchoé. Arch. Bot. Caen Bull. Mea 2: _ 31. 1928. , P. A. vAN bE. Floral induction, floral hormones and flowering. Meded. “Lanab Wageningen 72(9): 1-89. 1972. [Chapter 3, “Experiments with ome Crassulaceae,” 7-36. ee G. W., C. S. LEAveNWortH, W. D. Ginter, & H. B. Vickery. Studies in the metabolism of Crassulacean plants: changes in the composition of Bryophyllum pcan during growth. Pl. Physiol. 22: 1-19. 1947. [See six additional papers in this series in Pl. Physiol. 22: 205-227, 360-376, 477-493: 23: 123- ee 149-151; 24: 610-620 Raapts, E. The genus Kalanchoe (Crassulaceae) in tropical East Africa. Willdenowia 8: 101- 1977. Renicu, M. E. Roe a in Bryophyllum crenatum. Trans. Ill. Acad. Sci. 16: 183-197. 1923. 1978] SPONGBERG, CRASSULACEAE 247 RESENDE, F. Contribution to the physiology of development of the inflorescence and of the single flower (Bryophyllum and Kalanchoé). Portug. Acta Biol. A. 1949/51(Vol. R. B. Goldschmidt): 729-784. 1949-1951. . Hibridos intergenéricos e oo Kalanchoideae. I. Bol. Soc. Portug. Ci. Nat. 2a. 6: 241-244. 1 #l. & H. F. LINSKENS. On hermaphrodite a Here plants in eee re and its segregation. I. Portug. Acta Biol. A. 6: -178. 1960/1961 SALTMAN, P., G. Kunitake, H. Spotter, & C. Stitts. The dark fixation of CO, by succulent leaves: the first oe Pl. Physiol. 31: 464-468. 1956. [Experiments with Kalanchoé pinnata. | ScuwaBe, W. W. Evidence for a flowering inhibitor produced in long days in Kalanchoé Blossfeldiana. Ann. Bot. I. 20: 1-14. 1956 ——_ ects of photoperiod and hormone treatment on isolated rooted leav es of Kalanchoé Blossfeldiana. Physiol. Pl. 11: 225-239. 1958. ——. Flower inhibition in Kalanchoé Blossfeldiana. Bioassay of an endog- enous long-day inhibitor & inhibition by (+) abscisic acid and xanthoxin. Planta 103: 18-23. 1972. Sarma, G. K., & D. B. Dunn. Effect of environment on the cuticular features in Kalanchoé Fedschenkoi. Bull. Torrey Bot. Club 95: 464-473. 1968. & Effects of cool nights on flowering of Kalanchoé Fedschenkot. Trans. Missouri Acad. Sci. 3: 22-28. - SrrONVAL, C. Action of day-length upon the formation of adventitious buds in Bryophyllum tubiflorum Harv. Nature 178: 1357, 1358. 1956. Sopets, J. C. Uber die Entwicklung und Bewurzelung ruhender Meristeme bei Bryophyllum calycinum, B. crenatum und B. proliferum. Rec. Trav. Bot. Néerl. 31: 188-209. 1934. Spear, I. Some aspects of the physiology of caer in Kalanchoé Bloss- feldiana. (Abstr.) Pl. Physiol. 32(suppl.): xi, OST ——. & K. V. Tuimann. The interrelation eae CO, metabolism and photoperiodism in Kalanchoé. II. Effect of prolonged darkness and high temperatures. Pl. Physiol. 29: 414-417. 1954. Sropart, A. K., I. McLaren, & D. R. THomas. Chlorophylls and carotenoids of colourless callus, green callus and leaves of Kalanchoé crenata. Phyto- chem. 6: 1467-1474. 1967. Stoupt, H. N. Gemmipary in Kalanchoé rotundifolia and other Crassulaceae. Am, Jour. Bot. 25: 106-110. 8 SwINcLe, C. F. The easiest plant in the world to propagate. Jour. Hered. 25: 73, 74. 1934.* [Kalanchoé Daigremontiana. | TayLor, W. R. Chromosome morphology in Fritillaria, Alstroemeria, Silphium, and other genera. Am. Jour. Bot. 13: 179-193. 1926. [Kalanchoé pinnata, 2n = 40 or 38(?).] Tittson, A. H. The floral anatomy of the Kalanchoideae. Am. Jour. Bot. 27: 595-600. 1940. Wanput, M., & H. Y. Monan Ram. Morphogenesis in the leaf callus of Kalanchoé pinnata Pers. Phyton Buenos Aires 21: 143-147. 1964. Warpven, J. W. Contribucién para el estudio aay de los hibridos en Kalanchoideae, I. Portug. Acta Biol. A. 5: eS 8. . Variacién intraindividual del numero crom co en el meristema radicular de los hibridos de Bryophyllum calcinum iy ae Daigre- montianum. Bol. Soc. Argent. Bot. 7: 209-213 248 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 59 We cu, W. B. A study of the relation of cicatrization to evaporation from leaves of Bryophyllum. Trans. Ill. Acad. Sci. 32: 78. . Cicatrization in leaves of Bryophyllum calycinum. Bot. Gaz. 107: 95- 106. a Wirtscu, H. v., & A. Frtucer. Uber Polyploidieerhohung im Kurztag bei Kalan- choe Blossfeldiana. Zeitsche. Bot. 40: 281-291 ZIMMER, R. Phasenverschiebung und andere Storlichtwirkungen auf die endogen Tagesperiodischen Bliitenblattbewegungen von Kalanchoé Blossfeldiana. Planta 58: 283-300. 1962. ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 AN JAMAICA PLAIN, MASSACHUSETTS 02130 1978] KENG, PHYLLOCLADUS 249 THE GENUS PHYLLOCLADUS (PHYLLOCLADACEAE) HsuAN KENG THE CONIFEROUS GENUS Phyllocladus (Phyllocladaceae, formerly Podo- carpaceae) was established by L. C. and A. Richard in 1826 based on the Tasmanian species Phyllocladus rhomboidalis L. C. & A. Rich., and is now known as P. aspleniifolius (Labill.) Hooker f. (FIGURES 1-9). This genus is noted particularly for the presence of large, celery-leaf-like photo- synthetic organs (hence the common names “celery pine” or ‘“‘celery-topped pine’) which have been interpreted as phylloclades and from which the ceneric name is derived. The prevailing view of the origin and nature of the phylloclades is still dominated by the following statement made by Robertson (1906, p. 259) over half a century ago: “Phyllocladus is .. . characterized by the reduction of its true leaves to pointed scales, and the expansion of certain of its stem-branches into flattened leaf-like structures.” This implies that the phylloclade is a highly specialized, and therefore an advanced, character. My previous taxonomic and morphological studies on the genus were based mainly on a single Malesian species, Phyllocladus hypophyllus Hooker f. (Keng, 1963a, 1963b). From April to August, 1976, I visited Tasmania and New Zealand, and was able to examine living and preserved material and herbarium specimens of the other known species of Phyllo- cladus. In September and October, 1976, I also visited the Rijksherbarium, Leiden. the Netherlands, and examined a large number of herbarium speci- mens of P. kypophyllus collected from New Guinea and elsewhere in the Malesian regions. The main purpose of this extensive study tour was to re-examine my earlier suggestion (Keng, 1974) that the foliate phylloclade of this genus is a relict feature, possibly derived from or remotely related to the lateral branch systems of the recently established fossil taxon Pro- gymnospermae (or Progymnospermopsida). This was discussed in a paper presented to an I.A.P.T. Symposium in September, 1976, in Hamburg, and has been published elsewhere (Keng, 1977). In this article, the morphology and taxonomy of the genus PAyllocladus are discussed, and brief notes on the evolutionary trends and the paleogeographic distribution of the genus are given. MORPHOLOGY Hasir. Species of the genus and, in some cases, plants of the same species, range from shrubs or small trees to large trees. Phyllocladus as pleniifolius var. alpinus (previously known as P. alpinus) is usually a bushy shrub at high altitudes. Its lowland form, however, may become a small tree, reach- ing 8-9 meters. PAyllocladus glaucus and P. as pleniifolius var. as pleniifolius are small to medium-sized trees; the former can attain about 12 meters, 250 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 ‘2 a DK) ; LEX CF Oe ye, Figures 1-9. Phyllocladus pclae la var. aspleniifolius : 1, branch bearing whorled simple phylloclades and pollen cones; 2, pinnately lobed phylloclade with term age ud; 3, young branch with simple eae and pollen cones phyll une sen as sacs (in two vie ws. > 2. “DOr me not branch _ bearing ovulate cones; 6, ovulate cone on side of deformed simple Sp de aa 7, young seed cone; 8, branch bearing mature seed cones; 9, mature 1978 | KENG, PHYLLOCLADUS 251 and the latter can reach approximately 18 meters under suitable conditions. The two remaining species, P. trichomanoides and P. hypophyllus, become lofty trees, reaching a height of approximately 25 and 30 meters, respec- tively. Phyllocladus hypophyllus, however, like most of the other species, becomes stunted in subalpine forests. The bark of the species varies from greenish (Phyllocladus trichoman- oides) to reddish brown (P. hypophyllus) or blackish (P. aspleniifolius var. as pleniifolius), and is yellowish or reddish and fibrous within. It is smooth P. trichomanoides) or deeply fissured (e.g., P. aspleniifolius var. as pleniifolius), and is shed in large, thin flakes. Branches and branchlets are borne 3 to 5, rarely 8 or 9, in a false whorl. Trees of Phyllocladus can be readily identified in the field by their apparently whorled branches. Species of Phyllociadus, especially P. aspleniifolius var. alpinus, must be among the slowest growing conifers. A seven-year-old seedling of P. aspleniifolius var. alpinus in its natural habitat was only 6 cm. tall. It was fully covered with linear juvenile leaves, and one of its two withered cotyledons still persisted. A seedling only 12 cm. tall was estimated to be 25 years old. Trunks of P. aspleniifolius var. alpinus hardly exceed 30 cm. in diameter, yet may show up to 190 growth rings (Wardle, 1969). A wild seedling of P. aspleniifolius var. aspleniifolius of unknown age was brought into cultivation at the Botany Department of the University of Tasmania and, after about 15 years (Prof. W. D. Jackson, University of Tasmania, pers. comm.) has reached a height of 1.1 meters and has begun to produce staminate cones. Displayed at the same department is a large disc of wood with a diameter of about 45 cm. (maximum radius 29 cm., minimum radius 20 cm.). This came from a tree that was cut in 1944 and was estimated by carefully counting the annual rings to be 389 years old. PuytiocLapes (Cladodes). The unique “foliar” organ of the genus is a phylloclade or cladode, which emerges from the axil of an acicular leaf. Phylloclades are alternate, opposite, or more frequently in false whorls on stems and branches. They are thick and leathery, and are extremely vari- able in shape and size, apparently being easily modified by environmental changes; they can also be quite different in juvenile and adult plants. There are two types of phylloclades: simple phylloclades are found in Phyllocladus aspleniifolius var. aspleniifolius (FIGURE 10) and var. alpinus, while pinnately compound ones are characteristic of P. glaucus (FIGURES 13, 14), P. hypophyllus, and P. trichomanoides. Simple phylloclades are 1-4 cm. long and are mostly rhombic in out- line. Their margins are finely toothed or nearly entire, or sometimes shallowly to deeply 2- to 3-lobed on one or both sides in the lower part. They may also be pinnatifid, but are only very rarely pinnately compound. Some of the pinnatifid phylloclades bear a terminal bud which is covered with awl-shaped scales (F1GURE 11). These buds either remain dormant or give rise to a short shoot usually 1-4 cm. long. These short shoots are spirally covered with short, lanceolate, deciduous leaves, and terminate 2o2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 FicurEs 10-15. Simple and pinnately eae phylloclades. 10-12, Phyllo- cladus asplend folie var. aspleniifolius: 10, simple phyllo clades. without terminal bud, one undivided, the other bauer Beer 11, simple pinnately lobed phylloclade with terminal bud; 12, short shoot arising from pinnately lobed phylloclade. 13-15, : Rea 13, pinnate phylloclade with iartag re in place of terminal bud: 4, pinnate phylloclade with terminal bud: shoot with crown of a phylloclades arising from tip of pinnate phylloclade. 1978] KENG, PHYLLOCLADUS 253 in a group of 3 to 5 simple or occasionally pinnatifid phylloclades with a new terminal bud in the center (FicureE 12). The pinnately compound phylloclades can reach 10-15 cm. in length (in vigorous growth and especially in saplings they may be 30 cm. long), each of them consisting of 5 to 10 (to 12 or more) segments or pinnae distichously arranged in two rows along the rachis. In taxonomic litera- ture these individual segments or pinnae have often been misleadingly called phylloclades or cladodes. The pinnate phylloclades, like the simple ones, may or may not possess a terminal bud (Ficures 13, 14). In the former case, the terminal bud can develop into a short shoot 4-10 cm, (-15 cm. or more) long. This new short shoot is partly and sparsely covered with deciduous, linear-lanceolate scale leaves and is crowned with 3 to 5 (to 10) normally pinnately compound phylloclades, in addition to a terminal bud (Ficure 15). In the latter case, there is always a large, prominent terminal segment in lieu of the terminal bud; this terminal segment is often partly lobed or partially fused with one or both lateral segments below it (FicurE 13). MoNoECY AND DIOECY. In taxonomic literature, the same species of Phyllocladus has been described as monoecious or dioecious by different authors. For example, P. aspleniifolius var. alpinus was generally con- sidered as monoecious, but Wardle (1969), after careful observation, pointed out that dioecy is the usual condition. It is known that a Phyllo- cladus tree that normally produces pollen cones can occasionally produce a few ovulate cones, although the latter are often aborted to various degrees. Reduced ovulate cones have been found on the same branch with pollen cones in P. hypophyllus (Keng, 1962b) (Ficure 16). Abnormal pollen cones of P. aspleniifolius var. alpinus, each with an ovule at its base, were first reported by Robertson (1906) and again by Wardle (1969). During the present study (Ficures 20, 21), a well-developed pollen cone was found on a branch of P. trichomanoides that bore predominantly ovulate cones. Preserved material of P. glaucus collected by Dr. A. E. Orchard from Cranwell Park, Auckland, New Zealand, shows a number of interest- ing abnormalities. There are 10 to 15 normal pollen cones spirally ar- ranged on the axis of a dwarf shoot, and each is subtended by a lanceolate bract. At the top of the shoot, there are several pinnate phylloclades bearing reduced ovulate cones (F1GURE 18). Some of the pollen cones have reduced ovules at their bases. A most extraordinary teratological specimen from the same collection shows that some of these reduced ovules are in fact subtended by scales in turn bearing two minute pollen sacs It appears that young trees of most species of Phytlocloiue, with the possible exception of P. glaucus, bear either pollen or ovulate cones and thus are clearly dioecious. When the trees become older, the distinction is less clear. Dr. B. P. J. Molloy of the D.S.I.R., Lincoln, New Zealand, however, believes (pers. comm.) that sexuality of PAyllocladus gradually becomes more stable in older trees. Further observation is needed. 254 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 cas i~s—< aes Re Ficures 16-21. Some teratological specimens of Phyllocladus. 16, 17, P. hypophyllus: 16, pollen cones and underdeveloped ovulate cones arising from same branch; 17, pollen cone borne on partly deformed phylloclade. 18, 19, 1978 | KENG, PHYLLOCLADUS 259 PoLLEN CoNEs. The pollen cones are nearly sessile or short stalked and are usually crowded on the tip of an unfolding side branch. After the elongation of the new shoot from the tip of a side branch, it becomes apparent that these pollen cones are actually spirally arranged along the axis of the new shoot, and each cone is usually subtended by a bract. Old, dry pollen cones may remain on the tree for months or even years. The young pollen cones are purplish pink to sulphur yellow in color. There are from 2 to 5 (as in PAyllocladus aspleniifolius var. asplenifolius, FicurE 22) up to about 20 (as in P. glaucus, FIGURE 23) cones together. The pollen cones are cylindrical, 1.5-2 (in P. asplenitfolius) to 3-4 (in Soltetaiaie tr tatietaietoes Ficures 22-26. Branches of Phyllocladus with pollen cones: 22, P. aspleni- folius var. aspleniifolius, 23, 24, P. glaucus (after T. Kirk); 25, P. hypophyllus ; 26, P. trichomanoides. 256 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 P. glaucus) cm. long, and consist of approximately 30 to 50 (to 60 to 70) cone scales spirally arranged along an axis. The cone scales are triangular in shape, often with a pointed tip and toothed upper margins. Each cone scale bears a pair of pollen sacs at its base (FrGure 24). The pollen grains have two tiny bladder wings and are dispersed by wind. The stalk of the pollen cone is always associated with a bractlike struc- ture. From observation of teratological specimens of Phyllocladus hypophyllus (Keng, 1963b, 1974) (Ficure 17) and other species, it was suggested that the bractlike structure on the stalk might represent a much- reduced simple phylloclade. It probably means that the pollen strobili were originally borne at the end of a simple phylloclade (or on the seg- ments of a pinnate phylloclade). A similar situation still occurs in the ovulate strobili of modern species. OvuLAaTE Cones, In the species with simple phylloclades (i.e., Phyllocladus as plentifolius var. aspleniifolius (FIcURE 27) and var. alpinus), the ovulate cones are single or borne 2 to 4 together. They arise below or on one side of a slightly deformed cladode (Ficure 28, b), or are apparently terminal on a long stalk (FicURE 28, a), the entire cladode being reduced. In species with pinnate phylloclades (P. glaucus, P. hypophyllus, and P. trichomanoides), the ovulate cones are either seated in the notch of a bilobed segment (e.g., P. kypophyllus) (Ficures 31, 32) or attached to both sides of the rachis of a pinnate cladode, occupying the place of the lower lateral segments (e.g., P. glaucus) (F1GuRE 29). Although these are the normal situations, there is considerable variation. For example, in P. hypophyllus the notched segments may be reduced to various degrees while the main axis of a pinnate cladode becomes enlarged and columnar; in P. trichomanoides (FicurEs 34, 35), on the other hand, the segments under- neath the ovulate cones often become deformed or almost obsolete. The ovulate cones are sessile (e.g., Phyllocladus hypophyllus) (FIGURE 33) or short stalked (e.g., P. glaucus) (FicuRE 30). They are usually ovoid in outline and consist of few to many, more or less spirally arranged cone scales. Some of these scales (usually 2 or 3, sometimes more) possess a single axillary ovule; most of the others, however, are barren. In most species, normally only one (rarely two or three) ovule in an ovulate cone develops into a seed, but in P. glaucus a fair number of seeds (S to 10, sometimes up to 20) are produced in a single cone. SEEDS. The seeds are oval in shape, slightly compressed dorsiventrally, and pointed above, often with a tiny, crooked tip. They are generally very small. Among the four species, Phyllocladus trichomanoides has the small- est seeds (2-3 mm. long), and P. hypophyllus the largest (6-8 mm. long). A conspicuous, cup-shaped aril protrudes beyond the ripe cone scales and envelops the lower half or two-thirds of the seed (F1IGuRE 9). The aril is more or less symmetrical, but the upper edge is often dentate or irregularly lobed. Mature cone scales are fleshy or leathery, and more or less fused. The cone scales form an amorphous mass known as the “receptacle” in the species with fewer cone scales; in P. glaucus the cone scales remain dis- Ficures 27-36. Branches of PAyllocladus with ovulate and seed cones. 27, 28, ch w ie ospleniolin var, aspleniifolius: 27, bran ith both ovulate and young seed cones; 28, a, b, two ovulate cones, one sintended by ae nial the other eae terminal. 29, 30, °P. glaucus: 29, branch with seed cones e on oe part of pinnate phylloclades (after T. Kirk); 30, ovulate con ae hypophyllus: 31, ovulate cone in notches of pinnae of pinnate Sails aoa branch with seed cones in notches of pinnae of pinnate phyllocla es; Re ovu ae cone. 34-36, P. trichomanoides: 34, ovulate cones on one side oP e of pin- nate “phylloclade; 35, branch with seed cones; 36, ovulate con 258 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 tinguishable, and are more numerous (to about 20). Fully ripened scales usually turn bright red, in sharp contrast to the colors of the aril and seed, the former normally being white and the latter either chestnut brown (as in P. hypophyllus) or black (as in P. aspleniifolius and others). The mature seeds fall from the arils to the ground close to the parent tree. However, due to the highly contrasting colors of the seed, the aril, and the receptacle, as well as the small size of the seed, it is conceivable that the seed of Phyllocladus might also be dispersed by birds as suggested by Preest (1963). SEEDLINGS. The two cotyledons (Ficure 37) are usually early deciduous but may last for two to several years before turning brown and withering away; they always leave a pair of inconspicuous scars on the lower part of the axis. They are linear-lanceolate in shape, double-nerved, 1.5—2 cm. long, and shallowly notched or bilobed at the apices. The juvenile leaves are linear and single-veined, varying between 0.5 and 1 cm. in length. They are spirally arranged on the axis. Their apices are acute or rounded and are often sharp pointed at the extremity. Later-produced leaves be- come successively smaller, finally being lanceolate or triangular scale leaves usually 2-3 mm. long; phylloclades or cladodes are present in the axils of some of these scale leaves. In PAyllocladus aspleniifolius (FIGURE 38), the cladodes on seedlings are mostly simple, although toothed or crenate to various degrees in the margins. In P. hypophyllus (FIGURES 40, 41), the earlier-formed cladodes are simple; these are gradually re- placed by the deeply lobed ones. The lobation then becomes more pro- nounced until, finally, almost all the cladodes are pinnately compound. In P. glaucus (Ficure 39) and P. trichomanoides (Ficurrs 42, 43), even the earlier-formed cladodes are pinnatifid. In the latter species, the cladodes of a fully grown seedling are distinctly pinnate and with the pinnae further shallowly or deeply lobed, thus approaching a bipinnate state (FicuRE 43). The color of the seedlings varies from green (PAyilocladus as plenii folius) , to grayish green (P. glaucus, P. hypophyllus), to purplish (P. trichoma- noides), but this also depends on environmental conditions. Seedlings of nearly all species of Phyllocladus, especially those of P. asplenii folius var. aspleniifolius and var. alpinus, are highly polymorphic with respect to the shape, size, and dentation of the early-formed cladodes. This probably sug- gests that they are not homogeneous genetically. Dr. R. K. Crowden, of the University of Tasmania (pers. comm.), points out that in Tasmania, where forest fires are comparatively frequent, Phyllo- cladus aspleniifolius var. aspleniifolius is well adapted for survival. Seeds of this plant can probably remain in the forest soil for several years. After a forest fire, the vegetation is largely devastated, but a great number of seedlings of PAyllocladus often appear in the burned-over area. In con- trast to this, seeds of the New Zealand species appear to be very short lived (Dr. B. P. J. Molloy, pers. comm.). Seedlings of all the Phyllocladus species observed possess root nodules which contain symbiotic, nitrogen-fixing bacteria. These nodules are 1978] KENG, PHYLLOCLADUS 259 especially well developed in the Tasmanian and New Zealand species. This is probably one of the attributes that enable the seedlings to survive in a rather unfavorable environment. EVOLUTIONARY TRENDS The following hypothetical evolutionary trends seem to occur within the genus PAyllocladus. In the trends listed, the simpler condition is con- sidered to be the derived state; the more complex condition, the ancestral. PHYLLOCLADES. The seemingly bipinnate phylloclades, as seen in the seedlings of Phyllocladus trichomanoides (F1GURE 43), likely represent the ancestral form from which simple pinnate, pinnatifid, and finally simple phylloclades (as in P. aspleniifolius, FIGURE 10) were probably evolved. POLLEN STROBILI. Each pollen strobilus was probably originally seated in the top notch of a simple phylloclade, as shown in a teratological specimen of Phyllocladus hypophyllus (FicuRE 17). These simple phylloclades are usually either reduced and bractlike (FrcuRE 24, a) or obsolete. In P. glaucus, there are numerous (about 20) long-stalked, bracteate strobili, each composed of many (approximately 60 to 70) sporophylls. These strobili first appear to be crowded on the top of a short shoot and to surround the terminal bud (Figures 23, 26); later, after the activation of the terminal bud and the elongation of the new short shoot, it becomes clear that they are actually spirally arranged on the new short shoot (Ficures 18, 25). In contrast, P. aspleniifolius var. aspleniifolius (Ficures 1, 3, 4) has only 2 or 3 almost sessile strobili, with each strobilus composed of only 30 to 50 sporophylls. The latter probably represents the most highly evolved form. OvULATE sTROBILI. Each strobilus was probably originally seated in the notch of a pinna of a pinnate phylloclade, as shown in Phyllocladus hypo- phyllus (Ficures 31, 33) and in some cases also in P. trichomanoides (FicurE 35). These pinnate phylloclades thus remarkably and uniquely perform both vegetative and reproductive functions. There are 4 or 5 or more strobilus-bearing pinnate phylloclades together, arranged on the top of a short shoot in a false whorl. Each strobilus, presumably in the basic form, consists of a large number (to 20 or so) of fertile scales, as seen in P. glaucus (FIGURE 30). DISTRIBUTION OF SEXES. The trend seems to be from monoecious to dioecious. Nearly every species of PAyllocladus shows a certain degree of monoecism. The ancestral form probably possessed a number of simple phylloclades spirally arranged on a short shoot, each of which bore a terminal staminate strobilus. On the tip of the same short shoot, there was a crown of pinnate phylloclades, each with an ovulate strobilus on its pinna (cf. FIGURE 18). In short, among the four living species, it appears that Phyllocladus glaucus probably preserves more primitive traits than do the other species. This species is followed by P. hypophyllus and P. trichomanoides. Con- 260 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 FicurEs 37-43. Seedlings of Phyllocladus: 37, 38, P. aspleniifolius var. aspleni- ifolius; 39, P. glaucus; 40, 41, P. hypophyllus; 42, 43, P. trichomanoides. 1978] KENG, PHYLLOCLADUS 261 versely, P. aspleniifolius var. aspleniifolius and var. alpinus in many re- spects possibly represent the much-reduced or the highly evolved form. TAXONOMIC TREATMENT Phyllocladus L. C. & A. Richard, Comment. Bot. Conif. Cycad. 129. biG ty. des ee 0; Podocar pus Labill. Nov. Holl. Pl. Sp. 2: 71. ¢. 221. 1806, non L’Hérit. ex Pers. Brownetera L. C, Rich. Ann. Mus. Hist. Nat. Paris 16: 299. 1810, nomen nudum. Thalamia Sprengel, Anl. Kennt. Gewachse. ed. 2. 2: 218. 1817. Shrubs or small to large trees. Juvenile leaves at first linear, changing gradually to subulate and scaly. Phylloclades simple, pinnatifid, or pinnately compound, variable in shape and size, usually 2 to 5 or more in a false whorl. Plants with unisexual strobili mostly dioecious, some- times monoecious. Pollen cones 2 or 3 (up to 20), crowded in a cluster on the top of a short shoot, becoming spirally arranged along the new short shoot after its elongation; individual pollen cones consisting of 30 to 70 spirally arranged sporophylls, each bearing 2 pollen sacs; pollen grains winged. Ovulate cones terminal or marginal on fully grown or re- duced simple phylloclades, or on the pinnae of a pinnate phylloclade, the phylloclades 2 to 5 or more in a false whorl on the top of a short shoot; individual ovulate cones consisting of few, several, or numerous scales, usually only one or few, sometimes many (as in Pkyllocladus glaucus) of the scales being fertile; ovules erect, solitary in the axil of a scale. Seeds ovoid, dorsiventrally compressed, subtended by a cup-shaped, filmy aril and the often fused, succulent scales. Four species, distributed in Tasmania, New Zealand, and Malesia (from New Guinea to Luzon and Borneo). Type species: Phyllocladus rhomboidalis L. C. & A. Richard (= Phylloctadus aspleniifolius (Labill.) Hooker f.). KEY TO THE SPECIES OF PHYLLOCLADUS A. Phylloclades of adult plants mostly simple, sometimes shallowly or deeply bed, occasionally pinnatifid, very rarely truly pinnate. [Ovulate cones soli- tary or 2 to 4 on the margins and base of a modified (often much-reduced, with the cones thus seeming to be terminal) phylloclade, usually two or more such modified or reduced phylloclades together in a small cluster.] ............ 1. P. siete ian Phylloclades of adult pa always pinnate, consisting of 5 or 6 to many (about 12) distinct segmen B. Ovulate cones normally ore in the apical notch of a segment of the phylloclade, the segment varying from obcordate to oblanceolate, some- times columnar, very rarely totally reduced; mature seeds 5—8 mm. long mostly solitary, rarely in pairs, in an ovulate cone. .... 3. P. hypophyllus. Ovulate cones usually on the margins and base of a segment which is often highly modified (the lower segments in a pinnate phylloclade sometimes eo w 262 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 totally reduced, the ovulate cones thus appearing to be terminal on a side branch of the rachis); mature seeds 2-4 mm. long, | to many per cone. C. Ovulate cones consisting of about 10 to 20 fertile scales; er cones 1 or rarely 2 seeds; pollen cones 5 to 10 terminating a branchlet, se ENE Sass BAe G teed aida ue Wet ee howe Vex ais C. Ovulate cones e-eonereine of 2 or 3 fertile scales; mature cones with 1 or rarely 2 seeds: pollen cones 5 to 10 terminating a branchlet. pAb pecuee ss _.. 4. P. trichomanoides. 1. Phyllocladus aspleniifolius (Labill.) Hooker f. in London Jour. Bot. 4: 151. 1845 Podocarpus aspleniifolia Labill. Novae Holl. Op. 20:7 1e2 ced, 1e00; Type: Tasmania, J.-J. H. de Labillardiére (not seen). Phyllocladus rhomboidalis L. C. & A. Rich. ce Bot. Conif. Cycad. 23. t. 2. 1826. v7, & Ga. Phallocladus billardier: L. C. Rich. ex Mirbel in Mém. Mus. Hist. Nat. Paris 3: 48. 1825, nomen nudum. Thalamia aspleniifolia (Labill.) Sprengel, Anl. Kennt. Gewachse. ed. 2. 2: 218. Shrub or small to medium-sized tree, 6-18 m. high. Bark brownish to dark brown, deeply furrowed in old trees. Juvenile leaves on seedlings linear, 0.7-1 cm. long. Scale leaves on seedlings and on the short branches of adult plants triangular or subulate, 0.1-0.3 cm. long. Phylloclades on seedlings usually linear-oblong to rhomboid, 1.5—2.5 cm. by 1-1.2 cm., shallowly or deeply lobed, sometimes pinnatifid; phylloclades on adult plants mostly simple, rhomboid, 2.5-8 cm. by 1.5-2 cm., the apex acute or blunt, the base cuneate, sessile or stalked, the margins irregularly lacerate or occasionally pinnatisect, very rarely pinnately compound. Plants usually dioecious. Pollen strobili usually 2 or 3 (to 5) together, terminating lateral branches, cylindrical, 0.3-0.5(—1) cm. long, sulphur yellow, short stalked. Ovulate strobili solitary or several (2 to 4) at the side or base of a modified or highly deformed phylloclade, or sometimes the latter totally reduced, with 2 to 4 ovulate strobili together forming a seemingly terminal stalked cluster, each strobilus consisting of 2 or 3 (to 5) erect ovules each ua tended by a pinkish red, fleshy scale. Seeds ovoid, about 0.5 cm greenish black to black; aril white, irregularly lobed: receptacle purplish red and fleshy. KEY TO THE VARIETIES OF PHYLLOCLADUS ASPLENIIFOLIUS a Shrub or small to medium-sized tree, maximum eds about 18 m.; scale leaves on short shoots of adult plants subulate, 0.2-0.3 cm. long. Co nfined to SRS isa e Sone We nae oe hake a ean Gad es la. var. aspleniifolius. A. Shrub or small tree, maximum height about 9 m.; scale leaves on short shoots of adult plants triangular, less than 0.1 cm. long. Restricted to New Zealand. Ee RN Aeeea a ee en es Lae Gok ede oe ee ae eee nk as 1b. var. alpinus. 1978 | KENG, PHYLLOCLADUS 263 la. Phyllocladus aspleniifolius (Labill.) Hooker f. var. aspleni- ifolius. FicuRES 1-9. Phyllocladus aspleniifolius (Labill.) Hooker f. in London Jour. Bot. 4: 151. 1845: Henkel & Hochst. Synop. Nadelhdlzer, 371. 1865; Pilger in Engler, Pflanzenr. 4(5): 97. 1903: Curtis in Stones. Endemic Fl. Tasmania 3: 106. pl. 60. 1971; Curtis & Morris, Stud. Fl. Tasmania. ed. 2. 1: 2. fig. 2. 1975. Phyllocladus poi? L. C. & A. Rich. Comment. Bot. Conif. Cycad. 23. fig. 2. - Endlicher. Synop. Conif. 235. 1847; Hooker f. Fl. Tas- maniae ae eee Voy. pt. 3) 1: 358. 1859; Carriére, Traité Gén. Conif. ed. 2. 706. 1867; Parlatore in DC. Prodr. 16(2): 499. 1868; Curtis, Stud. Fl. Tasmania 1: 234. 1956. Phyllocladus billardieri L. C. Rich. ex Mirbel in Mém, Mus. Hist. Nat. Paris : 48. 1825, nomen nudum. Phyllocladus serratifolia Nois. ex Henkel & Hochst. Synop. Nadelholzer, 372. 1865. Podocarpus aspleniifolia Labill. Novae Holl. Pl. Sp. 2: 71. ¢. 221. 1806. Thalamia aspleniifolia (Labill.) Sprengel, An]. Kennt. Gewachse. ed. 2. 2: 218. 1817 This variety is characterized by its taller stature (maximum height about 18 m.) and by its longer (about 0.2-0.3 cm.) subulate scale leaves on the short shoots of adult plants. DiIstTRIBUTION. Endemic to Tasmania, Australia (Map 1). In temperate rain and wet sclerophyll forests, from 500 to 800 meters, largely confined to the western or the moist parts, occasionally descending to sea level in wet coastal areas. REPRESENTATIVE SPECIMENS. Australia. TASMANIA: Corinna, W. D. Jackson, J 146 (wo 2775); Cradle Mts., N. Sunderland (Ho 2776); Lake St. Clair, below Mt. Ida, N. 7. Burbidge (Ho 2768); King William Range, EZ. Rodway 134 (Ho 2767): Mt. Field Nat. Park, H. D. Gordon (wo 2756); W. M. Curtis (Ho 2751); Wishing Well, Mt. Wellington, J. Somerville (Ho 2766); Huon River, near Picton Hut, Somerville 288 (110 2806); near Franklin Bridge, Lyell Highway, J. Town- row (Ho 2782, 2801); Port Davey, D. H. Martin (Ho 2777, 2778). 1b. Phyllocladus aspleniifolius Asta ) Hooker f. var. alpinus (Hooker f.) H. Keng, comb. n Phyllocladus alpinus Hooker f. Fl. Novae-Zeland. 1: 235. t. 53. 1853; Carriere, Traité Gén. Conif. ed. 2. 708. 1867; Kirk in Trans. New Tenena Inst. 10: 382, 1878, Forest Fl. New Zealand, 199. ¢. 100. 1889; Pilger in Engler, Pflan- zenr. 4(5): 98. 1903; Cheeseman, Man. a iS aa Fl. 659. 1906, ed. 2. 120. 1925; Allan, Fl. New Zealand 1: 112. - Wardle in New Zealand Jour. Bot. 7: 76. 1969. Type: New pate an Island, Tongariro, Bid- will. Phyllocladus trichomanoides D. Don var. alpinus (Hooker f.) Parl. in DC. Prodr. 16(2): 498. 1868 This variety differs from the typical form of the species in the lower stature (maximum height about 9 m.) and in the shorter (less than 0.1 cm.) triangular scale leaves on the short shoot in adult plants. 264 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 re ~~ Map 1. Distribution of Phyllocladus aspleniifolius var. aspleniifolius in Tas- mania, Australia. DISTRIBUTION. Endemic to New Zealand (Map 2). In montane and subalpine forests and scrubs, usually at 500 to 1900 meters altitude on both North and South Islands, extending from approximately 37 degrees latitude to Foveaux Strait, but descending to sea level to the west of the coastal mountains of Nelson, the Southern Alps, and south of Southland. REPRESENTATIVE SPECIMENS. New Zealand. Nort IsLaNnp: King County, Pureora, T. W. Rawson (CHR 56455); Coromandel Peninsula, Moehau, L. B. Moore (CHR 40942, 40943); Raukumara Range, Hikurangi, 7. R. Fryer (CHR 185456); Almateatua, Maungaha-rutu Range, A. N. Druce (cHR 216842); Volcanic Plateau, Puke-o-nake, P. Wardle (cur 185468); Tongariro Nat. Park, Chateau, R. Mason (cHR 82609); Ruapehu Nat. Park, A. Cook (cHR 69111); Central Ruahine Range, near Waikanaka Hut, Wardle (cHR 185459); Ararua Range, Druce (CHR 190505, 190506); Wellington, Wilton’s Bush, Harrigan & Cook (CHR 83778). SoutH IsLanp: Nelson, Wanagapeka River, B. H. MacMillan (CHR 219942); Cobb Valley, Sylvester Stream, Moore (cur 151027); near 1978] KENG, PHYLLOCLADUS 265 q 44 Ja . a) a 4 a a 2] aa : ae Hy A De te 174 178 Map 2. Distribution of Phyllocladus aspleniifolius var. alpinus (a), P. glaucus (g), and P. trichomanoides (t) in New Zealan Nelson, Dun Saddle, Wardle (cur 179033); Marlborough, Mt. Stokes, J.G. Hay (CHR 112278); Lewis Pass, Paradise Lake, A. H. MacRae (cHR 185463); near =] n eS oO cS at) ce) e) Met jar oO v2) — ae jmm c Ry oe rg Fat) wm a es) A Vv ‘ : Melville (cur 130873); Mt. Cook, Hermitage, H. E. Connor (CHR 108109) ; Mt. Cockayne, Broken River, E. M. Chapman (CHR 258592); Karangama, 266 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Westland Nat. Park, Wardle (cur 179110); fiords, Stuart Mts., W. H. Thomson (CHR 88133); Expectation Stream, Caswell Sound, V. D. Zotov (CHR 71084) ; Fiordland, Lake Make, M. J. A. Simpson (CHR 115975); Southland, Nevis, A. J.D. Barker (cur 20950); Matukituki. Otago, J. A. Langbein (cHR 167031); N. Otago, Omarama, Moore (CHR 71226); Lake Ohau, Dobson River, Wardle (CHR 185470). In his Flora Tasmaniae (Vol. 1, p. 359), J. D. Hooker commented that the New Zealand alpine species (PAwlocladus alpinus) so closely resembles the Tasmanian species that he almost doubted “its being distinct.” T. Kirk (in Trans. Proc. New Zealand Inst. 10: 382, 383. 1878) stated that “T have no doubt that it [P. a/pinus| will ultimately prove identical with the Tasmanian P. rhomboidalis, . . . for although specimens from alpine habitats look very different to that plant, fruited specimens from low levels are undistinguishable.’ Recently, P. Wardle (1969) recognized two eco- types (alpine and lowland) of this plant and observed that the lowland ecotype of P. alpinus is very similar to the Tasmanian plant. He also mentioned three characters that distinguish them: maximum height, bud scales in the adult plant, and seed clusters. The last character (seed clusters usually lateral at the base of the phylloclade in New Zealand species and usually terminal on the phylloclade in Tasmanian plants) does not stand, so varietal rank seems most appropriate. 2. Phyllocladus glaucus Carr. Traité Gén. Conif. 502. 1855, ed. 2. 707. 1867; Kirk in Trans. New Zealand Inst. 1: 149. 1868, ibid. 10: 380. 1878, Forest Fl. New Zealand, 195. t#. 98, 99. 1889; Pilger in Engler, Pflanzenr. 4(5): 95. 1903; Cheeseman, Man. New Zealand Fl. 658. 1906, ed. 2. 120. 1925; Allan, Fl. New Zealand 1: 113. 1961. Type: New Zealand. North Island, Maungatawhiri, R. Mair. Phyllocladus trichomanoides D. Don var. glaucus (Carr.) Parl. in DC. Prodr. (2): 498. 1868 Small tree, to 12 m. tall. Bark dark brownish, deeply fissured. Juvenile leaves on seedlings linear, obtuse or acute, sometimes falcate, 0.5-1 cm. long. Scale leaves on seedlings subulate, about 0.5 cm. long: those at the base of a young terminal shoot of adult plant linear-lanceolate, 1.5—2 cm. long, thin-membranaceous, caducous. Phylloclades on seedlings usually deeply pinnatifid, 3-5 cm. long; those on adult plants pinnately compound, 12—20(—30) cm. long, consisting of 5 to 10 (to 12) segments, the seg- ments pale bluish green above and glaucous beneath when young, rhombic to ovate-flabellate, 2-5 cm. long, subsessile or stalked. Plants usually dioecious. Pollen strobili usually 10 to 20 crowded at the tips of the short branches, cylindrical, 1-2(-2.5) cm. long, stalked, the stalk stout. equal- ing or longer than the cone. Ovulate strobili 3 to 5, inserted on each side of the pinnate phylloclades at the tips of the short branchlets, occupying the place of the lower lateral segments; each strobilus ovoid, 0.5—0.8 cm. long, consisting of 10 to 20 spirally arranged fertile scales. Ripe seed cones pinkish red, 1-1.2 cm. long, with 8 to 20 seeds exserted beyond the 1978 | KENG, PHYLLOCLADUS 201 thickened scales. Seeds ovoid, compressed, black, 0.3 to 0.4 cm. long; aril cup-shaped, white, fleshy. Seed cone scales remaining integral; no re- ceptacle formed. DIstRIBUTION. Endemic to New Zealand (Map 2). In lowland and montane forests from sea level to 850 meters, rarely to 1000 meters on some mountain peaks, in the northwestern parts of North Island. REPRESENTATIVE SPECIMENS, New Zealand. NortH IsLtanp: Mangonui, Peria, Watts Bush, H. Carse 2343 (aK); near Whangarei, Mt. Maungatapere, Carse 641 (ak); Little Barrier Island, L. Cockayne 6296 (ax); Great Barrier Island, T. Kirk (weit 37769); Auckland, Birkdale, Carse 2337 (ak); Coromandel Peninsula. A. E. Wright 700 (ak); Te Aroha Mts., W. R. B. Oliver (WELT 15267); Rotorua, Mt. Tarawera, P. Hynes 110361 (ak); Lake Waikare, Ani- waniwa Falls, D. A. Crawford (CHR 65355). The following specimens are considered to be intermediate forms or hybrids: WELT 37791, collected by Thompson from Waitakere Range, New Zealand, is intermediate between Phyllocladus glaucus and P. tricho- manoides: CHR 186617 and CHR 186618, collected by A. P. Druce from Hine River, west of Waikareti, in Ureivera National Park, are between P. glaucus and P. aspleniifolius var. alpinus. Further examination is needed. 3, Phyllocladus hypophyllus Hooker f. Ic. Pl, WS. 52 %..60~ 1852; Carriére, Traité Gén. Conif. ed. 2. 706. 1867; Parlatore in DC. Prodr. 16(2): 499. 1868; Stapf in Trans. Linn. Soc. Bot. 4: 249. 1894: Pilger in Engler, Pflanzenr. 4(5): 99. 1903; Merrill, Enum. Philip. Fl. Pl. 1: 5. 1925; H. Keng in Gard. Bull, Sing. 207 123, fig. 1963; de Laubenfels in Jour. Arnold Arb. 50: 278. 1969. TYPE: Borneo. Mt. Kinabalu, alt. 8000 feet and upwards, H. Low. Phyllocladus hypophyllus Hooker f. var. protracta Warb. Monsunia 1: 278. 1900. Type: Mindanao, Warburg 14722. Phyllocladus protractus (Warb.) Pilger in Engler, Pflanzenr. 4(5): 99. 1903. Phyllocladus major Pilger in Bot. Jahrb. 54: 211. 1916. Type: New Guinea, Ledermann 9872. Small or large tree, up to 30 m. or more high. Bark reddish brown to dark brown, flaking, with hard, rough granules, breaking off in large scales. Juvenile leaves on seedlings linear, 0.5-0.8 cm. long. Scale leaves on the short branchlets of adult plant awl shaped, 0.2-0.3 cm. long. Phylloclades on seedlings spathulate to rhomboid, 2-8 cm. long, the apex acute or caudate, the base narrowly attenuate, the margin irregularly serrate, lacer- ate, or pinnatifid; phylloclades on adult plants pinnately compound, 10-15 cm. long, usually consisting of 5 to 10 segments, the segments oblanceolate, obovate to rhomboid, 1.5-4 cm. (-10 cm. in saplings) long, the upper surface dark green, the lower often glaucous. Plants usually dioecious. Pollen strobili often 2 or 3, sometimes to 15, in clusters on the tip of a short branchlet, sometimes mixed with sterile (but reduced) phylloclades or phylloclades bearing reduced ovulate strobili, yellowish brown, cylin- 268 JOURNAL OF THE ARNOLD ARBORETUM [voL, 59 drical, 1.2-1.5 cm. long, stalked. Ovulate strobili purplish, solitary or rarely 2 (or 3) together situated in the notch of a bilobed simple phyllo- clade, or in the notches of the bilobed segments of a compound phylloclade, sometimes these segments becoming columnar or obsolete: each ovulate strobilus ovoid or ellipsoid, consisting of about 15 coalescent scales of which only 2 or 3 are fertile, each bearing a simple, erect ovule. Seeds ovoid, chestnut brown, 5-8 mm. long, flattened, the aril whitish; receptacle yellowish brown, fleshy. Distr1BuTION. Endemic to Malesia (Map 3), from Luzon and Borneo, through the Celebes and Moluccas, to New Guinea. Scattered in montane rain and subalpine forests, on ridges and summits, from 1500 to 3000 meters, and occasionally also from 900 to 4000 meters in some ae areas. REPRESENTATIVE SPECIMENS. Philippines. Luzon: Mt. Benguet, Panai, E. D. Merrill 4753 (1); Benguet, Mt. Singakalsa, M.D. Sulit, PNH 7669 (L). Mtn- DANAO: Bukidnon, Mt. Katanglad, Swit, PNH 10052 (L); Davao, Mt. Apo, R. Kobbins s.n. (L). Borneo. SARAWAK: Baram, Mt. Mulu, J. A. R. ie 4544 (1); Melinau, Mt. Api, P. Chai, S 30366 (1): Mawas. Marigan Range, E. F. Brunig, S 9984 (1). Brunet: Mt. Ulak, P. Ashton, BRUN 1033 (L). SABAH: Mt. Kinabalu, : & M. S. Clemens 29743 (L); Chew, Corner, & Stainton, RSNB 710 (1); Kudat, Trusmandi, G. Mikil, SAN 31784 (1). KALIMANTAN: Mt. Semedum, #7. Hallier 697 (1); Banan, Bengkajang, bb 24777 (L). Celebes: Mt. Loemoet, Menado, P. J. Eyma 3621 (1); Enrekang, Sawito, collector unknown, bb 20782 (1); Labu, Malili, A. Burhi, bb 24089 (1). Moluccas: Batjan, de Haan, bb 23236 (L); Obi, de Haan, bb 23812 (L); Ceram, Mt. Sofia, E. Stresemann Map 3. Distribution of Phyllocladus hypophyllus in Malesia. 1978 | KENG, PHYLLOCLADUS 269 133 (1). Irian Jaya: Vogelkop Peninsula, Nettoti Range, van Royen & Sleumer 7403 (L); Kebar Valley, Neetjapaki Mts., C. Kalkman, BW 6373 (L); Wissel Lake, Eyma 4954 (1); Lake Habbema, L. J. Brass 10528 (L); Mt. Antares, Star Mts., Kalkman 4539 (L). Papua New Guinea. W. HIGHLANDS: Mt. Hagen, Hoogland & Pullen 5871 (v), Lake Inim, J. &. Flenley, ANU 2177 (1). MOROBE: Mannasat. Cromwell Mts., R. D. Hoogland 9482 (CaNB, L); Mt. Kaindi, Brass 29692 (CANB, L). S. HicHLANDS: Mt. Giluwe, &. Schodde 2014 (CANB, L). CENTRAL: Murray Pass to Woitape, Forman & Wardle, NGF 45587 CE 4. Phyllocladus trichomanoides D. Don in Lambert, Descr. Genus Pinus. ed. 3. 2: 159. 1832. A. Cunningham in Ann. Nat. Hist. 1: 211. 1838: Hooker, Ic. Pl. ns. 2: tt. 549-551. 1843; Endlicher, Synop. Conif. 235. 1847; Hooker f. Fl. Novae-Zeland. 1: 235. 1853; Carriére, Traité Gén. Conif. ed. 2. 705: 1867; Parlatore in DC. Prodr. 16(2): 498. excl. vars. 1868; Kirk in Trans. New Zealand Inst. 10: 381. 1878. Forest Fl. New Zealand, 9. ¢t. 6, 7. 1889; Pilger in Engler, Pflanzenr. 4(5): 97. 1903; Cheeseman, Man. New Zea- land Fl. 658. 1906, ed. 2, 119. 1925, Illustr. New Zealand Bl, 2: $1. 190. 1914: Allan, Fl. New Zealand 1: 113. 1961. LecroTyPE: New Zealand. North Island, Wangaroa, A. Cunningham (x). Medium-sized to large tree, 20-23 m. tall. Bark smooth and thick, greenish to dark gray or blackish. Juvenile leaves linear, acute, 0.8-1 cm. long. Scale leaves at the base of young shoots subulate, 0.2—0.3 cm. long. Phylloclades on seedlings usually pinnatifid to pinnately compound, 5-7 cm. long, stalked; those on adult plants pinnately compound, 3-8(-12) cm. long, consisting of 6 to 10 (to 12) segments, the segments broadly rhomboid or ovate-flabellate, 1.2-2.5 cm. long, the margin varying from broadly toothed to deeply laciniate. Plants usually dioecious. Pollen strobili 5 to 10 in a terminal cluster, cylindric, 0.8-1 cm. long, stalked, the stalks 0.3-1 cm. long. Ovulate strobili usually borne on one side of the segments of the pinnate phylloclade, these segments often reduced in various degrees, sometimes merely stalklike; each strobilus irregularly ovoid or subglobose, 0.2—-0.3 cm. long, usually consisting of 2 or 3 fertile scales, each possessing a single ovule. Seeds dark blue to black, ovoid, compressed, about 0.3 cm. long, solitary, exserted beyond the crenulate cupular aril and the succulent receptacle. Distrrpution, Endemic to New Zealand (Map 2). In lowland and montane forests, from sea level to 800 meters, rarely to 1000 meters in restricted areas, on most parts of North Island (from North Cape to Taranaki and Hawke Bay) and on the northern tip of South Island. REPRESENTATIVE SPECIMENS, New Zealand. NortH Istanp: North Cape, Kerr Point. H. D. Gordon 1993 (wELT); Mangonui, Kaitaia, Oinu, 7. B. Mat- thews 2328 (AK); Bay of Islands, Russell State Forest, A. EZ. Orchard 3794 (AK) ; Whangarei, Mangakahia Valley, Twin Bridges, K. Cooper 120085 (AK); Little Barrier Island, E. M. Smith 109127 (ak); Great Barrier Island, Cooper (CHR 206941); Anawhata Valley, L. B. Moore & L. M. Cranwell 91345 (AK); Waite- 270 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 mata, Waitakere Range, Orchard 3372 (aK); Manukau, Hunua, N. J. Pickett Thames, J. Adams 14276 (AK); Tauranga, McLaren’s Falls, W. J. Simpson (CHR 117400); Gisborne, Motu River, E. M. Chapman (cHR 258563): near New Plymouth, Barratt Lagoon, A. N. Druce s.m. (CHR 274998). SoutuH ISLAND: Marlborough, Okiwi Bay, H. Talbot (cre 270096); near Nelson, Mai Tai Valley, P. Wardle (CHR 179034); S. E. of Cape Farewell, Puponga, Druce (CHR 2735456). ECONOMIC USES There is not enough timber from any of the four species of celery pine to be of commercial importance. Nevertheless, the wood is white and even grained, and shrinks very little on drying. It is valuable for flooring, boat decks, railway sleepers, mine props and piles, and occasionally for local building purposes and agricultural implements. The bark contains a high percentage (20 to 25%) of tannin, and once was considered to be of creat value in tanning. According to Kirk (1889), the bark of two New Zealand species (PAyilocladus glaucus and P. trichomanoides) also yields a red dye which was formerly used by the Maoris for ¢ yeing their cloaks. joan PALEOQGEOGRAPHIC DISTRIBUTION Although PAyllocladus is now limited to Tasmania, New Zealand, and eastern Malesia, fossil records reveal its much wider distribution in geo- logic time. After enumerating and evaluating the paleobotanical data, Couper (1960) reached the following conclusions. (1) Macrofossils iden- tied on the basis of cuticular structure and cladode morphology as Phyllo- cladus are recorded from the Oligocene of continental Australia and the Oligocene, late Tertiary, and Quaternary of New Zealand. (2) Micro- fossils (pollen grains) are found sporadically in Tertiary sediments in Australia, Tasmania, and New Zealand (Map 4). In New Zealand they occur from the Oligocene onward, but in Australia they are found mainly in the Oligocene. They are also found in the lower Tertiary of Seymour Island in western Antarctica. (3) No fossils of Phyllocladus are known from the Cretaceous or Tertiary of northern South America, New Guinea, Borneo, or India. A phyllocladean affinity has been suggested for some fossil woods from southern South America, but these identifications are inconclusive. Recent pollen studies by Dodson (1974a, 1974b) indicate that Phyllo- cladus grew in western Victoria and southeastern South Australia during the Quaternary and became extinct only a few thousand years ago. Also according to the same author (pers. comm.), P. aspleniifolius var. as pleni- ifolius was formerly present on Flinders Island in Bass Strait (between mainland Australia and Tasmania). It disappeared from the island at the beginning of this century as a result of an influx of settlers. The absence of fossils in the tropics and in the Northern Hemisphere deserves special attention. It supports the view that the present extension 105 120 135 150 165 180 165 150 135 120 105 90 75 60 48 30 oe T 7 ‘Sv wv TS 14" ~~ a ‘ eS) 9 30) 7 . bas a ¢ Oo ote No tt callie mubcodelligsee cel steer tee eat ance Syeeeaeteee a ANP od NS ee eee teen we [Segue al i) ecanseaiaaail Malas MaMa fe, or a * s ° re o| dite .-, 15 15 . “ » ~ ‘ . ’ : ‘a te Pa Fee | ote : of = _ - 3 * : f A st, oh A ae i ~ jae ote veg . ” , =o" Sl = * i a em as ace 15 : -. 3 a ba e, : 2 es Ree ry . . . ols. * ~ ‘et s = ; + pacts: Bay ekctkeiteietets chebeieteher citi Renee te ey - --h--- een et elated Setehehiabenie ee Sater eer OO tals bhatt iiaiaiacaiaa tine (Rael DS ar A icine Sis 30 . hy . . s Pl ts teen fepe o” >, ea 6 a va e 45 Fe 7 i “tons erecee, — eee Sa i ee e e : e * % ° », le ¢ s ve ry B ° t 8 60 . . Cad mi sre Oe Ste e rio a Pep eee > ae | ee eee See eee Ed qeere me near Tze nesen4 | rd f= 90 1066 120 135 150 165 Eset 180 West 165 150 138 120 105 90 (4) 45 30 Copyrighted by Standard Process and Engraving Co., Berkeley, California. Map 4. Distribution of Phyllocladus, present (solid line) and paleogeographic (dotted lines). (Paleobotanical data based on Couper, 1960.) SAAVWIDOTIAHd ‘ONAN [S261 Lee 272 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 of Phyllocladus into eastern Malesia was a relatively recent development (Hair, 1963; de Laubenfels, 1969), Among the fifty or so genera of extant conifers, about thirty of them are confined to the Northern and fourteen (including Phyllocladus) to the Southern Hemisphere. Li (1953) has summarized the present range of conifers and emphasized the clear differentiation into northern and southern groups. This separation, according to Li, has been shown by paleobotanical records to have existed since the Paleozoic. Nearly all of the considerable number of monotypic or oligotypic genera (including Phyllocladus) show restricted or disjunct ranges, grow best in a damp mesophytic habitat, and are practically all confined to lands bordering the Pacific Basin. This is considered by some botanists (including Li) to support the continental drift theory as modified by du Toit, who believed in two ancient land masses, Gondwana and Laurasia. separated by the Tethys Sea. Krassilov (1974), however, feels that the history of conifers has little bearing on the problem of continental drift and suggests that the southern and northern conifers have been more effectively separated by equatorial climatic conditions than by water barriers. ACKNOWLEDGMENTS This study was carried out during a sabbatical leave from the Uni- versity of Singapore, and was made possible by an Inter-university Ex- change grant awarded by the British Council. Professors W. D. Jackson, of the University of Tasmania, Australia, and W. R. Philipson, of the University of Canterbury, New Zealand, kindly provided laboratory, li- brary, and other facilities. Specimens of the following herbaria were con- sulted: Auckland Institute and Museum. Auckland, New Zealand (ax): Botany Division, D.S.I.R., Christchurch, New Zealand (cuR); Dominion Museum, Wellington, New Zealand (WELT); Herbarium Australiense, Canberra, Australia (cans); Rijksherbarium, Leiden, the Netherlands (L); Tasmanian Herbarium, Hobart, Tasmania, Australia (Ho). To the authorities and personnel of all these institutions, I should like to extend my deep gratitude. In addition, I should also like to thank the following individuals for their useful discussions and willing assistance: N. M. Adams, M. Barker, E. J. H. Corner, R. K. Crowden, W. M. Curtis, J. W. Dawson, J. R. Dodson, H. Eichler, R. D. Hoogland, D. Hou, B. P. J. Molloy, L. B. Moore, D. I. Morris, A. E. Orchard, J. W. Parham, F. B. Sampson, P. Wardle, and many others. I wish to record my appreciation for the many valuable suggestions of those who reviewed the manuscript. Finally, I should like to thank my wife, Mrs. R. S. Ling Keng, for ac- companying me on the trips and for making numerous drawings from living and preserved materials. Only a few of these have been selected to il- lustrate this paper. 1978 | KENG, PHYLLOCLADUS 213 LITERATURE CITED Couper. R. A. 1960. Southern Hemisphere Mesozoic and Tertiary Podocarpa- ceae and Fagaceae a their palaeogeographic significance. Proc. Roy. Soc. London B, 152: 491- pE LAUBENFELS, D. J. nee A revision of the Malesian and Pacific rain forest conifers. I. Podocarpaceae, in part. Jour. Arnold Arb. 50: 274-369. Dopson, J. R. 1974a. Vegetation and climatic history near Lake Keilambete, West Victoria. Austral. Jour. Bot. 22: 709-717. _1974b. Vegetation history and water fluctuations at Lake Leake, south- eastern South Australia. I. 10,000 B. P. to present. bid. 22: 719-741. Harr, J. B. 1963. oo relationships of the southern podocarps. Pp. 401-414 ae L. Gressitt, ed., Pacific Basin biogeography. Bishop Mus. Press, Honolulu. Kenc, H. ae Pinacins hypophyllus Hook. f. Gard. Bull. Singapore 20: 123— : cee Aspects of morphology of Phyllocladus hypophyllus. Ann. Bot. n.s. 27: 69-78 . 1974. The phylloclade of Phyllocladus and its possible bearing on the branch systems of progymnosperms. /bid. 38: 757-764. 1977. eae and its bearing on the systematics of conifers. Pp. 235-251 in K. Kuprrzxt, ed., Flowering plants: evolution and classification of higher categories. cue -Verlag, Wien & New York. Kirk, T. 1889. The forest flora of New Zealand. xv + 345 pp. George Dids- bury, Wellington. Krassrtov, V. A. 1974. Podocarpus from the Upper Cretaceous of eastern Asia and its bearing on the theory of conifer evolution. Palaeontology 17(2): 365-370. Li, H. L. 1953. Present distribution and habitats of the conifers and taxads. Evolution 7: 245-261. Preest, D. S. 1963. A note on the dispersal characteristics of the seed of New Zealand podocarps and beeches and their bi ological significance. Pp. 415- 424 in J. L. Gressitt, ed., Pacific Basin biogeography. Bishop Mus. Press, Honolulu. Ropertson, A. 1906. Some points in the morphology of Phyllocladus alpinus ook. Ann. Bot. 20: 259-265. WaRDLE, P. 1969. Biological flora of New Zealand. 4. Phyllocladus alpinus Hook. f. Cals Mountain toatoa, celery pine. New Zealand Jour. Bot. 7: DEPARTMENT OF BOTANY UNIVERSITY OF SINGAPORE Bukit TiMAH ROAD SINGAPORE 10 274 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 WOOD ANATOMY AND PHYLOGENY OF PAEONIA SECTION MOUTAN JosEPH M. KEEFE AND MayNarb F. Mose Ley, JR. THE GENUS Paeonia consists of thirty-three species of herbaceous and woody plants. In his classic work on the taxonomy of Paconia, Stern (1946) separated the genus into three sections: MouTAN. PAEONIA, and OnaAEPIA. The section Mouran contains four species of shrubs native to western China. The section Paronta has twenty-seven species of perennial herbs and occurs from Spain and North Africa through temperate Asia into Japan. The section Onaep1a, also consisting of perennial herbs, contains two species native to western North America. De Candolle (1824) placed Paconia within the family Ranunculaceae. Bartling (1830), however, suggested that the genus belonged in a unige- neric family, Paeoniaceae, near the Magnoliaceae. Since then. the phylo- genetic position of Paeonia has been the subject of considerable debate. Bartling’s position has been supported by superficial studies of the wood anatomy (Worsdell, 1908; Kumazawa, 1935). Subsequent investigations of the cytology and floral anatomy, however, have indicated that a more natural position would be in the Dilleniales, near the Dilleniaceae or the Crossosomataceae (Corner, 1946; Lemesle, 1955; Eames, 1961). Many authors have agreed on the need for more detailed studies on all aspects of this genus. In particular, there has been little study of the secondary wood anatomy. Previous anatomical studies have been concerned only with the amphicribral condition of the vascular bundles (Worsdell, 1908), or with the occurrence of secondary ray tissue and the presence of scalariform perforation plates in the vessels (Kumazawa, 1935). The genus Paeonia is noteworthy in possessing a combination of both advanced and primitive characters. The most important of these combina- tions is the occurrence of both woody (primitive) and herbaceous (ad- vanced) species. Also, the Paeoniaceae is the only family with herbaceous members (advanced) that possesses scalariform (primitive) vessel perfora- tions (Eames, 1961). The gross morphology of the genus is predominantly primitive; however, a great number of advanced characters may be ob- served in the wood anatomy. Of special interest is the section MOoutTan, in which the ancestral woody habit has been retained. In addition, this section occurs in a region (western China) noted for its many relict taxa of great age. MATERIALS AND METHODS The secondary xylem of 59 samples of the four shrubby species of Paeonia section Mouran has been examined, Species of the section Mov- TAN are distinguished from the other sections of the genus by their peren- 1978] KEEFE & MOSELEY, PAEONIA Di nial, woody stems and by thin petals which are much longer than the sepals. The section is divided into two subsections (Stern, 1946): I. Subsection VAGINATAE 1. Paeonia suffruticosa Andrews. Samples investigated: twenty-six. Native to the mountains of central China; Kansu, Szechwan, and Shansi provinces; elevation 2130-4270 meters (7,000-14,000 feet). Subsection DELAVAYANAE Paeonia delavayi Franchet. Samples investigated: ten. Native to the Likiang Ranges of Yunnan and southwest Szechwan provinces; elevation 3050-3660 meters (10,000-12,000 feet). —_ 4 3. Paeonia lutea Delavay ex Franchet. Samples investigated: twelve. Native to the Tali Range, from northern Yunnan Province north- west into southeastern Tibet; elevation 2740-3660 meters (9,000— 12.000 feet). 4. Paeonia potaninii Komarov. Samples investigated: seven. Native to the northern border of Yunnan north to western Szechwan Province, elevation 2740-3050 meters (9,000-10,000 feet) In addition to the samples listed above, four ne were investigated: three of Paconia delavayi lutea and one of P. lutea X potaninii. All samples cited in this paper are identified by the senior sear ae s numbers. Herbarium specimens, which have accompanied some of the wood samples, have been deposited in the herbarium of the University of California, Santa Barbara, California. Five slides for microscopic examination were prepared from each sample, according to standard methods (Wetmore, 1932). Sections were softened in 48 percent hydrofluoric acid, embedded in celloidin, and cut on a sliding microtome at 15 pm. Sections were stained in Heidenhain’s iron alum haematoxylin and safranin. A complete collection of slides is filed with the herbarium sheets. The terminology used in the anatomical descriptions complies with that approved by the Committee on Nomenclature, International Association of Wood Anatomists (1933, 1964). The choice of diagnostic characters was made largely from the lists prepared by Record and Chattaway (1939) and Tippo (1941). The most frequent range category in each case was arbitrarily determined. All measurements were taken at random. The classification of wood parenchyma distribution utilized here is a synthesis of the one presented by Metcalfe and Chalk (1950) as amended by the International Association of Wood Anatomists (1951). ANATOMY OF THE SECONDARY XYLEM An outline of the measurements recorded for the secondary xylem is presented in TaBLeE 1. GrowTH rincs. Growth rings are always present (FicureE 23), although TABLE 1. Secondary xylem of Paeonia. JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 CHARACTER I. Imperforate elements (Tracheids, fiber- tracheids) a. Length b. Wall c. Pit-pair horizontal diameters thickness II. Vessels a. Length b. Distribution c. Transectional diameters d. Number of bars e. Width between bars f. End wall angle from the vertical g. Intervascular pit diameters III. Vascular rays a. Tangential frequency b. Width of multi- serlate rays c. Width of uni- serlate rays d. Height of uni- serlate rays e. Height of multi- serlate rays RANGE Most FREQUENT RANGE MEAN 160-1224 wm (short—medium) very thin-thick 2-8 um, 97-1191 um. (short—medium) 11-58/mm.2 16-88 um. (extremely small- moderately small) 1-7 (few intermediate) O wm. (narrow-wide) 445° 2-10 um. (minute-large ) 3-16/mm. 20-124 um. (2-8 cells) (very fine-broad) 28-37 um. (moderately fine) rather low) 160-3000 um. (extremely low- rather low) 33-640 um. (short) thin-thick 3-5 um, 250-500 um. 20-40/mm.? 28-44 um. (very small) 1-3 (few) S-12 um, (wide ) 15-30° (very oblique) nw (small) 6-10/mm. 28-56 um (2 ors cells) (moderately fine ) 541 um. (short) 4.1 pm. 380 em. (medium) 31/mm.? 36.4 xm. (very small) 6 »m (small) (very low) 620 um. (very low) 1978] KEEFE & MOSELEY, PAEONIA aA) they are occasionally poorly defined. The annual increment varies from relatively narrow to wide. IMPERFORATE ELEMENTS. The imperforate elements are tracheids and fiber-tracheids. No libriform wood fibers or gelatinous or septate fibers were observed. The walls of the imperforate elements vary from very thin to thick, but are most frequently within the limits of thin to thick (FIGURE 24). The pit-pairs are bordered and generally circular. The inner apertures of all pits are elongated, and those of the pit-pairs are crossed. The elon- gated inner apertures do not ordinarily extend beyond the outline of the borders (FicurEs 25, 31). One hundred length measurements of the imperforate cells of each of the 59 samples were taken at random. The lengths of the elements vary from 160 (short) to 1224 ym. (medium sized). The most frequent range is from 33 to 640 pm. (short), and the mean for the genus is 541 + 7.04 um. (short), with a standard deviation of 89.7 + 1.16 pm. The proportion of elements with spiral thickenings is quite variable from sample to sample. In some samples, most of the elements have spiral thickenings; in others, this condition is rare or absent. Occasionally, the thickening is a double one, with one coil pitched in one direction and the other one in the opposite direction. VessELs. The number of vessels per square millimeter in transverse aspect was calculated from 10 microscopic fields of each sample. Vessel distribution is predominantly of solitary and chain arrangements, but multiples also occur frequently (FicureE 23). Pore clusters occur oc- casionally or rarely in 36 of the samples. Paeonia is usually distinctly ring porous (Ficure 23), although some rings are indistinct or are only semi-ring porous. The early wood character- istically has larger and more frequent pores than are found in the late wood. The late wood has small pores which are usually fairly evenly distributed. In all species, the vessels are angular in cross section (FIGURES 3, 24). Exceptions were observed in scattered vessels in a few samples of P. lutea, where some of the larger vessels were circular or nearly so. Perforation plates in the end walls of the vessel elements are most commonly of the scalariform type, but other multiperforate forms, as well as simple perforation plates, also occur (Ficure 31). The number of bars in typical scalariform perforation plates varies from 1 to 7 (few to intermediate) (Ficures 7-10), and varies most frequently from 1 to 3 (few). The widths of the openings among the bars vary from 6 (narrow) to 20 ym. No borders occur around the perforations. Of a random count of 2950 perforation plates, 3.7 percent were found to be simple (Ficures 1-3). The simple perforation plates usually occur in vessels with relatively narrow diameters, and are of three main types. The first type (Ficure 1) has typically round openings which are most often encountered in the more advanced dicotyledons (‘‘simple perfora- tions” sensu Esau, 1965). These perforations occur in vessels with end 278 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Ficures 1-22. Perforation ae as seen in radial section: 1-4, 6-8, 10-15, 17-20, seen suffruticosa; 5, 9, 21, P. delavayi X lutea; 16, P. potaninii: 22, P. lute walls most closely approaching the horizontal. The second type of simple perforation is vertically elongate (FicurE 3). The last type appears similar to a scalariform perforation with the cross bar partially unformed (FicurEs 2, 31). In rare instances, a vessel end wall appears to have two distinct perforation groups, of which one might be simple and the other multiple (FicurE 5). This latter type might be considered a single scalari- form perforation plate, however. 1978 | KEEFE & MOSELEY, PAEONIA 279 With the exception of the herbaceous Paeoniaceae, scalariform perfora- tion plates in dicotyledons are found exclusively in woody taxa (Eames, 1961). Scalariform vessels have been observed by the writers in Paeonia californica, a perennial herb with fleshy roots. The most common pattern in Paeonia section Moutan could be classified as “typically scalariform” (FicureEs 9, 10); in this form the horizontal bars are arranged in a ladder- like series along the flattened, oblique end wall. This corresponds to the form occurring most frequently in dicotyledonous vessel perforation plates (“scalariform” sensu Esau, 1965) Ten percent of the perforation plates examined were atypical multiper- forate forms. These occur in four basic types. The first, and most common type includes typically scalariform bars with the addition of vertical bars or bar-fragments (Ficures 8, 11-14). We have named this type modified scalariform. The second type has some plates including irregularly retic- ulated perforation patterns in oval to circular (Frcures 15, 18-20), scat- tered (Ficure 21), or elongate (FicurE 22) configurations. In the third type the plates approach foraminate configurations (Gray & de Zeeuw, 1974) (Ficures 16, 17). The fourth type includes forms transitional be- tween scalariform and simple perforation plates (Ficures 2-6). Both the first and second types would be considered irregular reticulate plates according to Gray and de Zeeuw (1974). A similar and distinctive perforation pattern tends to appear in suc- cessive end walls, ad seriatim, of several members (five to ten) of a single vessel. Distinctive patterns were not observed to be repeated in radially successive vessel members: that is, in derivatives of the same cambial initial. This indicates that a particular perforation pattern may be ex- pressed in a vessel as a unit, rather than in the successive derivatives of a single cambial initial. The end walls of the vessel elements in tangential aspect are generally quite oblique, ranging from 4 to 45° off the vertical, but vary most fre- quently from 15 to 30° off the vertical (very oblique) (FIGURES 25; 20): The intervascular pitting of the vessels varies from transitional to alter- nate (Ficures 25-27). Often there is a single staggered row of pit-pairs in the vessel wall (Ficures 25, 30). The scalariform arrangement was observed in two samples of P. suffruticosa, although rarely so even in these cases (FIGURE 26). The size of the intervascular pit-pairs varies from 2 (minute) to 10 pm. (large), but ranges most frequently from 5 to 7 mm. (small), with a mean of 6 »m. Twenty pit-pairs from each of the samples were measured. The intervascular pit-pairs are predominantly circular in outline, but are occasionally elongate, oval, or scalariform. Generally, the pit-pairs have elongate and crossed apertures which in face view do not extend beyond the outlines of the pit borders. VESSEL-RAY PITTING. The pitting between the vessel elements and the vascular ray cells is variable. Alternate and opposite pitting are the most common, but the transitional arrangement was found frequently in most samples (FicURE 28). The vessel-ray pits are without borders. 280 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 OVieri@re- * — e¢ ' * teete est eel Nea An TAS t¢ #4 FIGURES 23-27. 23, Paeonia suffruticosa, transverse section showing a distinct growth ring boundary (distinctly ring-porous), the angular vessel elements in solitary, chain, and multiple arrangements (Keefe 2322), * 185. 24, P. lutea, transverse section demonstrating thick- to thin-walled imperforate tracheary elements, a growth ring boundary, and axial wood parenchyma (Keefe coe ae 4 1978] KEEFE & MOSELEY, PAEONIA 281 VESSEL LENGTH. One hundred length measurements of vessel elements were taken at random from macerated wood of each of the 59 samples. The length of the elements from the tip of one “tail” to the other varies from 97 (short) to 1190 pm. (medium). The most frequent range is from 250 to 500 pm. (short to medium), and the mean for the genus is 380 + 4.95 um. (medium) with a standard deviation of 89.7 + 1.16 wm. The vessel elements are often storied (FicurE 28). Spiral thickenings occur frequently in the vessel elements, and, as in the imperforate elements, they are often double. These occur most often in smaller vessels. VASCULAR RAYS. The frequency of rays occurring horizontally across a tangential section is from 3 to 16 per mm., and more commonly from 6 to 10 per mm. The vascular rays are both uniseriate and multiseriate. No compound or aggregate rays occur. The ray type is not constant for the section Moutan, or for a particular species. The rays, furthermore, do not closely agree with the classical patterns as set down by Kribs (1935). The rays are often heterocellular, but the vertically elongate cells occur randomly throughout a particular ray, rather than exclusively in the wings or distal portions of the ray (FrcurE 30). The classical pattern with elongate cells only at the distal portion of a ray was not observed in any of the samples. Three basic types of rays were observed. The most common was a modified Heterogeneous Type II (all names of ray types taken from Kribs, 1935). These consist of 40 to 99 percent uniseriate rays, which were in the low or extremely low class in height (Ficure 30). Generally speak- ing, the samples which possess a greater percentage of uniseriates tend to have the higher uniseriate rays. The multiseriate rays are most com- monly bi- or triseriate. Again, the samples with the higher percentage of uniseriate rays tend to have a greater proportion of biseriate rays than of wider ones One sample of Paeonia suffruticosa (Keefe 1301) has Homogeneous Type I rays. In this sample the uniseriate rays comprise 20 to 30 percent of the total number of rays. The multiseriate rays are almost entirely biseriate with very long wings. The height is in the extremely low class. Although these rays are fairly homocellular, the individual ray cells as seen in face view on tangential sections are intermediate between the tradi- tionally defined round and rectangular-shaped cells. Three samples, two of Paeonia suffruticosa (Keefe 1051, 1504) and one of P. potaninii (Keefe 1268), would agree with Kribs’s classification of 280. 25, P. suffruticosa, tangential section illustrating imperforate tracheary elements with circular-bordered pit-pairs and crossed aper ertures (arrow), vessel a arenchyma (K eefe 2332), < 360. 26, P. delavayt, poco section showing essel sana with scalariform intervascular pitting, and portion of vessel ele- ment with opposite to alternate intervascular pitting (Keefe 2359), *_ 340. 27, P. delavayi, tangential section vib vessel element with transitional inter- vascular pitting (Keefe 2359), X 3 282 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 59 wee=-- Pots — OS Se me _ RT aah) Sige exe, BBE WS Me Bata Seiecneac oi Pebe Lig <> Ci. ’ OT Rares |? th ey ai rd le gre gee ah aoe es eee o> Figures 28-31. 28, Paeonia lutea, radial section showing storied vessel ele- ments and vascular rays with square and upright cells (Keefe 2351), X 60. 29, P. potaninii forma alba, tangential section showing Homogeneous Type II vascu- lar rays, with infrequent uniseriate and nearly homocellular multiseriate vascular rays, and 3-celled strand axial wood parenchyma (arrow) (Keefe 1268), X 1978] KEEFE & MOSELEY, PAEONIA 283 Homogeneous Type II rays. These samples have uncommon uniseriate rays (10 to 20 percent of the total number of rays), and frequent multi- seriate rays which are usually wingless (FicuRE 29). The rays are homo- cellular, being composed entirely of angular to nearly rounded cells as seen in tangential section. For the section MouTAN as a whole, the widths of the multiseriate rays vary from 20 to 124 wm. (very fine to broad) or from 2 to 8 cells, as measured horizontally in tangential sections. The majority of multiseriate rays vary within the range of 28 to 56 pm. (moderately fine) or 2 or 3 cells wide, with the exception of the three samples with Homogeneous Type II rays. The uniseriate rays most frequently range from 28 to 37 pm. in width (moderately fine). Ray height for the genus has a mean of 620 pm. (very low) for the multiseriate rays, with a range of 160 to 3000 um. (extremely low to rather low). The mean height for the uniseriate rays is 570 pm. (very low), and the range is 128 to 2430 pm. (extremely low to rather low). In number of cells, the uniseriate rays vary commonly from 1 or 2 cells to 75 cells high. AXIAL WOoD PARENCHYMA. Both strand and fusiform types of axial wood parenchyma occur in all four woody species of Paeonia. The most common type is strand parenchyma, with each strand consisting of two cells (FicurE 30, arrow). Strands with three cells (FIGURE 29, arrow) or fusiform cells (F1cuRE 31, arrow) occur less commonly. The axial wood parenchyma is apotracheal and sparse, and occurs in a scattered diffuse pattern among the imperforate elements. In both shape and wall thick- ness, the axial wood parenchyma is often very similar to the imperforate elements. DISCUSSION Taxonomic uistory. The relationship of the genus Paconia has long been problematic. The genus has been placed in the families Ranuncula- ceae, Berberidaceae, and Paeoniaceae (and included in the orders Ranales, Paeoniales, Dilleniales, Theales, and Guttiferales) by various authors (TABLE 2). CoMPARATIVE ANATOMY OF THE SECONDARY XYLEM. The following characters of the secondary xylem of Paeonia are strictly primitive: (1 presence of tracheids and fiber-tracheids, but absence of libriform wood fibers (Bailey & Tupper, 1918; Bailey, 1936, 1953; Reinders, 1935); (2) absence of septate imperforate elements (Tippo, 1938; Metcalfe & Chalk, pa at a ee . 30 end walls; essentially Heterogeneous Type II vascular rays with heterocellular mu and 2-celled strand axial wood parenchyma (arrow) (Keefe 2332), X 275. 31, P. suffruticosa, radial section showing vessel with typical scalariform perforation plates with 1 and 2 bars, vessel with atypical simple perforation plate, and fusiform axial wood parenchyma (arrow) (Keefe 817), * 240. 284 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 TABLE 2. Summary of systematic treatments of the genus Paeonia. De Candolle, 1824 Bartling, 1830 Bentham & Hooker, 1862 Baillon, 1867 Engler & Prantl, 1887, 1915 Hallier, 1905 Worsdell, 1908 Lotsy, 1911 Engler & Gilg, 1924 Rendle, 1925 Heintze, 1927 Langlet, 1928 Kumazawa, 1935 Engler & Diels, 1936 Corner, 1946 Lawrence, 1951 Lemesle, 1955 Benson, 1957 Hutchinson, 1959 Eames, 1961 Melchior, 1964 Dickison, 1967 Thorne, 1968, 1974 Cronquist, 1968 Takhtajan, 1969 Ranunculaceae, near Magnoliaceae Ranunculaceae Ranunculaceae Ranunculaceae Ranunculaceae Berberidaceae Paeoniaceae, in Ranales, between Ranunculaceae and Magnoliaceae- Calycanthaceae Berberidaceae Ranunculaceae Ranunculaceae, between Lauraceae and Berberidaceae Paeoniaceae, in Paeoniales, with Ranunculaceae, Berberidaceae, Lardizabalaceae and Menispermaceae Berberidaceae (including Berberidoideae, Podophylloideae, and Paeonioideae) Paeoniaceae, in Ranales, near Magnoliaceae Ranunculaceae Paeoniaceae, in Dilleniales, with Actinidia Ranunculaceae, in Ranales Paeoniaceae, in Ranales, between Magnoliaceae and Calycanthaceae- Crossosomataceae Ranunculaceae, between Hernandiaceae and Berberidaceae Paeoniaceae, in Ranales, between Magnoliaceae and Helleboraceae Paeoniaceae, in Dilleniales, near Crossosomataceae Paeoniaceae, in Guttiferales, near Dilleniaceae Paeoniaceae, in Dilleniales, near Dilleniaceae Paeoniaceae, in Theales, under Dilleniineae Paeoniaceae, in Dilleniales, with Dilleniaceae and Crossosomataceae Paeoniaceae, in Paeoniales, between Dilleniales (including Dilleniaceae and Crossosomataceae) and Theales 1978 | KEEFE & MOSELEY, PAEONIA 285 1950); (3) moderately to very numerous vessels per square millimeter (Frost, 1930a); (4) vessel elements typically angular (Frost, 1930a) ; (5) vessel elements very small in diameter (Frost, 1930a); (6) vessel elements with scalariform perforation plates (Bailey & Tupper, 1918; Bailey, 1944; (Frost, 1930a, 1930b); (7) vessel elements with oblique end walls (Frost, 1930a; Bailey, 1944); (8) presence of uniseriate and multiseriate rays (Heterogeneous Type Ila) (Kribs, 1935; Barghoorn, 1940, 1941); and (9) axial wood parenchyma scattered, diffuse, and apotracheal (Kribs, 1937; Metcalfe & Chalk, 1950; Carlquist, 1961). The following characters of the secondary xylem of Paeonia are rela- tively advanced: (1) imperforate elements most frequently with thin to thick walls (intermediate categories) (Bailey, 1936, 1953; Reinders, 1935); (2) imperforate elements and vessels typically with circular-bordered pit-pairs having crossed inner apertures which do not extend beyond the border outlines (Bailey, 1936; Eames & MacDaniels, 1947); (3) length of imperforate elements ranging most frequently from 330 to 640 pum. (short) (Bailey & Tupper, 1918; Chattaway, 1936; Bailey, 1953); (4) vessel distribution commonly of pore chains and multiples in addition to the more primitive solitary condition (Tippo, 1946); (5) ring porosity or semi-ring porosity, ring porosity predominant (Frost, 1930a); (6) number of bars composing the scalariform perforation plates most frequently 1 to 3 (few) (Frost, 1930b); (7) the common presence of wide and (8) non- bordered perforations in the scalariform perforation plates (Frost, 1930b); (9) occasional simple perforation plates (Bailey & Tupper, 1918; Bailey, 1944; Frost, 1930a, 1930b); (10) opposite as well as alternate intervascular pitting (Bailey & Tupper, 1918; Frost, 1931); (11) alternate and opposite as well as transitional pitting between rays and vessels (Bailey & Tupper, 1918; Frost, 1931); (12) length of vessel elements ranging from 250 pm. (short) to 500 »m. (medium) (Bailey & Tupper, 1918: Chattaway, 1936; Bailey, 1953); (13) homogeneous rays (Types I and IT) (Kribs, 1935; Barghoorn, 1941); (14) presence of strand axial wood parenchyma, most commonly composed of two cells, rarely accom- panied by true fusiform axial parenchyma (Metcalfe & Chalk, 1950); (15) reduced vascular ray structure, with uni- or biseriate rays predominating (Metcalfe & Chalk, 1950); and (16) storied vessel elements (Metcalfe & Chalk, 1950). The following characters are specialized, but it has not been shown that they represent an advanced condition: (1) occurrence of spiral thick- enings in the imperforate elements (Tippo, 1946); (2) presence of spiral thickenings in the vessel elements (Frost, 1931); and (3) presence of atypical multiperforate perforation plates in the vessel elements (Mac- Duffie, 1921; Chalk, 1933; Gottwald & Parameswaran, 1964). The preceding lists support the hypothesis that the secondary xylem of Paconia cannot be considered primitive. Although several characters are of a primitive nature, a greater number are moderately specialized, and a few are highly specialized. In view of this evidence, one must tentatively 286 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 conclude that the family Paeoniaceae has reached a moderate level of specialization. It is difficult to attach any phyletic importance to the presence of the atypical multiperforate perforation plates. Atypical plates have been re- ported in many unrelated families. Aldridge (1964) described a definite abnormality of the regular perforation in tomatoes. This is the pheno- typic expression of a genetic aberration. Chalk (1933) and Gottwald and Parameswaran (1964) found plates composed of very numerous isodia- metric openings which are similar in both size and shape to the bordered pits on the side walls of the vessels. There is no evidence reported of fusion between these perforations within a plate. Multiperforate plates may occur intermittently in a plant and are usually found in taxa with otherwise simple perforation plates. This type undoubtedly has no significance in the evolution of the vessel as a whole, but rather is an anomaly in which the pattern of the intervascular pitting is imposed on the end walls. The atypical perforation plates found in Paconia represent modifications of scalariform or reticulate plates or are retentions of more primitive per- foration plates like those described for other taxa (Thompson, 1923; Dickison, 1967: Gray & de Zeeuw, 1974). Gottwald and Parameswaran (1964) described similar forms as occurring in the Dipterocarpaceae and believed that they were intermediate steps in the evolution of simple per- foration plates from scalariform ones. This is very possibly the case in Paeonia, where both scalariform and simple perforation plates exist. In his study on the secondary xylem of the Compositae, Carlquist (1959, 1960) occasionally found multiperforate plates, often in radial series, associated with the normally occurring simple perforation plates. This fact indicates that a single cambial initial tends to produce a series of atypical perforation plate patterns. This situation is different from the condition in Paconia, where similar atypical patterns were found to occur in successive end walls of a single vessel, rather than in radial series. Atypical multiperforate plates have been reported sporadically, but not infrequently, in scattered families of angiosperms, including both primitive and advanced taxa, as well as in ferns and other lower groups. MacDuffie (1921) states that this condition is particularly prevalent in dicotyledons with herbaceous and climbing habits, but this is not confirmed by a search of other papers treating the subject (Chalk, 1933: Thompson, 1923; Web- ber, 1936; Bailey & Howard, 1941: Heimsch, 1942: Young, 1955: Carl- quist, 1959, 1960, 1961; Eames, 1961: Gottwald & Parameswaran, 1964; Dickison, 1967). TABLE 3 shows a comparison of the various characters of the secondary xylem of Paeonia with those of other taxa with which it has been variously allied in the past. In summary, on the basis of the secondary wood anatomy alone, the woody species of Paconia are similar to, but more advanced than, the Dilleniaceae and Actinidiaceae, and are similar to, but slightly more primitive than, the Calycanthaceae and Annonaceae. Borderline taxa, in which there is a possibility of relationship, but much less than in com- 1978] TABLE 3. Comparison e IMPERFORATE ELEMENTS Type Length P Spiral thickenings P VESSELS Porosity + Pore distribution P Pore diameter A Cross-sectional outline + Perforation plates P Number of bars Pp Intervascular pitting P Ray-vessel pitting P Ray-vessel pits bordered + Length P Spiral thickenings + VASCULAR RAYS Type P Widths of multiseriates A Sheath cells common fe AXIAL WOOD PARENCHYMA Type + Number of cells in strand p Parenchyma distribution A Percentage of characters 33 identical to those of Paeoni Percentage of characters 52 of a lower phyletic level L than those of Paeonia Percentage of characters 15 of a higher phyletic level é ; han those of Paeonia + Family Winteraceae 60 KEEFE of secondary Canellaceae & MOSELEY, Annonaceae A Myristicaceae Calycanthaceae xylem PAEONIA of Paeonia ca v a B oS Dilleniaceae Crosso Rosaceae I+ 25 Spiraeoideae and Rosoideae other 7) 4 iss] ou 2 2 2 ee [ou + P P + + + P P - A A A A A A (0) 0 A A A A + + P 3s + + + + A A + + + + oO 0 A A 32 32 26 32 42 36 taxa. 2 3 Qo og soon v a] x (2) e < + + P ae) Bp P - P P Pp A + Bs (0) P A P P P ie) P 0 P 0) P P + + P P + A A + Oo + 0) P A+ 26 «38 58 50 16 12 P Character of the family in question is less specialized than that of Paeonia A Character in question is more specialized than that of Paeonia 0 Knowledge of character is lacking, or phyletic nature not understood 207 Saurauiaceae Guttiferae n question has essentially the same secondary xylary feature as found in Paeonia 288 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 parison to the families noted above, are the families Magnoliaceae and Berberidaceae, the rosaceous subfamilies Maloideae and Prunoideae, and the ranunculaceous genus Clematis. About one-third of the secondary xylary characters of these latter taxa lie at a phyletic level similar to those of Paeonia. CORRELATION OF TAXONOMIC AND MORPHOLOGICAL FEATURES WITH XYLEM CHARACTERS. It is unwise to speculate concerning phyletic relationship solely on the basis of secondary wood anatomy (Bailey, 1951, 1953). Considering that similar structures appear in quite distantly related taxa due to the frequent occurrence of parallel evolution, no single line of evidence should be considered conclusive. Since a taxon which retains primitive secondary wood characters cannot be derived from one which is uniformly advanced, Bailey (1957) has pointed out that the evidence obtained from secondary wood anatomical studies is often more helpful in negations than in positive assertions of close alliance. Bailey (1951, 1953, 1957) has also stressed that meaningful phylo- genies can be constructed only when comprehensive studies, taking into account evidence from all organs and parts of the plant body, have been evaluated. For this reason, Taste 4 has been prepared to show differences and similarities of features other than those of the secondary wood, between Paeonia and each of the taxa with which it has been linked. As familial taxonomic descriptions are readily available, they are not included here. The principal works used in compiling Taste 4 are those by Rendle (1925), Gunderson (1950), Lawrence (1951), Hutchinson (1959), and Melchior and Werdermann (1964). The characters chosen for comparison are those for which information is available in these standard works. The morphological characters of the Paeoniaceae that have been shown to be primitive (Hutchinson, 1959: Eames, 1961: Davis & Heywood, 1963) include: woody and perennial habit (section Movuran);_ spiral arrangement of leaves; bisexual, solitary, and actinomorphic flowers with many free parts, spirally arranged; hypogynous stamens; several free car- pels; and follicular fruit. Those characters of Paeonia which are commonly considered advanced or specialized include herbaceous habit (sections PAEONIA and ONAEPIA): deciduous and pinnately compound leaves; ab- sence of stipules; fasciculate stamens; and tricolporate pollen. In addition, Camp and Hubbard (1963) found an abundant vascular supply subtend- ing the ovules of Paeonia lactiflora (section PaEONIA). From this they sug- gest that the ovules in Paeonia are specialized through reduction from an ancestral type. It can be seen, therefore, that the Paeoniaceae may be considered to have predominantly primitive characters. The relatively specialized wood anatomy, however, possibly invalidates the placement of the family in the traditional positions near families that have similar primitive morphology, but that possess dissimilar primitive wood anatomy. TABLE 4. Comparison of morphological-taxonomic Habit woody Vines absent Deciduous Stipules absent Leaves alternate Leaves pinnately compound Flowers solitary Flowers bisexual Bracts present Sepals not fugacious Sepals 5 Petals 5-10 Stamens many Stamens spiral Stamens hypogynous Stamens fasciculate Stamens centrifugal Anthers opening lengthwise Disc present Nectaries present Stigma sessile Apocarpous Carpels 2-5 Ovules several Ovules anatropous Placentation parietal Fruit a follicle Pollination by beetles Pollen 3-colporate Endosperm copious Embryo small Nucellus crassinucleate Integuments 2 Aril present Raphides absent Basic chromosome number Number of characters h some members Percentage of characters identical to those of Paeonia in at least some members v a co) o vy 4 ® a o 0. o a w u a a ] © uy oO J 7 y 7 a oa o 9 i) we) no rey S ° o c a 6 ec ow v o «A a o ¢ ¢ = Me o G a % Ee > oe = = o < a + + + = +o + = = t= + +t + + os - + = + * 3 = i) sf + + + = jab + - - = 0 0 + = 0 = 0 a + - + = - + to = £ + = = t+ o- + fF +t - + + + + > 10 13-13 7, 7 19 13 23. 23) 23- 23:23 w wa 100 66 75 73 73 65 + Characters identical - Characters different Characters identical I+ OQ Character unknown v « wv a e ov oO 3 i] Vv 3) 7) a bu v % 3 Oo on ce) 3) v < % uw} 2 a a) vo ro a v i c 3 El —| = ood % 3) 3] be c 2) ec = u v cy} >» 3 ol = o = a E oo % % eon Oo «2 a a + + + + + + - + + = + + + = + = + + + - + - + + + + + + + + + - Oo + O - + = + + - + - + + - 5 Se Sees 2s + + + + + - = = + = + Fig lates tei UG achie ihe + + + + + ee + + - + + + eh ese ey Wee et + + + + 15 26 18 24 21 28 47 74 51 71 62 85 to those of Paeonia from those of Paeonia Crossosomataceae Rosaceae character-states. v Ll uv U0 ou od oO ° v uv Uv 0 A wt ° a8 a ° wn” [4 + + + + + + + + - + = te + + + af + + + + + + + + + + + + ie} 10) 7 + t+ + + + e- e + + + + + + a 1s + 8, 7, 9 8 25. 223. 68 68 Maloideae Prunoideae at least in some members to those of Paeonia Theaceae 60 Actinidiaceae 58- 52 Saurauiaceae He Guttiferae 290 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Members of the family Dilleniaceae possess the highest number of mor- phological characters similar to those of Paeonia (85 percent, TABLE 4). In addition, it is one of the five families most possibly related to Paeonia on the basis of the secondary wood anatomy alone. Morphologically, the families are extremely similar, differing, in characters considered, only by the presence in Dilleniaceae of occasional climbing forms, usually simple leaves, a free stigma, a basic chromosome number of x = 8 to 13. and the absence of a disc and nectaries. Paeonia can be seen to be more specialized than species in the family Dilleniaceae, especially in its deciduous habit and in its secondary wood, which exhibits such specialized features as ring or semi-ring porosity, short imperforate and vessel elements, a reduction in the number of bars in the scalariform perforation plates (from 20 to 130 down to | to 3), and a reduced vascular ray structure (TABLE 3). Takhtajan (1969) considers Paeonia the component of an unigeneric family and order, thereby separating it somewhat more from the Dilleni- aceae than do the present writers. His separation is based upon the follow- ing morphological characters: (1) thick, fleshy carpels, (2) broad stigmas, (3) a lobed and unusually prominent staminal nectariferous disc. (4) a massive outer integument, (5) seed coat characters, and (6) the peculiar embryogeny described by Yakolev and Yoffe (1957), Cave, Arnott, and Cook (1961), and Walters (1962). Because of the frequency with which Paeconia has been placed within the Ranunculaceae, a discussion of the relationship between these two taxa seems appropriate. The similarities between the Paeoniaceae and Ranun- culaceae (and often with the “Ranales” in general) include: usually peren- nial and herbaceous habit; alternate, more or less divided leaves: bisexual flowers with free and spiral or spirocyclic parts: numerous stamens: free and numerous carpels; follicular or achenial fruits: and seeds containing copious, oily endosperm with a small embryo at the apex. The separation of Paeonia from the Ranunculaceae, however, is strongly indicated in TaBLe 5, in which the major and fundamental differences between the two taxa are listed. Of the differences separating Paeonia and the Ranunculaceae. several characters of Paeonia also separate that genus from the “Ranales” in general (sensu Eames, 1961): amphicribral vascular bundles, fasciculate and centrifugal stamens, and several characters of the ovule, pollen, and secondary wood. The “Ranalian” families that are most similar to the Paeoniaceae on the basis of the secondary wood anatomy (TABLE 2) are the Calycanthaceae and the Annonaceae, with one-half of the total characters used similar to those found in Paeonia. The Calycanthaceae, however, are quite dissimilar to Paeonia in morphological features (TABLE 4) and are therefore unlikely to be related. The family Annonaceae is chiefly tropical and seems to be a reduced or specialized taxon closely related to the Magnoliaceae. Characters which show this relationship are the whorled perianth, exstipulate leaves, and ruminate endosperm, In these features, the Annonaceae are most closely 1978 | KEEFE & MOSELEY, PAEONIA 291 TABLE 5. Comparison of Paeonia and Ranunculaceae. CHARACTER Paeonia Ranunculaceae 1. Habit Woody and herbaceous Herbaceous perennials, perennials woody vines (Clematis), or rarely shrubs (Xanthorrhiza) 2. Vascular bundles Amphicribral Frequently amphivasal 3. Perianth Gradation gradual Perianth sharply delimitation 4. Receptacle 5. Vascular supply to perianth 6. Petal derivation 7. Sepals 8. Stamens 9. Subgynoecial disc 10. Ovule development 11. Seed 12. Pollen 13. Seed germination 14. Basic chromosome number 15.* Imperforate elements 16.* Vessels es Sheath cells 20. Accessory cortical bundles from leaves through bracts, sepals, and petals Somewhat concave Sepals and petals receiving few to several vascular traces From bracts Persiste nt Fasciculate, developing With outer integument strongly developed and thick; nucellus crushed, hypostase invariably present; antipodal cells ephemeral Aril present 3-colporate Hypogeal Xylary fibers and fiber- trachei Small and often solitary, with scalariform and simple perforation plates Absent in vascular rays Apotracheal With little or no sclerenchyma Present delimited from ‘leaves and bracts Convex Sepals typically receiving three traces, petals receiving one trace Probably from stamens Fugacious Free, developing centripetally Absent Outer integument similar to the inner oe ent; nucellar epider persisting and leat the embryo sac, hypostase varely present; antipodal cells ene Aril absent 3-colpate Epigeal (except in part of Clematis) n Libriform wood fibers Large and usually clustered, with simple perforation plates only Common in vascular rays Paratracheal With many fiber and stone cells Absent * Items 15 through 18 present a comparison of the secondary xylem of Paeonia with that of Clematis. The latter genus provides the only basis for comparison in this respect. 292 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 related to the more advanced genera of the Magnoliaceae, such as Lirioden- dron. The Magnoliaceae and Annonaceae resemble one another in so many ways that the two families are doubtless derived from the same ancestral stock, with the Annonaceae being somewhat more advanced. This advance- ment has brought the Annonaceae within its alliance up to a phylogenetic level similar to that of Paconia. Because of the basic differences between Paconia and the “Ranales”, and the probability of close relationship be- tween the Magnoliaceae and the Annonaceae, the similarities of Paeonia and the Annonaceae are probably the result of parallel evolution, rather than of close relationship. According to TaBLe 3 (wood anatomy), the Theales, including the Theaceae, Actinidiaceae, and Saurauiaceae, seem to be closely related to the Dilleniales. Morphologically, the family Theaceae represents a direct syncarpous derivative from the more primitive order Dilleniales. The alliance of the Dilleniaceae to the Theaceae (Thorne, 1968, 1974) is sup- ported by anatomical considerations. The greatest agreement between the anatomy of the two families exists between the subfamily Dillenioideae and the tribes Actinandrieae and Ternstroemieae in the Theaceae (Dickison, 1967 The Guttiferae are most often placed in their own order in close associa- tion with the Theales. The Guttiferae (Theales and Dilleniales) are also seen as advanced syncarpous types, with opposite leaves, and free or vari- ously connate stamens. The stamens are often united into specialized bundles (phalanges). Morphologically, the Guttiferae show little relation- ship to Paeonia, and the secondary wood anatomy is completely dissimilar (TABLE 3). These two orders, the Theales and the Guttiferales, were more likely derived from a dilleniaceous stock in one evolutionary direction. The secondary wood of the Actinidiaceae shows a phylogenetic level similar to that of the secondary wood in Paeonia, but this is not necessarily indicative of close relationship. The different direction taken by these taxa may be seen in the development of carpels and stamens that are more united, a lianous habit (in Actinidia and several genera of the Dilleniaceae), and opposite leaves (Guttiferae). Members of the family Rosaceae are characterized by the usual presence of stipules, a generally pentamerous flower, a hypanthium (in most genera), and the near absence of endosperm in the seed. This family represents a large and eminently successful modification from the “Ranales” by the development of a hypanthium. It has been suggested that the Dilleniales represent the link between the “Ranales” and the Rosales (Porter, 1959). This view, however, is questionable on the basis of specializations found in the Dilleniales which apparently have been retained by other taxa arising from the same ancestral stock, but which are not found in the Rosales (e.g., the centrifugal stamens which are found in the Dilleniales and Theales, but not in the Rosales). Characters found in the Rosaceae which show no similarities to any condition in the Dilleniaceae or in Paeonia include the hypanthium, leaves with stipules, free and centripetal 1978] KEEFE & MOSELEY, PAEONIA 293 stamens, axile placentation, nearly absent endosperm, and the absence of arillate seeds. In total number of morphological characters, Paeonia and the Rosaceae might be considered somewhat similar (TABLE 4), but the important differences already mentioned do not support such a conclusion. Also, treating the four subfamilies separately, as was done in the anatomi- cal treatment, reduces the number of morphological similarities to less than that shown by the family as a whole. Although the Maloideae and Prunoideae exhibit fewer morphological characters similar to those in Paconia than do the other subfamilies, they are anatomically closer to the Paeoniaceae than are the Spiraeoideae or Rosoideae. This is further evi- dence of the probability that the two taxa are not closely related. The monogeneric family Crossosomataceae has been included either in the Rosales (Gunderson, 1950; Thorne, 1968) on the basis of the hypan- thium, or in the Dilleniales (Hutchinson, 1959; Eames, 1961) on the basis of the free follicular carpels and strongly arillate seeds. Metcalfe and Chalk (1950) and Dickison (1967), however, question the placement of Crossosoma within the Dilleniales because of their rather dissimilar anat- omy. This view is supported by the lack of similarities between the sec- ondary xylary anatomy of Paeonia and that of Crossosoma (TABLE 3). In general, the secondary wood of Crossosoma contains more highly advanced features than does that of Paeonia. Both of these taxa are probably northern derivatives and end-line specializations of larger tropical groups. CONCLUSIONS e secondary wood of Paeonia shows both primitive and advanced features, with the majority of characters being relatively specialized or advanced. The comparative morphology of the taxa previously held to be related to Paeonia clearly indicates that relationship to the Dilleniaceae is very close. The secondary wood anatomy does not negate this conclusion, but further suggests that Paeonia is most probably best placed as a valid and distinct unigeneric family within the Dilleniales, somewhat more specialized than the Dilleniaceae. There is a great deal of evidence from morphology, wood anatomy, and cytology to show that Paeonia should not be linked with the Ranunculaceae; furthermore, similar evidence nearly as strongly negates close relationship with the other taxa discussed above except for the Theales. It is noteworthy that the authors’ association of Paeonia with the Dil- leniaceae agrees with the concepts of Corner (1946), Eames (1961), Melchior and Werdermann (1964), Dickison (1967), Thorne (1968, 1974), Cronquist (1968), and Takhtajan (1969). The evidence assembled by the writers, however, does not indicate that the relationship of Paeonia with the Crossosomataceae is as significant as is implied by the systems of Eames (1961), Cronquist (1968), and Takhtajan (1969). Closer to our views are those of Corner (1946), Dickison (1967), and Thorne (1968, 1974) that suggest a relationship of Paeonia, Dilleniales, and Theales, ex- cluding Crossosomataceae. The proportions of similar anatomical and 294 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 59 morphological-taxonomic characters assembled in TaBLeEs 3 and 4 indicate that Paeonia should be included as a distinct family within the Dilleniales. Our assemblage of characters also shows the probability of a close relation- ship between the Dilleniales (including only the Dilleniaceae and Paeoni- aceae) and the Theales. e Paeoniaceae probably arose from dilleniaceous ancestors in the Indo-Malayan region, the present center of distribution of the Dilleni- aceae, and migrated northward into western China. The ancestral types still persist in the higher mountains of western China as the section Moutan. The more specialized herbaceous forms have spread west across temperate Eurasia and northeast into western North America. ACKNOWLEDGMENTS Grateful thanks are due to Dr. Katherine Esau, for helping to interpret the perforation plate types, and to Mr. Robert Gill for technical assistance. LITERATURE CITED ALprRIDGE, N. H. 1964. Anomalous vessel elements in wilty-dwarf tomato, Bot. Gaz. 125: 138-142. Bartey, I. W. 1936. The problem of differentiating and classifying tracheids, fiber-tracheids, and libriform wood fibers. Trop. Woods 45: 1 The development of oe in wee and its significance in ee apo research. Am. Jour. Bot. 31: —428. 19 The use and abuse of tonal Fas in the study of phylogeny and ae erae Phytomorphology 1953. Evolution of tracheary cee of jana plants. Am. Jour. Bot. 40: 1957. The potentialities and limitations - wood anatomy in the phy- logeny ane classification of angiosperms. Jour. Arnold Arb. 38: 243-254. : R. A. Howarp. 1941, The comparative ioe of the Icacinaceae. II. Vessels, Jour. Arnold Arb. 22: 171-187. & W.W. Tupper. 1919, Size variation in tracheary cells; I. A comparison between the secondary xylems of vascular cryptogams, gymnosperms and angiosperms. Proc. Am. Acad. 54: 149-204. BAILLON, H. 1867. Histoire des plantes. Vol. I. 476 pp. Hachette, Paris. BARGHOORN, E. S. 1940. The ontogenetic development and phylogenetic special- ization of rays in the xylem of dicotyledons. I. The primitive ray structure. Am. Jour. Bot. 27: 918-928. . 1941. The ontogenetic development and phylogenetic specialization of rays in the xylem of dicotyledons. II. Modification of the multiseriate and uniserlate rays. /bid. 28: 273-282 BarTLING, I’. G. 1830. Ordenes naturales plantarum. 498 pp. Sumtibus Diete- richianis, Gottingae. Benson, L. 1957. Plant classification. 688 pp. D. C. ee Boston. BENTHAM, G., & J. D. Hooker. 1862. Gen. Pl. 1: 1-15. L. Reeve & Co., London. Camp, W. H., & M. M. Hupparp. 1963. Vascular spl and structure of the ovule and aril in paeony and of the aril in nutmeg. Am. Jour. Bot. 50: 174-178. 1978 | KEEFE & MOSELEY, PAEONIA 295 CANDOLLE, A. P. pe. 1824. Prodr. Syst. Nat. Reg. Veg. 1: CarLouist, S. 1959. Wood anatomy of Helenieae (Compositae). “Trop. Woods 111: 19-39. _ 1960. Wood anatomy of Cichorieae (Compositae). bid. 112: 65-91. _ 1961. Comparative plant anatomy. 146 pp. Holt, Rinehart & Winston, New York. Cave, M. S., H. J. Arnott, & S. A. Coox. 1961. Embryogeny in cs pape peonies with reference to their taxonomic positions. Am. Jou . 48: 3 404. CuHaLk, L. 1933. Mul pemonate plates in vessels with special reference to the Bignoniaceae. Forestry 16-25 CuattTaway, M. M. 1936. Relation between ae and cambial initial length in dicotyledonous woods. Trop. Woods 46: 16-20 CoMMITTEE ON NOMENCLATURE, International eo of Wood Anatomists. LOS Si eee glossary of terms used in wood anatomy. Trop. Woods 107: 1- : cs “Multilingual glossary of terms used in wood oe 186 pp. Verlanganstalt Buchdruckerei Konkordia Winterthur, Zur Corner, E. J. 1946. Centrifugal stamens. Jour. Arnold Arb. 17 423-437. Cronquist, A. 1968. The oe and classification of flowering plants. 396 pp. Houghton Mifflin Co., Davis, P. H., & V. H. peraes vi Principles of angiosperm taxonomy. 556 .D. Van Nostrand, Princeto DICKISON, W. A067. Cope morphological studies in Dilleniaceae; I. Wood anatomy. Jour. Arnold Arb. 48: 1-29. EAMES, - J. 1961. Morphology of the ste et 518 pp. McGraw-Hill, New Yo —— : a H. MacDanic.s. 1947. An introduction to plant anatomy. 427 pp. McGraw-Hill, New York. Esau, K. 1965. Plant anatomy. 767 pp. John Wiley & Sons, New York. ENGLER, A.. & L. Diets. 1936. Syllabus Pflanzenfam. ed. 11. 419 pp. Gebruder Borntraeger, Berlin. —____ & E, Giic. 1924. Syllabus Pflanzenfam. eds. 9, 10. 420 pp. W. Engel- mann, Leipzig. Frost, F. H. 1930a. Specialization in secondary xylem of dicotyledons. II. Origin of vessels. Bot. Gaz. 89: 67-94. _1930b. Specialization in secondary xylem of es II. Evolution of the end wall of vessel segment. /bid. 90: 198- _ 1931. Specialization in secondary xylem of sctyedons III. Specializa- ion of lateral wall of vessel segment. /bid. 91: GotrwaLp, H., & N. PARAMESWARAN. 1964. ee Geftsdrshbreshungen in der Familie Dipterocarpaceae. Zeitschr. Bot. 52: Gray, R. L., & C. H. ve Zeeuw. 1974. Terminology tr multiperforate plates in vessel elements. Bull. Int. Assoc. Wood Anat. 2: GUuNbERSON, A. 1950. Families of dicotyledons. 237 pp. Gane Botanica, Waltham. Hauer, H. 1905. Provisional scheme of we natural (phylogenetic) system of flowering plants. New Phytol. 4: 151- Hetmscu, C. 1942. Comparative aes a the secondary xylem of the “Cruinales” and “Terebinthales” of Wettstein with reference to taxonomic grouping. Lilloa 8: 83-198. 296 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Herntze, A. 1927. Comofyternas fylogeni. 170 pp. Hakan Ohlssons Boktryckeri, Lu nd. HutcuHinson, J. 1959. Families of flowering plants. I. Dicotyledons. 792 pp. The Clarendon Press, Oxford. INTERNATIONAL AssoctaTioN or Woop AnaTomisTs. 1951. Parenchyma. News Bull. Feb. 1951: 10-18. Kriss, D. A. 1935. ete lines of structural specialization in the wood rays of dicotyledons. Bot. Gaz. 96: 547-557. 1937. Salient lines of structural pap sar in the wood parenchyma of dicotyledons. Bull. Torrey Bot. Club 64: 177-187. Kumazawa, M, 1935. The structure and ae - Paeonia. Bot. Mag. Tokyo 49: 306-315. LANGLET, O. 1928. Einige er a uber die Zytologie der Berberidazeen. Sv. Bot. Tidskr. 22: 169- LAWRENCE, G. H. M. 1951. ae of flowering plants. 823 pp. Macmillan, New York LeMESLE, R. 1955. Contribution a l'étude de quelques arae de dictotylédones considérées comme primitives. Phytomorphology Lotsy, J. P. 1911. Vortrage iiber botanische ee ee Vol. II. 815 pp. G. Fischer, Jena. MacDurrig, R. C. 1921. Vessels of the Gnetalean type in angiosperms. Bot. Gaz. 71: 438-445, Metcuior, H., & E. WERDERMANN. 1964. A. Engler’s Syllabus der Pflanzenfami- lien, Zw lfte Auflage, Band II. 666 pp. Gebriider Borntraeger, Berlin METCALFE, C. R., & L. CHark. 1950. Anatomy of the dicotyledons. Vols. > 1500 pp. The Clarendon Press, Oxford. Porter, C, L. 1959. Taxonomy of flowering plants. 472 pp. W. H. Freeman, San Francisco. PRANTL, K. A. 1888. Ranunculaceae. Jn: A. ENGLER & K. A. PRANTL, Nat. Phase te 2: 43-66. Gebriider Borntraeger, Berlin. Recorp, S. J., . M. Cuattaway. 1939. List of anatomical features used in ee Ree irae woods. Trop. Woods 11-16. REINDERS, E, 1935. Fiber-tracheids, libriform wood ve and systematics in wood anatomy. Trop. Woods 44: 30-36. RENDLE, A. B. 1925. The classification of flowering plants. Vol. II. Dicotyledons. 636 pp. University Press, Cambridge, Englan STERN, F. C. 1946. A study of the genus Paeonia. ‘155 pp. Royal Horticultural Society, London. Takytajan, A, 1969. Flowering plants: origin and dispersal. (Translated by . JEFFREY.) 310 pp. Oliver & Boyd, Edinburg THompson, W. P. 1923. The oe of the Aiterent types of angiosperm vessels. Ann. Bot. 37: 183- ee R. F. 1968. Synopsis ed a putaively phylogenetic classification of the flowering plants. Aliso 6(4): 57-66. . 1974. A phylogenetic classification of the Annoniflorae. Jbid. 8: 147- 9 Trppo, O. 1938. ae anatomy of the Moraceae and their presumed allies. Bot. Gaz. 100: 1-99. 1941. A list of agnostic a for descriptions of dicotyledonous woods, Trans. Ill. Acad. 34: 106. 1978] KEEFE & MOSELEY, PAEONIA 297 . 1946. The role of wood anatomy in phylogeny. Am. Midl. Nat. 36: 362-372. Watters, J. L. 1962. eer and gametophyte selection in Paeonia californica. Am. Jour. Bot. 49: 787-794 WEBBER, I. E. 1936. Systematic anieuy of the woods of the Simarubaceae. Am. Jour. Bot. 23: 577-587. Wetmore, R. H. 1932. The use of celloidin in botanical technic. Stain Tech. : 36-62. WorsDELL, W. C. 1908. The affinities of Paeonia. Jour. Bot. London 46: 114-116. YAKOVLEV, M. S., & M. D. Yorre. 1957. On some peculiar features in the embryology af Paeonia L. Phytomorphology 82. Youna, R. L. 1955. The xylem anatomy of ee (Julianiaceae). Trop. Woods 101: 29-43. J. M. K. M. F. M. B1ioLocy DIVISION DEPARTMENT OF —— SCIENCES GLENDALE COMMUNITY COLLEGE UNIVERSITY OF CALIFORNIA GLENDALE, CALIFORNIA 91208 SANTA BARBARA, een 93106 AND Los ANGELES COUNTY MusEUM OF NATURAL HISTORY 298 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 A SPECTACULAR BUCHNERA (SCROPHULARIACEAE) FROM COLOMBIA D. PHItcox DURING A STupDy visit to the Gray Herbarium and Arnold Arboretum, in preparation for a revision of the extra-African Old World Buchnera, the author discovered several specimens of an unidentified plant appear- ing to represent a species of that genus from South America. The de- scriptive labels, which gave the flower color variously as scarlet, vermilion, orange-red, or red, cast doubt on its identity. Hitherto, all other species of Buchnera have had flower colors in the range from mauve through blue to pinkish lilac or white, but never red. However, following my eventual study of this plant together with the excellent color photographs accom- panying Zarucchi 2143, I have no hesitation in describing it as new. Buchnera rubriflora Philcox, sp. nov. Ab species aliis Buchnerae in floribus rubris et capsulis leviter com- pressis differt. Annual or perennial herb 20-60 cm. tall. Stem slender, erect, simple or branched from the base, minutely pubescent with adpressed antrorse hairs to subglabrous. Leaves: cauline, 12-30(—45) mm. long, 1—2(—4) mm. wide, opposite to subopposite, narrowly linear, entire, 1-nerved, glabrous to minutely scabrid; basal, 8-22 mm. long, 4-7.5 mm. wide, obovate to obovate-oblong, narrowed into petiole-like base up to 2 mm. long, pin- nately nerved, all nerves prominent beneath, with secondary nerves strongly reticulated. Inflorescence slender, lax to compact, up to 8 cm. long; bracts ovate, up to 2.5 mm. long, shortly ciliate, otherwise glabrous; bracteoles linear, shorter than bracts. Calyx erect, to about 4.5 mm. long including teeth 2 mm. long, glabrous except for shortly ciliate teeth, clearly 10-nerved with no intermediate nervation, teeth becoming patent in fruit. Corolla bright red, tube slender, up to 5 mm. long, glabrous, strict to slightly arcuate, densely villous at throat, internally shortly pilose, lobes about 4 mm. long, 3 mm. wide, broadly obovate, shallowly emarginate. Stigma clavate, style persistent in fruit. Capsule to about .5 mm. long, compressed-globose, rarely equaling but not exceeding calyx. Hasitat. Dry, sandy savannas up to 450 meters. DISTRIBUTION. Known only from Vaupés, Colombia. REPRESENTATIVE SPECIMENS. Colombia. VaupEs: Rio Kuduyari, Cerro Yapo- oda, ca. 450 m., Schultes & Cabrera 14211 (a, holotype), Schultes, Baker, & Cabrera 18543 (a); Rio Karurd, Mesa de Yambi, Savanna Goo-ran-hoo-da, Schultes & Cabrera 19132 (cH), 19141 (GH); Rio Kubiyu, Mitt and vicinity, Zarucchi 2143 (GH, K). HERBARIUM, RoyYAL paces GARDENS Kew, RICHMOND, SURRE ENGLAND 1973] MARTINEZ-HERNANDEZ ET AL., TILIACEAE 299 POLLEN OF TROPICAL TREES. I. TILIACEAE E. Martinez-HERNANDEZ, P. FERNANDEZ, AND S. LozANo SINCE THE FAMILY Tiliaceae is predominantly tropical, and some of its members are important elements of tropical ecosystems, we have chosen to study and describe the pollen of a few of them. We worked with five species of five different genera, representatives of the flora of tropical Mexico, each one described and illustrated in the Manual para la Identifica- cién de Campo de los Principales Arboles Tropicales de México, by T. D. Pennington and J. Sarukhan. These species are: Apeiba tibourbou Aublet, Belotia mexicana (DC.) K. Schum., Carpodiptera ameliae Lundell, Helio- carpus donnell-smithii Rose, and Luehea speciosa Willd. Our descriptions are accompanied by photographs taken with a Zeiss light microscope and with an AMR scanning electron microscope (SEM). The specimens studied under the light microscope were obtained from the Herbario Nacional de México (mExu), while those studied under the SEM were from the collections of the Harvard University Herbaria (a' or cH2). The specimens from MEXU were acetolyzed (Erdtman, 1952) and mounted with glycerine jelly; those from A or GH were coated with carbon and gold-palladium. Pollen of most of the genera studied has been described before by various authors: Apeiba by Sharma (1969), Belotia by Erdtman (1952), Carpodip- tera by Erdtman (1952) and by Chaudhuri (1965), and Luehea by Mohl (1835). The pollen specimens are deposited in the pollen reference collection of the Paleobotanical Laboratory in the Harvard University Herbaria build- ing and at the Palynological Section of the Museo de Micropaleontologia at the Instituto de Geologia Universidad Nacional Autonoma de México. DESCRIPTIONS Apeiba tibourbou Aublet. Ficures 1-5, 25, 32. APERTURE. Brevitricolporate, colpus transversalis. Grains slightly aspidor- ate. When acetolyzed the colpus membrane formed by the nexine is mod- ified. After acetolysis the remaining structure, the sexine, is psilate under the light microscope. With the SEM this structure is scabrate in unace- tolyzed grains. Polar index, 0.7; colpus transversalis, 7 um. 3 wm.; colpi, 9.38 wm.; pore, 9 wm. & 4 pm. ORNAMENTATION. Tectum perforate to heteroreticulate; lumina 0.3—0.8 ym., distributed at random; muri simplibaculate. Stratification clear, colum- ?Arnold Arboretum of Harvard University, Cambridge, Mass. ? Gray Herbarium of Harvard University, Cambridge, Mass. Ficures 1-12. 1-5, Apeiba oe (from Pennington & sees 9065, Tuxtepec, Oaxaca, México), x 1000. 1, 2, polar view: 1, high focus; 2, median optical section. 3-5, equatorial eke 3, high foc see ee eee aper- ture) ; 4, optical section (note effect of aac on sexine of os); 5, broken grain showing well-developed colpus transversalis in com onan aperture. 6-10, Belotia mexicana (from Pennington & Sarukhdn 9278, Puebla, México), X 1000. 6, polar view, high focus. 7-10, ies view: 7, high focus (colpus eat covered by the sexine) ; 8, hig us (colpus transversalis com- plete); 9, high focus; 16 optical section. is 12. Carpodiptera ameliae (from Matz ida $.n. eae México), Xx 1000, polar view: 11, high focus; 12, median optical section. FicurES 13-24. 13-15, Carpodiptera ameliae (from Matuda s.n., Tenosique, México), 1000, equatorial view: 13, high focus; 14, optical section; 15, high focus. 16-20, Heliocarpus donnell-smithiu (from Matuda Si. Acocoya agua Chiapas, México), X 1000. 16, 17, polar view: 16, high focus; 17, median optical section. 18-20, equatorial Siew 18, 19, high focus: 20, rica optical section. axaca, México), X 1000: 21, polar view, optical section on one side and high focus on others; 22, 23, high focus; 24, optical section (note clear stratification of sexine on one side of grain). 302 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 ellae easily seen. Exine, 1.5 »m. (ectosexine, 0.5 pm., endosexine, 0.75 pm., mexine, 0.70: qiii) 2 Belotia mexicana (DC.) kK. Schum. Ficures 6-10, 26, 32. APERTURE. Tricolporate, colpus transversalis. The colpus membrane is psilate when the grain is acetolyzed, but verrucae are present close to the colpus transversalis. Without acetolysis the colpus membrane is scabrate (SEM sample). Polar index, 0.2; colpus transversalis, 24 pm. * 6.8 pm.; colpi, 25.04 wm.; pore, 3.82 pm. — u ORNAMENTATION. Semitectate, supraheterobrochate. Lumen size decreas- ing toward the apertures. There is a wide range between the sizes of the lumina (0.3-0.8 »m.). Stratification clear; muri simplibaculate. SEM photographs show the tectum to be perforate within the lumina. Colum- ellae visible. The nexine disappears at the level of the colpus transversalis. The ornamented sexine remains when the grains are acetolyzed. ee 1.6 pm. (ectosexine, 0.4 pm.; endosexine, 0.8 jm.; nexine, 0.4 pm. Carpodiptera ameliae Lundell. Ficures 11-15, 27, 33. APERTURE. Tri-brevicolporate and tetra-brevitricolporate (less frequently). Aspidorate. Pore oval, meridianally elongated. Pore membrane scabrate when acetolyzed (light microscope sample). Before acetolysis the grains have a notch located in the ora region of the aperture. This notch is lost upon acetolysis, and instead it is possible to see an exoaperture. Unace- tolyzed grains are endoaperturate. The membrane of the colpi is psilate (acetolyzed grains). There is a homogeneous layer constituting the endo- sexine at the pore level. Polar index, 0.7; colpi, 4.25 ~m.; pore, 1.89 um. ORNAMENTATION. Infratectate, perforate, microreticulate when acetolyzed (light microscope sample). F ossulate when unacetolyzed (SEM sample) (Walker & Doyle, 1975), microrugulate. Muri simplibaculate. Exine 1.6 »m. The ratio between the ectosexine, the endosexine, and the nexine is ie2e., — Heliocarpus donnell-smithii Rose. Ficures 16-20, 28, 29, 34. APERTURE. Tricolporate, endoaperturate. Colpus peu psilate (aceto- lyzed grains). The grains have a notch placed at the ora of the colpus when they are unacetolyzed. The pore is ae a meridianally elongated. Polar index, 0.2; colpi, 29.24 »m.: pore, 4.32 um. & 6 pm. ORNAMENTATION. Tectate, supraheteroreticulate. Two classes of lumina are found. The first kind decreases in size toward the aperture; the second is smaller and is distributed uniformly across the surface among those of the first type. Muri simplibaculate. The tectum is perforate inside the lumina (SEM sample). Columellae are easily seen. At the level of the pores, the ectosexine is thin, while the endosexine is thick. Exine, 1.6 pm. (ectosexine, 0.3 »m.; endosexine, 0.8 »m.; nexine, 0.5 pm.). 1978] MARTINEZ-HERNANDEZ ET AL., TILIACEAE 303 Ficures 25-28. Scanning eee micrographs of pollen grains: 25, Apeiba tibourbou (from Dressler & Jones 211, Mss ae Tuz, Snes a 6, Belotia mexicana (from Pennington & Sarukhdn 0278); arpodipte eliae (from Matuda 3626, Tabasco, México); 28, Bae donna smithii Gon Martinez-Calde- rén 452, Tuxtepec, Oaxaca, México); all X 2000). 304 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Ficures 29, 30. Scanning electron micrographs of pollen grains: 29, Helio- carpus donnell-smithii (from Martinez- Ree Nia 30, Luehea speciosa (from Purpus 1917, Veracruz, México); both x 2 1978] MARTIiNEZ-HERNANDEZ ET AL., TILIACEAE 305 Figures 31, 32. Scanning electron micrographs of pollen grains: 31, Apezba tibourbou (from Dressler & Jones 211), X 10,000; 32, Carpodiptera ameliae (from Matuda 3626), X 5000 306 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Ficures 33-35. Scanning electron micrographs of pollen grains: 33, Belotia mexicana (from Pennington & Sarukhdn 9278), X 10,000; 34, Heliocar pus donnell-smithii (from Martinez-Calderén 452), X 10,000; 35, Luehea speciosa (from Purpus 1917), < 50,000. 1978] MARTINEZ-HERNANDEZ ET AL., TILIACEAE 307 Luehea speciosa Willd. Ficures 21-24, 30, 35. Aperture. Tricolporate. Colpus membrane scabrate (SEM sample). Colpus transversalis 14-16 ym. long and 4-5 pm. wide. Polar index, 0.4; colpi, 23.92 pm.; pore, 5.93 »m. The unacetolyzed grains have a notch at the ora of the colpus that is lost during acetolysis. ORNAMENTATION. Tectate, supraheteroreticulate. Lumina of the supra- reticulum measure from 0.8 to 1.4 »m. Rugulate (SEM sample). Muri simplibaculate; columellae easily seen. Exine, 1.6 pm. (ectosexine, 0.4 pm.; endosexine, 0.8 »m.; nexine, 0.4 pm.). DISCUSSION The species of Tiliaceae studied are, in general, tricolporate. Using this feature we can determine two groups among the species observed: (8) brevitricolporate (Apeiba tibourbou and Carpodiptera ameliae), and (2) tricolporate (Belotia mexicana, Heliocarpus donnell-smithii, and Luehea speciosa). The pollen grains of C. ameliae can also be tetracolporate. In all except Carpodiptera ameliae and Heliocarpus donnell-smithi, the aperture is a colpus transversalis (Faegri & Iversen, 1964). They have in common a colpus membrane constituted by the endexine. Carpodiptera ameliae has a pore membrane in addition to the other structures. Helio- carpus donnell-smithii and Carpodiptera ameliae are endoaperturate. The pollen grains of these species were studied after two different treat- ments, with and without acetolysis. This produced remarkable differences between the apertures and ornamentation of grains of the same species (see text). The apparent changes were actually the result of loss of structures that had obscured other characters. The polar index, besides other features, seems to be important as a distinguishing character. We found a correlation between the polar index and the form of the grains. Depending on the structure, there are tectate grains (A peiba tibourbou) or semitectate grains (Belotia mexicana); some may have supraorna- mentation (e.g., Belotia mexicana). e width of the exine is very uniform, but differences occur in the ectosexine-endosexine-nexine ratio. Sizes of the different structures, as well as shape, association, polarity, and symmetry, are shown in TABLE 1 SUMMARY Although the Tiliaceae has representatives in the temperate zone that have been well studied, the representatives of the family in the tropics have been ignored from the palynological standpoint. They are very important as components of tropical ecosystems. For this reason we have studied their pollen morphology, and that of some of the more important elements of tropical tree flora of Mexico. TABLE 1. Pollen characteristics of five species of Tiliaceae. A peiba Belotia Carpodiptera Heliocarpus ueh tibourbou mexicana ameliae donnell-smit hii speciosa Size mean range mean range me range an range me range (um.) (um.) (um.) — (um.) (um.) (um.) (um.) (um.) (um.) (um.) POLAR AXIS 24.8 24 -26 34.64 32.8-38.4 20.8 16 -22 32.96 28 -36.8 30.5 24 -40 EQUATORIAL AXIS 33.43 28.8 —36.4 29,82 26.4-32.4 25.02 22.8 -28 23.72 20 -30.4 22.82 19.8-25.02 LuMINA 0.5 0.3 - 0.8 -~ 0.3— 0.8 0.4 - 1.5 1.2- 0.5 0.4 0.6 (large lumina) 0.8 0.6- (small lumina) Co.pr 9.38 8 -12 25.04 20 -28.8 4.25 4. —6 29.24 25.2-34.4 23.92 16.8-28.8 PORES 1.89 1.2 -24 3.82 2 - 5.6 1.89 1.2 - 2.4 4.32 3.2- 5.2 0.76 0.4— 2.8 9.5 TRANSVERSALIS x 4.0 9 -10 24 x 6.8 - - - - - 15 x 4.5 14 -16 long long ~ 5 wide EXINE: EcTOSEXINE 0.5 0.4 - 0.4 - 0.3 - 0.4 - ENDOSEXINE 0.75 - 0.8 - 0.8 ~ 0.8 - 0.8 = NExI O25 - 0.4 0.4 0.5 - 0.4 - Shape oblate prolate oblate prolate prolate Association monad monad monad monad monad Polarity isopolar isopolar isopol isopolar isopolar Symmetry ial radial radial or bilateral radial radial 80¢ WOLAYOLUY GIONNV AHL AO IWNwNAOL 6S “10A] 1978] MARTINEZ-HERNANDEZ ET AL., TILIACEAE 309 Five species in five different genera are included in this paper: A peiba tibourbou, Belotia mexicana, Carpodiptera ameliae, Heliocarpus donnell- smithii, and Luehea speciosa. All of them are tricolporate but some are brevitricolporate, indicating variability in the polar index. The ornamenta- tion and structure of pollen in the species studied appear to be character- istic of the genera. ACKNOWLEDGMENTS We would like to thank the several members of the staff at Harvard University who made the realization of this paper possible. Special thanks are due to Dr. Alice Tryon for her valuable assistance. Also we would like to thank Professor R. M. Tryon, Professor E. S. Barghoorn, B. Tiffney, E. Seling, of the Museum of Comparative Zoology, and the Arnold Arbore- tum-Gray Herbarium library personnel. REFERENCES Cuavupuurt, S. K. 1965. Pollen morphological studies of the order Malvales. I. Bot. Soc. Bengal 19: 147-158. ErptMaN, G. 1952. Pollen morphology and plant taxonomy: angiosperms. Pp. 337— 394. Hafner Publ. Co., New York. Farcri, K., & J. — 1964. Textbook of pollen analysis. P. 237. Hafner Publ. Co., New Yo Mon, H. 1835. Sur i structure et les formes des graines de pollen. Ann. Sci. Nat. Bot. ser. 2. 3: 148-180; 220-236; 304-346. PENNINGTON, T. D., & J. SARUKHAN, 1968. Manual para la identificacién de campo de los principales ae eg de México. v + 413 pp. Inst. Nac. Invest. Forest., S.A.G., M SHarMa, B. D. 1969. Pollen morpoloey of Tiliaceae in relation to plant taxon- omy. - — as ee WALKER, J., & J. Doy Le. 1975. an bases of angiosperm phylogeny: palynology. Ann. Missouri Bot, Gard. 62(3): 664-723. E. M.-H. anp S. L. P. F. DEPARTAMENTO DE GEOLOGIA DEPARTAMENTO DE BOTANICA INSTITUTO DE GEOLOGIA INSTITUTO DE BIOLOGIA UNIVERSIDAD NACIONAL UNIVERSIDAD NACIONAL AUTONOMA DE MEXICO AUTONOMA DE MEXICO APARTADO PosTAL 70-233 APpARTADO PosTAL 70-233 Mexico 20, D.F., Mexico Mexico 20, D. F. Mexico o 09 7 | y t Jaa ‘ ia : i? : _ : : = :. on 37 | : os i a — - a . - _ pi = r - _ : = ry - - aS ea) Wt a = _ eee tO — cea Fete . 1, ys eg — Preliminary Announcement Thirteenth International Botanical Congress Sydney, Australia. 21-28th August, 1981 The Programme will consist of 12 sections — molecular, metabolic, cellu- lar and structural, developmental, environmental, community, genetic, systematic and evolutionary, fungal, aquatic, historical, and applied botany. There will be plenary sessions, symposia, and sessions for sub- mitted contributions (papers and posters). Chairman of the Programme Committee:—Dr. L. T. Evans. Field Trips will include visits to arid and semi-arid regions, eucalypt forest, rain forest, heath, coastal vegetation (e.g. Great Barrier Reef, mangroves) etc., and specialist trips. Chairman of the Field Trips Com- mittee: —Prof. L. D. Pryor. First Circular, containing details, will be mailed in August, 1979. Send your name and full address, preferably on a postcard, to ensure your in- clusion on the mailing list. Enquiries should be sent to the Executive Secretary, Dr. W. J. Cram. Congress address — 13th I.B.C., University of Sydney, N.S.W. 2006, Australia. Sponsored by the Australian Academy of Science. Journal of the Arnold Arboretum July, 1978 CONTENTS OF VOLUME 59, NUMBER 3 The Genera of Crassulaceae in the Southeastern United States. STEPHEN -A. SPONGBERG . --2>... . 5°. Se. ~~ 198 The Genus Phyllocladus (Phyllocladaceae). HsuaN KENG . .. oe tee oe eee Wood Anatomy and Phylogeny of Paeonia Section Moutan. JOSEPH M. KEEFE and MAYNAROSE. MOSEERY «IR, . .¢ 10271). ame. | TA A Spectacular Buchnera (Scrophu- lariaceae) from Colombia. DERE COKE. See Sa a. a abe 2, Se Re eee ee Pollen of Tropical Trees. I. Tiliaceae. E. MARTINEZ-HERNANDEZ, PA RERNANDEZ, -and-S: LOZANO... ., .# . op cert... 1909 SSS Volume 59, Number 2, including pages 105-196, was issued May 3, 1978. JOURNAL or tH: ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 59 NUMBER 4 US ISSN 0004-2625 Journal of the Arnold Arboretum Published quarterly in January, April, July, and October by the Arnold Arboretum, Harvard University. 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 1-45, reprinted, and some back numbers of volumes 46-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, U.S.A. EDITORIAL COMMITTEE B. G. Schubert, Chairman S. A. Spongberg P. F. Stevens C. E. Wood, Jr. ASSISTANT EDITOR E. B. Schmidt Printed at the Harvard University Printing Office, Boston, Massachusetts COVER: The stylized representation of fruits 8 various Leguminosae, form- ing the design of our new co for Vol | on terial from plants growing the Arnold Arboretum of Harvard University. The des was planned and executed by Karen S. V re, wh also drawn several ed xec of the preceding covers as well as various tes used in the Journal of the Arnold Arboretum and on the offprint cov The three Sa ara of the Leguminosae are represented by the fruits shown in the desi The genera are, top row (left to right), Albizia (Mimosoideae), Gleditsia RC aceHeinigldeae), Caragana (Faboideae), Gymnocladus (Caesal- pinioideae), and lower row, right, Colutea (Faboideae). The family Leguminosae, =e second most important family of flowering plants, is to be the subject o oe onan Legume Conference at the Royal Botanic Gardens, Kew, ae Surr y ecient in July, 1978. This cov- er is our salute to the complete success ss that Conference. Second-class postage paid at Boston, Massachusetts JOURNAL OF THE ARNOLD ARBORETUM VoL. 59 OCTOBER 1978 NUMBER 4 GENERIC LIMITS IN THE TRIBE CLADOTHAMNEAE (ERICACEAE), AND ITS POSITION IN THE RHODODENDROIDEAE Bruce A. BouM, Scott W. Brim, RICHARD J. HEBDA, AND P. F. STEVENS THE CLADOTHAMNEAE COMPRISE a group of four species, all ornamental shrubs, in the Ericaceae subfamily Rhododendroideae. Although they are all fairly well known, being in cultivation, the four species have been given s. There is no consistency in the usage of these names, four generic name as the arrangements listed below indicate. SPECIES EPITHET | racemosa paniculata _ bracteata pyroliflorus REFERENCE Elliottia Elliottia Elliottia Cladothamnus | Hooker, 1876; is Co peland, 1943 f Elliottia Elliottia Botryostege Cladothamnus | Stapf, 1934 a Elhiottia Tripetaleta Tripetaleia Cladothamnus | Drude, 1897; s ox, 1948 iS) Elhiottia Tripetaleia Tripetaleia Tripetaleia Stevens, a4 fa (combinatio rs C. Gas in 0 Tripetaleia, not published) In the following discussion, Copeland’s nomenclature will be followed. This paper is an attempt to resolve the generic limits in the group by using evidence from anatomy, chemistry, morphology, and palynology,’ and combining this information with what is already known about floral anatomy. The evidence presented here suggests that there is only a single * The sections of morphology and anatomy are by Brim and Stevens; the pollen was examined by Hebda in con the former The introduction and general discussion, written by eae are acceptable to all. © President and Fellows of Harvard College, 1978. ai2 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 genus, Flliottia, and it also provides a firmer basis for understanding the phylogeny of the Ericaceae subfamily Rhododendroideae as a whole. The Cladothamneae were first recognized as a separate tribe within the Rhododendroideae by Copeland (1943) on the basis of evidence from morphology, floral anatomy, and embryology. Four species were included: Cladothamnus pyroliflorus Bongard, from northwestern North America (Map 2); lliottia racemosa Muhl. ex Elliott, from southeastern North America (Map 3); and F. paniculata (Siebold & Zucc.) Bentham & Hook- er f. and £. bracteata (Maxim.) Bentham & Hooker f., both from Japan (Map 1). A fifth species, Tripetaleia yvakusimensis Nakai, has been de- scribed from Japan, but it was considered by Kydégoku (1962) to be a synonym of 7. paniculata Siebold & Zucc. (£. paniculata). On the basis of numerical analysis, Watson, Williams, and Lance (1967) suggested that the species that they studied (C. pvroliflorus, FE. paniculata, and E. brac- teata) would be better included in the Rhodoreae D. Don. Stevens (1971) followed Copeland in his circumscription of the Cladothamneae. Whether placed in a separate tribe or not, the four species included in the Clado- thamneae are clearly more closely related to one another than they are to any other members of the Rhododendroideae. Ericaceae subfamily Rhododendroideae tribe Cladothamneae Copeland. Shrubs or small trees. Indumentum of unicellular hairs only, plant some- times almost glabrous. Leaves deciduous, convolute in bud, entire. Inflo- rescence terminal, + monotelic (cymose), eperulate; + foliaceous bracts and “‘bracteoles” usually present, the former sometimes occurring alone, without the latter; pedicel continuous with the calyx. Flowers 3- to 5- (or 6-)merous; calyx well developed, the sepals connate or free; corolla polypetalous; stamens 5 to 10, the filaments flattened, the anthers with resorption tissue, dehiscing by long slits, the pollen nie in very young bud, the tetrads with viscin threads; ovary glabrous, 3- to 5-locular, pla- centation axile, the placentae apical; style impressed into the apex of the ovary, with an expanded collar around the + prominent stigma. Capsule septicidal; seeds variable in size and shape, the cells of the testa thin walled, polygonal, with large pits (3—)6—-15 pm. across, the embryo with a poorly developed chalazal haustorium. Typr: Cladothamnus Bongard. GROWTH, MORPHOLOGY, AND ANATOMY VEGETATIVE ORGANS GROWTH AND MoRPHOLOGY. Like most of the Ericaceae, the four mem- bers of the Cladothamneae prefer acid habitats. Flliottia bracteata, E. paniculata, and Cladothamnus pyroliflorus are shrubs less than three meters tall, and, although £. racemosa is apparently at least initially a tree, Knight (1938) noted that most plants in the wild are shrubby because they sprout vigorously from the base after damage. All taxa have the same basic ar- chitecture: the stems are basically orthotropic, each shoot is terminated 1978 | BOHM ET AL., CLADOTHAMNEAE OLS by an inflorescence, and growth is continued by the development of axil- lary buds. This growth pattern basically conforms to Leeuwenberg’s model (nomenclature that of Hallé & Oldeman, 1970) and is similar to that of many Ericaceae, especially the great majority of the subfamily Rhododendroideae (Temple, 1975). Shoots of the four species can readily be distinguished from one another, even in winter, although the differences between them are mostly trivial. The bark of Cladothamnus pyroliflorus begins to exfoliate on twigs that are one year old; this character is not nearly so marked in other species, although in all the phellogen is initiated inside the ring of pericyclic fibers of the stem. Elliottia paniculata has sharply angled twigs, the angles being decurrent from the bases of the leaves; the twigs of &. racemosa are al- most terete (the inflorescence less so); the other two species are inter- mediate. Not surprisingly, . paniculata has the most prominent foliar buttresses. The vegetative buds of Elliottia racemosa and Cladothamnus pyroli- florus consist simply of paired, mucronate bud scales. However, in £. bracteata, and even more so in E. paniculata, the outer pair of scales does not enclose the mucronate inner scales, so the bud scales are imbricate. All species have inflorescences terminating the current year’s growth, and the terminal buds of vegetative shoots are functional; there is, how- ever, rather surprising variation in the growth patterns of the four species. Often in Elliottia paniculata, less frequently in E. bracteata, very rarely and perhaps only exceptionally in E. racemosa, and never, apparently, in Cladothamnus pyroliflorus, growth of the vegetative shoot ceases after a few weeks, and a more or less perulate terminal resting bud, with promi- nent and well developed scales, is formed (e.g., see Nakai, 1922, under Tripetaleia paniculata). Growth soon starts again, and small leaves and an inflorescence are produced (FicureE 1, D). n the two other species of the Cladothamneae, Elliottia racemosa and Cladothamnus pyroliflorus, there is usually no obvious cessation of plant growth, and the foliage leaves and bracts are not clearly distinguishable. However, in E. racemosa elongation of the main axis slows down soon after the axis is initiated and then speeds up again. In EF. bracteata the growth of the inflorescence is frequently similar to that of C. pyroliflorus (but see above), although specimens placed in Elliottia (Tripetaleia bracteata var. longiracemosa Nakai) suggest that the transition between vegetative growth and the growth of the inflorescence may sometimes be abrupt (Ficurr 1, B). In these specimens the leafy vegetative shoot has large buds in the axils of the leaves and is rather clearly distinct from the in- florescence, which is long, the lower 4.5—7(-10) cm. of the axis bearing small leaves in the axils of which are neither flowers nor obvious buds. In E. paniculata and E. bracteata both continuous and discontinuous growth may occur on a single plant during the same season. The growth pattern of Elliottia racemosa and E. bracteata has not been observed before in the Ericaceae and may be unique in the family (a comparative study of growth patterns and inflorescence type and develop- 314 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 ~NS Oo > Zz \ O £ PP bg poe Ass a N LLLLLL Ww ‘ he a " NY ( Sete Vp pps A Se 2 LP PPL i oe a N ao Ficure 1. Inflorescences in the Cladothamneae (diagrammatic): A, Clado- thamnus pyroliflorus; B, Elliottia bracteata var. longiracemosa; C, E. bracteata normal); D, E. paniculata; E, E. bracteata, unusual branched inflorescence; F, G, E. racemosa (circles indicate flowers; ellipses, buds; triangles, axis not obviously terminated by a flower). \ (Abe, SEE a 1978 | BOHM ET AL., CLADOTHAMNEAE a1 ment (see below) throughout the family is much needed). Although Lems’s (1962) interesting study of shoot development in the Andromedeae did not include a similar pattern, the closest approach to it seems to be in that tribe. In Pieris japonica (Thunb.) D. Don a leafy shoot, a resting bud, and then another leafy shoot terminated by an inflorescence are frequently all produced in the same season; however, the flowers of the inflorescence do not open until the following year (W. Judd, pers. comm. ). All four species of the Cladothamneae are deciduous. (Stevens, 1971, reported that Elliottia racemosa was evergreen, but the living plant ob- served was growing in a greenhouse.) The leaves of E. racemosa have long petioles, while those of Cladothamnus pyroliflorus are sessile; the other two species have leaves with at most short petioles. Distribution of indumentum and the shape and texture of the leaf blade allow each spe- cies to be recognized by its leaf alone (see Key). Anatomy. The wood anatomy of Elliottia racemosa, E. paniculata, and Cladothamnus pyroliflorus was examined by Cox (1948). In this respect, E. racemosa was found to be more distinct from the other two species than the two species were from each other. In £. paniculata there was a slight preponderance of scalariform perforation plates with more than nine bars, and there were round pits in the wood parenchyma and on the side walls of the vessels. In the other two species there was a strong preponderance of perforation plates with fewer than eight bars, the pits were round or elongate-elliptic, and there were also differences in medul- lary ray type. Elliottia racemosa and C. pyroliflorus were similar, differing mainly in the percentage of vessels with porous perforation plates. How- ever, it is difficult to evaluate the significance of these differences since the variation between the species that Cox considered in the natural genus Befaria Mutis ex L. was as great as that between the three species of the Cladothamneae that he examined. Perhaps, more importantly, many of the characters that Cox used both vary infraspecifically and also depend on the age and the general health of the part of the plant from which the material examined was taken (see Carlquist, 1961, for references). The pith of all four species is heterogeneous, with clusters of thick- walled cells interspersed with clusters of larger, thin-walled cells. (Watson, 1965: Stevens, 1971; Gris, 1870, 1872; and Copeland, 1943, give descrip- tions of the pith that are in basic agreement.) However, the thick-walled cells toward the periphery of the pith tend to be especially well developed. This is particularly marked in EF. paniculata, in which the pith approaches the Calluna type, with small, thick-walled cells in the periphery and larger, thin-walled cells restricted to the center. Copeland (1943) found that in E. racemosa most of the pith consisted of thick-walled cells, the thinner- walled cells being relatively inconspicuous. Stevens (1971) noted that the vascular bundle in the petiole of Elliottia racemosa is closed, with the phloem surrounding a ring of xylem, while that of the other species was arcuate. However, in the smaller leaves of E. racemosa, especially those produced at the beginning of growth and 316 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 those produced later just below the inflorescence, the vascular bundle is open, although it is deeply arcuate. In some intermediate-sized leaves it may be open at the base of the petiole, but then it becomes closed higher up. In the other species it is always shallowly arcuate. There are a number of differences in lamina anatomy between the four species. Most of these are relatively trivial, and include differences in the thickness of the cuticle, sinuosity of the anticlinal epidermal cell walls (an environmentally labile character), and length of the palisade mesophyll cells. All species have a single layer of palisade cells, although in Clado- thamnus pyroliflorus and Elliottia racemosa there is often a more or less well-developed second layer. The variation in the shape of the midrib bundle mirrors that of the petiole: in E. racemosa it is closed to deeply arcuate, while in the other species it is shallowly arcuate. The stomata are always on the abaxial surface of the lamina. In C. pyroliflorus and E. bracteata the stomata are anomocytic, while in the other two species the majority of stomata are paracytic (Stevens, 1971). The cuticle in E. brac- teata is almost reticulately ridged, while in the other species it is striate (most prominently so in &, paniculata) to smooth. INFLORESCENCE STRUCTURE. The inflorescences of the four members of the Cladotham- neae are at first sight rather different one from another, and, with the partial exception of Cladothamnus pyroliflorus, they have often been called paniculate or racemose (i.e., polytelic: Hooker, 1876; Copeland, 1943; Wood, 1961; Stevens, 1971; Yamazaki, 1975). However, the in- florescences in all four species are basically similar and are more or less clearly monotelic (cymose). In the following discussion, unqualified use of the term “bract” will refer to a leafy structure on the main axis that usually subtends a first-order paracladium (a flower or a more complex, branched system that repeats the structure of the main axis of the flower- ing system); the bract may be empty. The term “second-order bract”’ will refer to a comparable structure on a first-order paracladium; and “third-order bract” to a leafy structure borne on a second-order para- cladium. (This nomenclature follows that of Troll; see Weberling, 1965, for a convenient summary.) Cladothamnus pyroliflorus (Ficurr 1, A) has the smallest inflorescence in the Cladothamneae. The 1- to 2- (to 4-) flowered inflorescence, a botrys, has bracts that are almost indistinguishable from the leaves, and any paracladia, which are always single flowers, have a pair of opposite to alternate second-order bracts that are usually called bracteoles because of their position. Between the paracladia and the terminal flower there are usually two or more bracts that do not subtend flowers; small buds are sometimes visible in the axils of these bracts. The terminal flower opens first. Elliottia bracteata (Ficure 1, C) has an inflorescence with essentially the same structure as that of Cladothamnus pyroliflorus, although the leafy 1978 | BOHM ET AL., CLADOTHAMNEAE Sy bracts are smaller and are more clearly distinguishable from the foliage leaves, and the inflorescence is larger, having 3 to 15 flowers. The terminal flower often opens first, although in larger “Taflereeenees the lowest flow- ers open first, then the terminal flower, and, finally, the remaining ones. Between the two uppermost paracladia there is sometimes a bract lacking an axillant paracladium; this bract is out of phyllotactic sequence, as are any bracts above the uppermost paracladium. In other inflorescences the first bract that does not subtend a paracladium is inserted at, or almost at, the level of the uppermost paracladium, and in yet other inflorescences the first bract is inserted well above the uppermost paracladium (FIGURE 1, C, arrows). In these latter two cases regular phyllotactic sequence is maintained. In the first case the phyllotactic sequence would be regular if the first bract were inserted above the uppermost paracladium, so it seems that this first bract can be variable in position. In some inflorescences there are no empty bracts between the terminal flower and the uppermost paracladium. Elliottia paniculata has a profuse, branched thyrse that may have over 40 flowers (FicurE 1, D). The main axis of the inflorescence and the first-order paracladia terminate in flowers that open before the flowers immediately below them. The lower first-order paracladia of the inflores- cence bear single flowers along their length and alternate second-order bracts between the terminal flower and the lateral flowers, while the upper first-order paracladia have more or less paired lateral flowers and usually lack second-order bracts between the terminal flower and the uppermost second-order paracladia. Where bracts are present between a terminal flower and a paracladium, they usually continue the phyllotactic spiral of the axis below them, although there is a tendency for the sterile first- order bracts on the main axis to be out of sequence (see above). All sec- ond-order paracladia have two subopposite third-order bracts and a ter- minal flower. The main axis or the first-order paracladium sometimes lacks terminal flowers, and instead there are a number of small bracts of the appropriate order as well as abortive paracladia (axillant flowers). In a few cases none of the paracladia has more than a single flower. When both such variations occur on a single inflorescence (terminal flower lack- ing only on the main axis), a ‘‘racemose”’ inflorescence results (see below, FE. racemosa). The lowermost paracladium of one robust inflorescence of E. bracteata examined had a pair of second-order bracts, possibly bud scales, at its base (FIGURE 1, E, arrow). Vigorous, branched inflorescences on a plant of FE. bracteata growing at the Arnold Arboretum were similar in structure to those of E. paniculata (FicurE 1, E). e main axis of Elliottia racemosa is usually not terminated by a flow- er, but there are a number of bracts subtending abortive paracladia toward the end. The only inflorescences we saw that were terminated by flowers were on a late-flowering plant at the Arnold Arboretum, but they ap- parently occur sporadically in nature (Wood, 1961, p. 22 (footnote) ). The first-order paracladia have either a single terminal flower and two subopposite second-order bracts (‘“bracteoles’”: FicuRE 1, F), or two op- 318 JOURNAL OF THE ARNOLD ARBORETUM [VvOL. 59 posite second-order paracladia as well, these in turn having a terminal flower and two subopposite third-order bracts (FrGuRE 1, G). More com- plex inflorescences have not been seen, and there are no bracts between terminal flower and paracladium on the lateral branches. The flowers to- ward the middle of the inflorescence open first (FicurE 2, A), and the terminal flowers on the three-flowered branches in an inflorescence like that shown in FicurE 1, G open before the lateral flowers (this latter has been observed before (e.g.. Wood, 1961)). ‘hus, the inflorescences of all four species are fundamentally monotelic (cymose), although they differ greatly in size and in overall appearance. The small, simple type of inflorescence of Cladothamnus pyroliflorus can be related to the more elaborate inflorescences of Elliottia racemosa and E. paniculata by the lengthening of the main axis, with the terminal flow- E 2, Flower types in the Cladothamneae. A, B, Elliottia racemosa: A, ireceie (note flowers in the middle of the inflorescence opening first) less spreading petals and stamens i, paniculata (note spreading petals and filaments that are initially erec eg 1978 | BOHM ET AL., CLADOTHAMNEAE 319 er consequently developing later and eventually not developing at all, and the development of both more and also higher orders of paracladia. The inflorescence shows a comparable trend in the reduction of foliation in both the bracts and the sepals. There is also a trend in the regularity of insertion of bracts on paracladia: in C. pyroliflorus the bracts tend to be alternate, but in E. racemosa they are rather strictly opposite. FLOWER Morphology. There are several differences between the members of the Cladothamneae in the numbers and arrangement of the parts of the flower (see TABLE 3). Some other floral differences are discussed below. Calyx. The two species with free sepals, Elliottia bracteata and Clado- thamnus pyroliflorus, also have a very much longer calyx than do the other two species (/. bracteata, ca. 5 mm. long; C. pyroliflorus, ca. 15 mm. long: E. paniculata and EF. racemosa, less than 2 mm. long). Corolla color. In Elliottia racemosa the petals are white, while in E. paniculata and EF. bracteata, although basically white or pale greenish, they are more or less pink or streaked with pink toward the apex. The petals of Cladothamnus pyroliflorus are usually completely copper col- ored, although they are apparently sometimes greenish white (an albino form?). Curvature of the stigma and protection of the nectary. Stevens (1971) noted that the petals of Elliottia bracteata (Ficure 2, C) and Clado- thamnus pyroliflorus were broadly spreading, as were the stamens. In- sects visiting flowers of these species landed on the petals and flattened filaments and received pollen on their backs from that attached to the strongly curved stigma as they took nectar from the exposed nectary at the base of the flower. It was suggested that in E. racemosa and E:. panicu- lata the pollination mechanism was similar, although it was noticed that petals of the former were not so reflexed; only herbarium material of the latter was seen. Observations of vigorously growing plants of Elliottia racemosa and E. paniculata at the Arnold Arboretum show that the style is not strongly re- curved and the nectary is not exposed. However, the method of nectar protection differs in the two species. In . racemosa (F1GuRE 2, B; Santa- mour, 1967, fig. 54) the petals are not particularly strongly reflexed and form a cup that encloses the nectary. The edges of the flattened filaments have spaces between them, and the filaments themselves are little bowed at their bases. In £. paniculata (FicurrE 2, D) the petals are reflexed and do not form a cup, but the filaments are bowed at their bases and in con- tact laterally, thus forming a tube around the base of the ovary and the ary. Some details of the pollination of Elliottia racemosa were given by Wood (1961), and, although field studies of the pollination of the Cladotham- 320 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 neae (and most other Ericaceae) are badly needed, the following addi- tional information based on the plants growing at the Arnold Arboretum seems of interest. No visitors to /. paniculata or E. bracteata were seen, but numerous bees (Bombus and Apis spp.) were working E. racemosa (see also Santamour, 1967, fig. 52). When the pollen is shed in the young bud, it collects in the space between the petals and the stigma (the style is, as yet, short). The style elongates considerably in later bud and, just before the petals open, is compressed into an S-shape. The petals, with their somewhat hooded tips in which the stigma rests, tend to spring open violently when late buds are touched with twigs. Small masses of pollen held together by viscin threads were seen flying through the air when this happened; observations are needed to see if this is an important part of the pollination process. Visiting insects landed on and clung to the petals. ‘he stigma, to which pollen is attached, touched the ventral and lateral surfaces of these insects as they probed for nectar between the filaments; the clefts on the stigmatic lobes are initially closed. Seeds, The seeds of £lliottia racemosa are shieldlike; the embryo and endosperm occupy the central part and are surrounded by a wing composed of testa only (see Wood, 1961, for illustrations). The cells of the testa that forms the wing and that which covers the central part are similar, both having large pitted areas, and are similar to the testa cells of the other members of the Cladothamneae. Pollen Methods. For examination under the light microscope, pollen grains were first acetolyzed (Faegri & Iversen, 1964) and were then passed through an alcohol dehydration series before being mounted in silicone oil on glass slides. Measurements for each species were made on specimens from at least two different sources (see TABLE 1). Values were obtained for tetrad diameter, length of the furrow as exposed on the surface of the grain, width of the furrow margin, and exine thickness at the poles of grains, Both acetolyzed and unacetolyzed pollen samples were prepared for SEM observation. Pollen was air-dried and mounted directly on copper conducting tape that was glued to an aluminum stub. Gold was evaporated onto the material in a Mikros model evaporator, and the stubs were then observed in a Cambridge stereoscan. — Results. The pollen of all four species is borne in tetrads (FrcurE 3, A). The range of tetrad diameter is 40-70 ym. (TaBLE 1), with the diameters of these four species forming a series of overlapping ranges. Elliottia paniculata has the smallest tetrads, with means between 40 and 45 pm., whereas the other three species have tetrads with means between 40 and 55 ym. However, one of the two samples of E. racemosa (Harper s.n.) has larger tetrads that vary considerably in size, the mean diameter being 65.6 wm. Tetrad size is especially variable in this species; Santamour 1978 | BOHM ET AL., CLADOTHAMNEAE 321 Table 1. Variation in pollen measurements in the Cladothamneae. Elliottia bracteata* Elliottia paniculata+ Elliottia racemosa” Cladothamnus pyroliflorust+ 1 2 3 4 5 6 fi Average tetrad diameter** 51,5 47.8 Bowe 42.3 44.7 65.6 523 48.0 53.9 Range of tetrad diameter 45-59 44-52 51-57 36-50 42-48 59-74 7-60 42-54 49-57 Standard deviation 2.8 - - 1.8 - 4.1 es a5 = Sample size 100 10 25 100 10 27 02 100 10 Average grain diameter 35.1 S552 35.0 30 7.9 37.2 33 36.0 Range of grain diameter 29-40 31-39 33-37 6 27-32 44-54 33-4 29-38 34-40 Standard deviation 2.1 - = 3 = 3.4 - 2.0 - Sample size 102 10 10 100 10 8 100 102 10 Furrow length - 9.9 = = 8.3 - 9.5 10:2 - Margin width - 3.6 = = 3.9 - 4.3 44 oe Average exine thickness 1.2-1.6 1.2-1.6 1.2-2.0 1.€ 1.6-2.0 1.6-2 2.0-2 2.0-2.4 1.2-2.0 Sample size 10 10 20 10 10 10 10 10 10 *Elliottia bracteata - 1. Bohm, 1975, Garden; 2, Faurie, 1905, Japan; 3, Murot, 1955, Japan. +Elliottia paniculata - 4, Bohm, 1975, Garden; 5, Muroi, 1955, Japan. °Eliiottia racemosa - 6, Harper, 1901, Georgia, U.S.A.; 7, Lee, 1960, Georgia, U.S.A. ++Cladothamnus pyroliflorus - 8, Bohm, 1975, B.C., Canada; 9, Peterson, 1959, B.C., Canada. **A1] measurements in micrometers (am.). (1967) found less than 6 percent fertility in the material that he ex- amined, and the collapse of one or two members of the tetrads was par- ticularly noticeable in this study. Furrow length (8.3—-10.2 »m.) and margin width 3.6—4.4 ym.) are very similar for all four species. A pore aia Sauna cae to form a fusi- form, transverse furrow (FrcuRE 3, B) in all members of the group. The is form mostly by a thickening of the endexine (Ficure 3, D). This thickened area is surrounded by a peripheral endexinal crack ae 3, C, surface view; Ficure 3, B, transverse view) which is usually connected to a net- work of endocracks es cracks: FicurE 3, C) distributed over the visible surface of the gra The exine is ee with the ektexine usually from 1.5 to 3 times thicker than the endexine. It consists of a thin foot layer and narrow columellae. Under high magnification both the light micrographs (Fic- URE 3, E-H) and SEM micrographs show that the clavobaculate sur- face ornamentation is produced by the tops of the columellae, some of which are fused (Ficure 4, B, D, F, H). Elliottia racemosa has slightly finer surface ornamentation (Ficure 4, D) than do the other three species. oe2 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 59 FicurE 3. Details of pollen of the Cladothamneae. A, Cladothamnus pyro- liflorus, tetrad, 760; B, Elliottia racemosa, elongate pore crossing furrow, 2000; C, Elliottia bracteata, surface of two adjacent furrows showing dif- ferentiated margins, with peripheral endexinal crack (double arrow) surround- 1978 | BOHM ET AL., CLADOTHAMNEAE O28 Exine thicknesses of Elliottia racemosa, E, paniculata, and Cladotham- nus pyroliflorus are similar (means, 1.7—2.2 pm.), but the exine of the three collections of E. bracteata is somewhat thinner (means, 1.4, 1.4, and 1.5 »m.). Comparison of the SEM micrographs of the unacetolyzed tetrads (Ficure 4, A, C, E, G) clearly shows that the tetrads of FE. bracteata collapse much more severely than do those of the other species, probably because of this thinner exine. Collapse can also be seen under the light microscope and has already occurred in herbarium specimens. In some cases acetolysis somehow “reinflated” the grains, destroying the differ- ence between E. bracteata and the other species. The viscin strands observed earlier in the Cladothamneae (e.g., Cope- land, 1943) can be clearly seen in the SEM photographs (FIGURE 4, A, Col eG)s Conclusion. The pollen morphology is very similar in all four species. Although Elliottia bracteata has a somewhat thinner exine than do the other species, all of the other essential features are sufficiently similar to indicate very close relationships among the four species. Hence, pollen studies provide no evidence for placing the species in more than one group (see also General Discussion). Cytology. Santamour (1967) reported a number of » = 11 for Elliot- tia bracteata. Embryology. The chalazal haustoria in Elliottia racemosa and Clado- thamnus pyroliflorus were essentially moribund according to Copeland (1942), and their absence was used to characterize the tribe. However, Yamazaki (1975) found small, but definite, chalazal haustoria in £. panicu- lata and E. bracteata. Dissection of mature seeds of E. bracteata showed a large micropylar haustorium (pers. obs., P. F. S.), although there was no trace of a chalazal haustorium. All species were reported to have a large micropylar haustorium. Any difference between the species in the size of the micropylar haustorium is probably one of degree only. The micropylar haustorium is much larger than the chalazal haustorium in all three species, whereas in most of the family both haustoria are prominent. Anatomy. Copeland (1943) examined the floral anatomy of Elliottia racemosa, E. paniculata, and Cladothamnus pyroliflorus. In the first two species he found that there were separate traces to the individual sepals and petals; each of these traces later divided into three. The stamens were likewise supplied by single traces, and the carpels were vascularized by carpel dorsal, carpel ventral, and septal bundles. Branches of the carpel ing margins, and reticulum of endexinal cracks (single arrow), x 1248; D, El- liottia paniculata, transverse section of furrow, showing endexinal thickening (tip of double arrow) and two endexinal cracks (indentations at tips of single arrows), < 2000, E-H, surface ornamentation, heads of clavobacula showin as black spots, X 2000: E, C. pyroliflorus,; F, E. racemosa; G, E. paniculata; H, E. bracteata. (For collections examined, see Figure 4.) 324 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 E 4. Comparison of the pollen of the Cladothamneae. A, C, E, G, scan- ning ee micrographs of pollen tetrads showing viscin strands and furrows 1978 | BOHM ET AL., CLADOTHAMNEAE 6) dorsal and the septal bundles vascularized the nectary. Copeland found Cladothamnus pyroliflorus to be basically similar, although its receptacle was depressed and the axis hence more compressed. Up to three vascular bundles supplied the sepals, with the lateral sepal bundles arising from other sepal bundles or from petal bundles, or even originating unattached in the receptacle. The bundles to the antepetalous stamens arose from the petaline bundles, while the vascular bundles supplying the antesepalous stamens sometimes originated from the main sepal bundle. The nectary was not vascularized. Yamazaki (1975) recently examined the floral anatomy of Flliottia bracteata and E. paniculata (as Tripetaleia). His account disagrees in detail with that of Copeland (1943) and contains a novel suggestion con- cerning the origin of the petals in these two species. In &/. bracteata Ya- mazaki found six petal traces arising from four gaps: two traces from ad- jacent gaps supplied the adaxial (upper) petal, while a pair of traces arose from each of the two lower gaps and entered the two abaxial (lower) petals. In E. paniculata he found that the adaxial petal was supplied by two traces from adjacent gaps, the abaxial petals being vascularized by a single trace that later forked into three. Other details were as given by Copeland (1943). In abortive flowers the upper petals frequently were supplied by a single bundle that forked into two; this bundle arose from a single gap. The vascular supply to the petals of E. bracteata suggested to Yamazaki that each petal might have been derived from two by fusion; the vascular supply of E. paniculata seemed to him to be more specialized than that of E. bracteata. FLAVONOIDS A few years ago a study of the flavonoids of the Cladothamneae was undertaken with a view to applying the data to the systematic problems within the tribe. The flavonoids of Cladothamnus pyroliflorus were de- termined at that time (Bohm & Saleh, 1972). Since then, material from the other three taxa has become available. Cladothamnus has been reex- amined by improved methods, and the structures of the majority of flavo- noids present in the other species have been determined. Several reports of chemical constituents of members of the tribe have been published by other workers. Kondo ef al. (1963) reported terpene derivatives, paraffins, fatty alcohols, ou quercetin from Tripetaleia panicu- lata ‘lliottia paniculata). In a series of papers by Yasue and co- workers (1971a, 1971b, 1973, 1974; Scie et al., 1974), a number of long-chain fatty alcohols and ketones, sterol derivatives, simple phenolic CE), scale bar = 10um.; B,D, Fy e, scanning electron micrographs of surface ornamentation of pollen grains, scale bar = 1 wm. ladothamnus pyroli- florus (from Bohm, 1975, Canada) ; cD Elliottia racemosa (Lee, 1960, Georgia, U.S.A.); E, F, else paniculata (Muroi, apa liottia bracteata (Bohm, 1975, garden). Note in G m ae collapse of tetrad in E. bracteata; in B, D, F, and H, fused Peaobacnl (FC). 326 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 compounds, and flavonoids were described from various tissues of the same species. The only flavonoid glycosides reported were quercetin 3-O-arabino- side and quercetin 3-O-galactoside. Harborne and Williams (1973) re- ported kaempferol, quercetin, and myricetin in hydrolyzed extracts of Cladothamnus pyroliflorus and Elliottia racemosa in their extensive survey of the Ericaceae. Material. Cladothamnus pyroliflorus was collected in Mt. Seymour Pro- vincial Park, British Columbia, Canada, at the site from which material was obtained for the earlier work on this species (Bohm & Saleh, 1972). Elliottia racemosa, from Bulloch Co., Georgia, was supplied by Dr. John Bozeman, of Georgia Southern College, who has retained a voucher speci- men. Eilliottia paniculata was obtained from two sources: 1) Dr. G. Mu- rata, Kyoto University, who collected the material from its native habitat; and 2) my garden (B. A. B.). Elliotia bracteata was also available from my garden. Voucher specimens have been prepared; living material of the latter two taxa is being maintaine Extraction and isolation. Dried material was extracted by repeated soaking with 80 percent methanol, usually with three or four changes of solvent; the extracts were pooled. Fresh material was blended with 80 percent methanol at room temperature in a Waring blender. (The ma- terial extracted included leaves, stems, and flowers.) The ground mass was filtered and reextracted with 80 percent methanol by letting it stand, with occasional stirring, for two hours. Three or four changes of solvent gave complete extraction as judged by color loss of the leaf material; the extracts were again pooled and evaporated im vacuo at 30°. The resulting aqueous suspension was filtered through a bed of Celite Analytical Filter Aid. Any chlorophyll remaining at this stage was removed by shaking a few times with cyclohexane. The aqueous solution was saturated with NaCl and extracted ten times with ethyl acetate in a separatory funnel. In order to be sure that the more polar constituents had been extracted, two extractions with water-saturated n-butanol were performed. The ethyl acetate and n-butanol extracts were compared by two dimensional thin layer chromatography (TLC); in all cases in this study the two ex- tracts were the same and were combined. Isolation and purification of the flavonoid glycosides were accomplished by the methods described in detail by Wilkins and Bohm (1976). Struc- tural analyses were based upon standard ultraviolet spectrophotometric methods (Mabry ef a/., 1970). The proton magnetic resonance (PMR) spectrum of quercetin 3-O-arabinoside (determined as the pertrimethylsilyl ether) was obtained using a Varian HA-100 instrument. Tetramethyl- silane was used as the internal standard. Results. Twenty-two flavonol glycosides have been isolated and identi- fied from the four species examined (TaBLE 2). The majority of these compounds are derivatives of kaempferol and quercetin, with quercetin being the major aglycone present in all taxa. Three myricetin glycosides 1978 | BOHM ET AL., CLADOTHAMNEAE p27 TABLE 2. Flavonoid distribution in the Cladothamneae. Cladothamnus Elliottia Elliottia Elliottia pyroliflorus racemosa paniculata bracteata K-3-Q-arabinoside + + slit 3-O-glucoside + + + + 3-O-galactoside + + + 3-O-gln 4 + tr + 3-O-rha + 3-O-rutinoside = +. a 3-O-gal-7-O-rha + 3-O-ara-7-O-rha 4 Q-3-O-ara +a +a a 3-O-gle “pe a a + 3-O-gal + a. as 3-O-gln + - tr z 3-O-rha +. 3-O-rutinoside + -E +. 3-O-rha-ara - 0-O-gal-7-O-rha a 3-O-ara-7-Q-rha + 7-O-rha = IR-3-O-rutinoside + M-3-O-ara + + = 3-O-gle + Bie A 3-O-gal + Explanation of abbreviations: K = kaempferol; Q = quercetin; IR = iso- rhamnetin; M = myricetin; ara = arabinose; glc = glucose; gal = galactose; gln = glucuronic acid; rha = rhamnose; a = isomeric pair of glycosides; + = present; tr = trace only; no entry = absence of compound. were found. A small amount of isorhamnetin-3-O-rutinoside was found in Elliottia paniculata, Within the tribe, Cladothamnus pyroliflorus is the only species that makes 3, 7-di-O-glycosides (Bohm & Saleh, 1972). Both kaempferol and quercetin occur as the 3-O-galactosyl-7-O-rhamnosides and 3-O-arabinosyl-7-O-rhamnosides. Reexamination of C. pyroliflorus showed a small amount of quercetin-7-O-rhamnoside not noticed before (Bohm & Saleh, 1972). Three species accumulate pairs of isomeric quercetin-3-O-monosides. Cladothamnus pyroliflorus had a pair of quercetin-3-O-arabinosides and a pair of 3-O-glucosides. Elliottia racemosa and E. paniculata had pairs of 3-O-arabinosides, while . paniculata had a pair of 3-O-glucosides as well. The predominant member of the arabinoside pair from E. paniculata was shown to be the 6-D-pyranoside by PMR. The predominant member of the glucoside pairs was also the 8-D-pyranoside; this was determined on the basis of chromatography against standards known to possess this con- figuration. In all cases the pyranosides chromatographed more slowly than did the isomeric compounds. 328 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Elliottia paniculata possessed a quercetin-3-O-glycoside derivative that ran in aqueous systems on polyamide plates approximately halfway be- tween the 3-O-glucoside a the 3-O-rutinoside. Total acid hydrolysis yielded quercetin, arabinose, and rhamnose, while partial acid hydrolysis yielded arabinose and a compound dic dngaihebie from quercetin-3-O- rhamnoside, so the diglycoside is quercetin-3-O-arabinosylrhamnoside. The position of attachment of the “outer” sugar was not determined. Careful search failed to detect this compound in the other taxa. Discussion. Full use of chemicals for systematic purposes requires knowledge of their biosynthesis and interrelationships. In the case of fla- vonoids, much is known about the pathways by which the various types are made. Details of some of the steps have been obtained by studying the inheritance of flavonoid pigments and by investigating substrate spe- cificity of enzymes of the pathways. By looking at systems in other plant groups, it is possible to discuss the flavonoid glycoside differences seen in the Cladothamneae in terms of single biochemical steps. Although neither enzymatic nor genetic studies of the Cladothamneae flavonoids have been reported, enough of these other systems is known to suggest that the for- mation of flavonoids follows much the same route in all plant groups. Some pertinent genetic and enzymatic results are reviewed below The formation of the aglycone types found in the Cladothamneae can be summarized by the following abbreviated sequence (flavonol A-rings are omitted for clarity). ; vricetin querce 3° -O-methylat ic Tt OCH, It is known that different genes control 3’ and 5’ hydroxylation of antho- cyanins as well as flavones and flavonols in several plants (Harborne, 1967). Harborne (1967) also listed several plants in which O-methylation was shown to be controlled by a single gene. Ebel ef al. (1972) presented evidence showing that specific flavonol O-methyltransferases exist in cul- tured cells of parsley. Flavonoid monoglycosides are formed by transfer of the sugar moiety of a nucleoside diphosphate sugar to one of the hydroxyl groups of a flavo- noid. Enzymatic and genetic evidence suggest that each such event is under strict control. Using preparations from Phaseolus vulgaris, Marsh 1978 | BOHM ET AL., CLADOTHAMNEAE 329 (1960) demonstrated the enzymic formation of quercetin 3-O-glucuronide. Similarly, Barber and Chang (1968) showed that an enzyme from Leu- caena glauca catalyzed the biosynthesis of quercetin 3-O-rhamnoside. Harborne (1967) has reviewed the genetics of glycosylation reactions in a variety of plants. Typical of the results are the observations of Lawrence and Sturgess (1957) and Harborne (1963) on the inheritance of flower color in Streptocarpus hybrida. Separate genes control 3-glycosylation and 5-glycosylation of anthocyanins. Separate genes also control xylosyla- tion to give anthocyanin 3-sambubiosides and rhamnosylation (of 3, 5- diglucosides) to give the 3-rutinoside-5-glucosides. Additional support for the stepwise formation of diglycosides comes from the work of Barber (1962), who showed that the formation of quercetin 3-O-rutinoside occurs in Phaseolus aureus by successive enzyme reactions. In recent work with cell cultures of parsley, Sutter and coworkers (Sutter e¢ al., 1972; Sutter & Grisebach, 1973) have shown the existence of two separable glycosylat- ing enzymes with different positional and substrate specificities. Uridine diphosphate glucose (UDP-glucose) : flavone/flavonol 7-O- glucosyltrans- ferase was specific for the 7-position and would not accept 3-O- glycosides as substrates. The second system, UDP-glucose : flavonol 3-O-glucosyl- transferase, was specific for the 3-position but would accept flavonol 7-O- glycosides as substrates. The flavonoid glycosides of the Cladothamneae are based upon the com- mon aglycones kaempferol, quercetin, isorhamnetin, and myricetin (TABLE 2). Elliottia bracteata has kaempferol and quercetin glycosides only; Cladothamnus pyroliflorus and E. racemosa make kaempferol, quercetin, and myricetin glycosides, while E. paniculata goes one step further and ac- cumulates the three hydroxy flavonols as well as the O-methylated flavonol Re ues Thus, on the basis of aglycone elaboration, E. bracteata E. paniculata can be distinguished from one another and from the ie two taxa, which are indistinguishable. In the case of monoglycoside composition, Elliottia bracteata can be dis- tinguished from the other three taxa by its simplicity, only glucosides and glucuronides being found. The other taxa make both of these, as well as arabinosides and galactosides. Elliottia racemosa can be distinguished from the other taxa by its ability to make quercetin 3-O-rhamnoside. Clado- thamnus pyroliflorus has kaempferol 3-O-rhamnoside and quercetin 7-O- rhamnoside. Elliottia paniculata has the capacity to make quercetin 3-O- rhamnoside, but this compound is converted to the 3-O-arabinosylrhamno- side: no monorhamnoside was detected. Elliottia bracteata does not make monorhamnosides, but does have the capacity to make rhamnosylglucosides (rutinosides). nteresting differences exist in the diglycoside fraction of the four spe- cies. The three species of Elliottia contained 3-O-rutinosides, but Clado- thamnus pyroliflorus did not. A striking feature of C. pyroliflorus is its capacity to make flavonol-3-O-galactoside-7-O-rhamnosides and 3-O-ara- binoside-7-O-rhamnosides. It seems reasonable to postulate that this rep- resents a significant difference in rhamnosyltransferase enzymes compared 330 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 to the other three members of the tribe. In Elliottia rhamnosylation occurs at position 3 of the flavonol or at the 6” position of the flavonol-3-O-glu- cosides. Cladothamnus pyroliflorus appears to be able to carry out the first of these reactions (it had kaempferol-3-O-rhamnoside), but it does not convert the product further. It can also make the 3-O-glucoside, but further rhamnosylation to the rutinoside does not occur. Finally, C. pyroliflorus was found to have quercetin-7-O-rhamnoside, unique in the tribe. In line with the specificity findings of Sutter and coworkers (1972, 1973), this last compound could serve as substrate for enzymes transferring arabinose and galactose to the 3-position of the 7-O-rhamnosides. One further distinction can be based upon diglycoside data. Elliottia paniculata and EF. bracteata have in common the capacity to synthesize rutinosides, but they can be distinguished from one another on the basis of the formation of quercetin-3-O-arabinosyl-rhamnoside in the former but not in the latter. GENERAL DISCUSSION Despite our rather considerable knowledge of the variation in the Clado- thamneae, the main points of which are summarized in TaBLe 3, the de- cision as to how many genera should be recognized is still not an easy one. When the Cladothamneae are considered in terms of unique characters or character states that one or more species may have, it is clear that both Elliottia racemosa and Cladothamnus pyroliflorus are individually more different from the other three species than are E. bracteata and E. panicu- lata, although the species pairs E. bracteata-C. pyroliflorus and E. panicu- lata-E, racemosa each have a number of characters in common. Elliottia racemosa differs from all other Cladothamneae in its more or less closed petiole and lamina bundle (and perhaps also in characters of wood anatomy: cf. Cox, 1948), in its inflorescence, which usually lacks a terminal flower, and in its thick-walled capsules with flattened, winged seeds that are shed only tardily. Cladothamnus pyroliflorus differs from the other three species in its copper-colored petals, sessile gynoecium, and unvascularized nectary, as well as in several aspects of the flavonoid chem- istry (see also TABLE 2). Of the latter, the occurrence of 7-O-rhamnosyla- tion and the associated “diglycoside replacement” is perhaps most impor- tant. The species pairs Elliottia bracteata-Cladothamnus pyroliflorus and F.. paniculata-E. racemosa have in common stomatal configuration, calyx size and type, and flower type. There are numerous other differences in morphology, anatomy, and flavonoid chemistry (aglycone structure and richness and complexity of glycosylation types) that allow the ready recognition of each species and that help to give them their very distinctive individual appearances. These would support the recognition of genera if suggested by other evidence, but of themselves are insufficient to characterize genera. Of greater signifi- cance are the numerous characters that all four species have in common, at least some of which might be expected to vary if one were dealing with 1978] BOHM ET AL., CLADOTHAMNEAE Table 3. Variation of some characters among species of the Cladothamneae. Elliottia Elliottia racemosa paniculata Two periods of growth per season = +(-) Buds with imbricate bud scales = + Petiole bundle closed or deeply arcuate + - Stomata without accessory cells = a Leaves long-petiolate + - Lamina rounded at the apex, but mucronulate - - Inflorescences branched +- + Inflorescences usually with a terminal flower -(+) +(-) Sepals free, foliaceous - - Calyx lobe/sepal number 4-5 (3-)5(-6) Petal number 4-5 3-4 Corolla basically white + + Stamen number 8-10 (5-)6-8 Gynoecium stipitate + +4 Nectary exposed, style curved more than 90° = Number of ovules per loculus 8-10 intermediate Capsule thick-walled + - Seeds strongly flattened and winged + - Nectary vascularized + + Myricetin glycosides Bz + Isorhamnetin 3-O-rutinoside = + Monoglycosides - arabinosides and galactosides only - - Diglycosides - 3-0-rutinosides + + 3, 7-di-0-glycosides = = Quercetin 7-0-rhammoside - - Explanation of symbols: + = present; ++ = present, very strongly developed; - = absent; Elliotria bracteata numerous numerous () = uncommon state. Cladothamnus pyroliflorus different genera; these characters are summarized in the tribal description in the introduction. The significance of the characters that typify the three groupings above is much vitiated by the basic similarity of all four species. The similarity in the pollen of the four species is perhaps to be pected in ex the Ericaceae; see the limited amount of variation found by Oldfield (1959) in his study of some European taxa. Elliottia racemosa, suggested by Stevens (1971) as being distinct, can be recognized only by its long-petiolate leaves, woody capsules, and flat- aoe JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 tened seeds. The inflorescence is similar in structure to that of E. panicu- fata, and the difference in leaf anatomy can be used only to support a genus otherwise well founded: the variation in leaf anatomy found in the Cladothamneae as a whole is common at the infrageneric level elsewhere in the Ericaceae (e.g., Befaria, Rhododendron, and Kalmia). Cladothamnus pyroliflorus is perhaps the most distinctive of the four species in its general appearance. Copeland (1943), following Hooker (1876), distinguished Cladothamnus in a key by its sessile leaves that were rounded at the apex, and by its flowers, which were terminal or from the axils of the highest leaves and had 5 or 6 petals (lliottia had petiolate leaves that were acute at the apex, paniculate inflorescences, and flowers often with 3 or 4 petals). It will be clear from Taser 3 and the preceding discussion that the two latter characters are of little significance; FE. brac- teata has very shortly petiolate leaves that are often broadly rounded and mucronate at the apex and, in these characters, are like leaves of C. pyro- liflorus. Thus C. pyroliflorus, rather like E. racemosa, is distinctive main- ly because it stands at one end of a number of trends in the Cladotham- neae: leaf type, inflorescence elaboration, calyx size, numbers of flower parts. In flower color, sessile gynoecium, and non-vascularized nectary (the last two might be indirectly connected), C. pyroliflorus stands some- what apart from the other Cladothamneae, and it is the only species with a flavonoid chemistry distinctive enough to support generic rank, but the morphological differences are too slight to maintain it as a monotypic ge- nus. The characters distinguishing Cladothamnus pyroliflorus and those dis- tinguishing Elliottia racemosa from the rest of the Cladothamneae are ra- ther different. The inflorescence type of E. racemosa is clearly derived, and the winged seed also seems to be a specialization of the ellipsoid seed found elsewhere in the Cladothamneae. Flowers in the Ericaceae are ba- sically 5-merous, and the 4-merous flowers of E. racemosa with their stipi- tate gynoecia are probably derived, although less so than those of E. brac- teata and FE. paniculata. The closed petiole bundle of E. racemosa is pos- sibly less derived than the open condition found in the rest of the genus. In C. pyroliflorus, on the other hand, it is possible that only the vegetative characters are derived, and even in these characters C. pyroliflorus ap- proaches £. bracteata. Unfortunately, not enough is known of the chem- ical characters to enable discussion concerning whether or not they are de- rived. The suggestion that there are two groups in the Cladothamneae, Elliot- tia paniculata-E, racemosa and E. bracteata-Cladothamnus pyroliflorus, is a novel one, although Stapf (1934) approached this arrangement when he suggested that his genus Botryostege (E. bracteata) was closer to C. pyroli- florus than to the other two species, which he placed in Elliottia. However, it should be remembered that the nectary is protected in different ways in E. paniculata (by the bowed filaments) and E&. racemosa (primarily by the petals). It is also possible that the presence of a large, foliaceous calyx may be functionally correlated with a flat-lying corolla, the large calyx 1978 | BOHM ET AL., CLADOTHAMNEAE 333 providing additional support to the platform on which the insects land. In view of the more obvious differences between E. bracteata and C. pyroli- florus, the recognition of two genera, each with two species, has little to recommend it. The most satisfactory (or perhaps least unsatisfactory) taxonomic treatment of the Cladothamneae is as a single genus, Elliottia, containing four very distinct species that are related in a somewhat reticulate fash- ion. The names of the species are listed below, and their main synonymy given. Hooker (1876) placed the three species that he included in Elliottia (he did not include Cladothamnus pyroliflorus ) in separate, unnamed sec- tions. Nakai (1922) placed the Japanese species in separate sections with- 45 130 135 140 145 Distribution of Elliottia in Japan: dots, E. bracteata (from Horikawa, oe Fae E. paniculata (from H. Hara, 19 58 Ne 334 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 in Tvripetaleia: sect. EUTRIPETALEIA (properly TRIPETALEIA) included TL. paniculata; sect. ScHIzoCALYX Nakai included 7. bracteata. The pres- ent authors feel that, rather than formally placing each species in a sep- arate section or subgenus, it seems sufficient to remember that all four species are very distinct. Elliottia as circumscribed here has a disjunct distribution, with Elliottia bracteata and E. paniculata growing in Japan (Map 1), E. pyroliflora found scattered along the northwest coast of North America from Alaska to Oregon (Map 2), and E. racemosa occurring in the southeastern United States (Georgia and, formerly, South Carolina: Map 3). This and similar disjunctions (about 50 genera are involved — see Wood, 1972, for refer- ences) probably result from the disruption of a widespread early Tertiary flora by climatic change and orogenic events (Graham, 1972). The four species of Elliottia grow in more or less acid and moist habi- tats, as is common in the Ericaceae. Elliottia bracteata grows in grass- lands immediately surrounding snowbeds (Ishizuka, 1974) and in mon- Maps 2, 3. Distribution of Elliottia in North America: 2 (left), Z. pyroliflora (partly from Hultén, 1968, and Szczawinski, 1962); 3 (right), E. racemosa. 1978 | BOHM ET AL., CLADOTHAMNEAE 335 tane broadleaf-evergreen forests, and E. paniculata grows in similar habi- tats, but at somewhat lower altitudes. Flliottia pyroliflora grows in moist forests, near streams and bogs, along shores, and near the tree line as thickets in openings in hemlock-spruce forests (Calder & Taylor, 1968). Elliottia racemosa seems to do best on moist, well-drained sand in mixed woods rather near streams, but it also grows in such habitats as oak-dom- inated forests, ridges, and Pinus australis savannas (Wood, 1961). Elliottia Muhlenberg ex Elliott, Sketch Bot. South Carolina Georgia 1: 448. 1817 Cladothamnus Bongard, Mém. ee ae St.-Pétersb. Sci. Nat. VI. 2: 155. 1832. Tolmeia Hooker, Fl. Bor.-Am. 2: 834. Tripetaleia Siebold & Zuccarini, a. an Akad. Math. Phys. Kl. 3: 731. 1843. Botryostege Stapf, Kew Bull. 1934: 194. 1934. KEY TO THE SPECIES OF ELLIOTTIA ry Inflorescences usually with fewer than 4 flowers; sepals ' . mm. lon petals copper colored; ovary sessile. ................... elas cee eee nearly always with more than 3 flowers; sal ae most ca. —" m. long; petals white, tinged with red; ovary stipitat ee, often + rounded at the apex, mucronate; a of free sepals ca. 5 mm, long; nectary exposed in open flower. .......... 1. E. bracteata. bo Lamina cuneate to acute at the apex, ae or not mucronate; calyx connate, less than 2 mm. long; nectary concealed in open flow 3% es subterete; petiole 0.3-1.8 cm. long; capsule as sed flat- COS is wheat en eiw ee eee ea ae . racemosa 3s aie strongly angled; petiole less than 4 mm. lon pee iat woody; seeds not flattened: 2. Joe. pas i ee ees ks _E. paniculata. 1. Elliottia bracteata (Maximowicz) Hooker f. in Bentham & Hooker f. Gen. Pl. 2: 598. 1876; Tripetaleia bracteata Maximowicz, Bull. Acad. Sci. St.-Pétersb. III. 11: 432. 1876; Botryostege bracteata (Maximowicz) Stapf, Kew Bull. 1934: 194. 1934. SyNrTyPEs: Japan. Yeso [near Hakodate], variis locis, Maximowicz; Nambu, 1865, Tschonoski. la. Elliottia bracteata (Maximowicz) Hooker f. forma bracteata. 1b. Elliottia bracteata (Maximowicz) Hooker f. forma longiracemosa (Nakai) Brim & P. F. Stevens, comb. nov. Tripetaleia bracteata Maximowicz var. longeracemosa Nakai, Bot. Mag. Tokyo 44: 530. 1930; Botryostege bracteata (Maximowicz) Stapf forma longe- racemosa (Nakai) Hara, Bot. Mag. Tokyo 50: 562. 1936. Type: Japan. Yeso, in monte Apoi, prov. Hidaka, T. Nakai. 2. Elliottia paniculata (Siebold & Zuccarini) Hooker f. in Bentham & Hooker f. Gen. Pl. 2: 598. 1876; Tripetaleia paniculata Siebold & 336 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Zuccarini, Abh. Bayer. Akad. Math.-Phys. Kl. 3: 732. 1843. Type: Japan, Siebold. T. yakusimensis Nakai, Bot. Mag. Tokyo 40: 485. 1926. Type: Japan. Ya- kushima, in monte Yaedake, Auvmura. 3. Elliottia pyroliflora (Bongard) Brim & P. F. Stevens, comb. nov. Cladothamnus pyrolaeflorus Bongard, Mém. Acad. Sci. St.-Pétersb. Sci. Nat. VI. 2: 155. 1832; Letophyllum pyrolaeflorum (Bongard) Dippel, Handb. Laubholzk. 1: 436. 1889. Type: | Alaska.] New Archangel, Mertens. Pyrola fruticosa Esch. ex Ledebour, Fl. Ross. 2: 924. 1824. Nomen.” Tolmeia occidentalis Hooker, Fl. Bor.-Am. 2: 44. 1834. Type: northwest coast of America, Menzies. 4, Elliottia racemosa Muhlenberg ex Elliott, Sketch Bot. South Caro- lina Georgia 1: 448. 1817. SyNnrypes: Georgia. Waynesborough, Burke County, a@non.; Oconee, Jackson. The reassessment of the generic limits in the Cladothamneae raises some points of more general importance in considering the evolution of the Ericaceae as a whole. Befaria, with its terminal inflorescences and poly- petalous flowers with large and variable numbers of parts, has often ex- plicitly or implicitly been considered to be the most primitive member of the Rhododendroideae and/or the Ericaceae as a whole (Copeland, 1943, and references; N. Hara, 1958), and the Cladothamneae have been placed close to the Befarieae. Reevaluation of the available evidence shows that there is no necessity to consider Befaria and Flliottia to be more closely related to each other than to any other Rhododendroideae, and that it is very difficult to assess the position of either genus in any phyletic system of the family. Although Stebbins (1974, p. 267) cites the Ericaceae as a family in which all members have indeterminate (racemose, polytelic) inflorescences, this is clearly not so; however, monotelic inflorescences seem to be un- common outside the Cladothamneae. Befaria racemosa Vent. (Rhododen- droideae - Befarieae) very occasionally has three-flowered branches with a terminal flower, the inflorescence having a structure similar to that shown in Ficure 1, F (R. & M. Kral 6715). In Ledothamnus Meissner (Rhododendroideae - Phyllodoceae) there are reports of terminal inflo- rescences and sometimes single flowers (e.g., Copeland, 1943), but the flowers in the inflorescences examined were all axillary; careful observa- tion is necessary to confirm that the single flowers are truly terminal. In the Vaccinioideae - Andromedeae, Craibiodendron W. W. Sm. has flowers apparently terminating the main axes (although these are usually broken off in herbarium specimens) and certainly terminating the lateral axes of * Baillon (Adansonia 1: 197. 1860) described a plant from “California” that Doug- las called Pyrola fruticosa. Although Baillon clearly considered P. fruticosa to be a species of Cladothamnus, he did not make the combination of P. fruticosa under Cladothamnus. 1978 | BOHM ET AL., CLADOTHAMNEAE 337 the axillary inflorescences. Occasional terminal flowers occur in Agarista Don (W. Judd, pers. comm.). Terminal flowers are also reported from the Monotropoideae (Copeland, 1941) There seems to be a consensus that the cymose, monotelic inflorescence is a more generalized (primitive) type than the racemose, polytelic inflo- rescence. Although there is some discussion as to whether a single, terminal flower or a three-flowered dichasium is the least specialized inflorescence type of all (see Rickett, 1944; Stebbins, 1974), the few-flowered, highly foliated inflorescence of Elliottia pyroliflora (FiGuRE 1, A) is either the least specialized type (Rickett, 1944) or very close to it, and can be con- sidered the least specialized type in the Ericaceae as a whole. The inflorescences are also rather unspecialized in the rest of the genus, but there is apparently a transition to the polytelic inflorescence common in the rest of the family, with a concomitant reduction in foliation, increase in number of flowers, and increase in the regularity of insertion of the highest order bracts (the “bracteoles” of the rest of the family). It is rela- tively easy to derive at least the great majority of inflorescence types in the Ericaceae from those found in Elliottia. The inflorescences of Befaria are of a derived type, being usually more compact, less foliated, and ap- parently almost entirely polytelic; the single exception is noted above. Much of the other evidence as to the primitiveness of Befaria and its close relationship with the Cladothamneae is unsatisfactory. Copeland himself noted (1943) that the extremities of the carpels were more per- fectly fused in Befaria than in the Cladothamneae, although the vascular bundles supplying the various organs of the flower in B. racemosa were all separate (this was not so in the Cladothamneae). There seems to be no reason to consider the reduction of the chalazal haustoria of the embryo in the Cladothamneae to be derived rather than the converse. Copeland considered that the presence of resorption tissue in the anthers and of fila- ments (viscin threads) in the pollen was evidence of the relationship of Befaria and the Cladothamneae, since he thought that such a combination of characters did not occur elsewhere in the Rhododendroideae. However, such a combination also occurs in the Phyllodoceae (Stevens, 1971); the mechanism of anther dehiscence in the Epigeaeae and Diplarcheae is un- known. The presence of an endothecium in the anther of Befaria (but not in Elliottia) may be an unspecialized character. Although Befaria may have more than the 5 sepals, petals, and carpels, and the 10 stamens that are common in the Ericaceae, together with infraspecific variation in the numbers of these organs, this again cannot be considered a@ priori a primi- tive character; such a combination is a derived state in Rhododendron section VirEYA, and species of Lyonia that have more than the normal number of floral parts are also derived (W. Judd, pers. comm.). Elliottia pyroliflora perhaps has comparable variation in the number of floral or- gans, although in Elliottia as a whole there also seems to be a reduction trend toward trimery, culminating in Elliottia bracteata. There are good indications that polypetaly may be a derived character in some Ericaceae (Ledum, Vaccinium L., and the Pyroloideae: Copeland, 1947; Leins, 333 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 1964); the transition from polypetaly to sympetaly in the Ericaceae is clearly neither difficult nor indicative of major phylogenetic differences. Finally, Cox (1948) noted that the wood anatomy of Befaria was not notably primitive, and, although he thought that it was rather different from the other Rhododendroideae examined, the evidence given hardly supported such a view. There are several other differences between the Befarieae and Clado- thamneae (Stevens, 1971). Befaria, so far as has been examined, has per- haps primitive trilacunar nodes (apparently not constantly so in B. race- mosa) and tetracytic stomata that are oriented more or less at right angles to the long axis of the leaf; Elliottia has unilacunar nodes and anomocytic or paracytic stomata that show no particular orientation. The two differ in leaf vernation, Befaria having revolute leaves, Elliottia convolute. The entire absence of other than unicellular hairs in Flliottia is unusual for the Ericaceae and is perhaps a derived character; Befaria has multicellular hairs as well. The testa of Befaria, with its elongated, thin-walled cells and very fine pitting, is similar to that of most of the Rhodoreae and some of the Phyllodoceae; that of Elliottia, with its broadly pitted, polygonal cells, is unique in the Rhododendroideae. Recent evidence on the distribu- tion of phenolics is also interesting (Harborne & Williams, 1973). The two species of Cladothamneae that they studied were distinct from the other Rhododendroideae in lacking all six phenolic constituents charac- teristic of that subfamily; Befaria was readily characterized by the ab- sence of gossypetin and the predominance of 3, 5-O-methyl flavonols and dihydroflavonols. Harborne and Williams (1973) did not find kaemp- ferol in Befaria, although it occurs in all the Cladothamneae. It is clear that we do not have enough knowledge to talk very con- structively about relative advancement in the Rhododendroideae. How- ever, Befaria and Flliottia, both isolated genera, are clearly not close to one another since they have in common no characters that necessarily suggest immediate relationship. Befaria is in some ways (phenolics, inflo- rescence, indumentum, testa) closer to other Rhododendroideae than is Elliottia. Each genus appears to have some ‘“‘primitive”’ characters, but the final position of the two genera in any phyletic scheme must take into account both the findings on inflorescence types and those on polypetaly; indeed, all characters must be treated much more critically. ACKNOWLEDGMENTS The flavonoid and pollen studies were supported by grants from the National Research Council of Canada, for which we are sincerely grate- ful. Special thanks are due to Drs. J. Bozeman and G. Murata, for sup- plying plant material, and to the staff of the Royal Botanic Garden, Edin- burgh, for the opportunity to use their collections, both living and pre- served. Dr. W. Judd and Professor C. E. Wood, Jr., have made a number of useful suggestions during the course of the work, and Dr. A. E. Rowse made valuable comments on the pollen work. Robin Lefberg drew Figure 1 and the maps. 1978 | BOHM ET AL., CLADOTHAMNEAE 339 LITERATURE CITED BARBER, G. A. 1962. Enzymic glycosylation of quercetin to rutin. Biochemistry 1: 463-468. & M. T. Y. Chang. 1968. The enzymic glycosylation of quercetin to quercitrin. Phytochemistry 7: 35-39. Boum, B. A., & N. A. M. SALLEH. 1972. me sai of Cladothamnus pyro- laeflorus, Canad. Jour. Bot. 50: 2081- J. A, & R. L. Tayior. 1968. oo at the Queen Charlotte Islands. art 1. xiii + 659 pp. Queen’s Printer, Ottawa. CarLquist, S$. 1961. Comparative plant anatomy. 146 pp. Holt, Rinehart, & Slane H. F. 1941. Further studies on Monotropoideae. Madrofio 6: 97- “1943. A study, anatomic and taxonomic, of the genera of the Rhodo- dendroideae. Am. Midl. Nat. 30: 533-625 1947. Observations on the structure and classification of the Pyroleae. Madrofio Dy DS=LO2, Cox, H. T. 1948. Studies in the comparative anatomy of the Ericales 1. Erica- ceae subfamily Rhododendroideae. Am. Midl. Nat. 39: 220-245. Drupe, O. 1889. Ericaceae. Jv: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. 4(1): 1-48. EBEL, J., K. Haniprock, & H. GrisEeBACH. 1972. Purification and properties of o-dihydricphenol meta-O-methyltransferase from cell-suspension cultures of parsley and _ relation to flavonoid biosynthesis. Biochim. Biophys. Acta 268: 313-3 FarecRI, K., & J. Soe 1964. Textbook of pollen analysis. ed. 2. 237 pp. afner, New Yor Grauam, A. 1972. Outline of the origin and historical recognition of floristic affinities between Asia and eastern North America. 7m: A. GRAHAM, ed., Floristics and Cres as of Asia and eastern North America. xii 278 pp. Elsevier, New Gris, A. 1870. An sae Se de la moelle dans les Ericinées. Bull. Soc. Bot. France 17: 11 , bBe2: aia | un mémoire sur la moelle des plantes ligneuses. Ann. Sci. Nat . V. 14: 34-79. HALLE, F,, & R. re A. OLDEMAN. 1970. Essai sur l’architecture et a oe de croissance des arbres tropicaux. xii + 178 pp. Masson, — ie 1958. Distribution maps of flowering plants in japan ~ an 100 + 4 pp. Inoue Shoten, Tokyo. Hana, N. 1958. Structure of vegetative shoot apex and development of the leaf the Ericaceae and their allies. Jour. Fac. Sci. Univ. Tokyo 7: 367-450. ee J. B. 1962. Plant polyphenols. IX. The glycosidic pattern of antho- cyanin pigments. Phytochemistry 2: 85-97. Comparative biochemistry of the flavonoids. viii + 383 pp. Academic Press, New Y C. A. WILLIAMS. 1973. A chemotaxonomic survey of flavonoids and simple phenols in the Ericaceae. Bot. Jour. Linn. Soc. 66: 37-54. J. D. 1876. Ericaceae. J: G. BENTHAM & J. D. Hooxer, Gen. PI. 4. Horixawa, Y. 1972. Atlas of the Japanese flora. 12 + 500 + vii pp. Gakken, Tokyo. 340 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 Hvutren, E. 1968. Flora of Alaska and neighboring territories. xxii + 1008 pp. Stanford University Press, Palo Alto. IsnizuKa, K, 1974. Mountain vegetation. Jn: N. Numata, ed., The flora and vegetation of Japan. x + 294 pp. Elsevier, New York. Knicut, W. A. 1938. A rare American shrub. Arnold Arb. Bull. Pop. Inf. 6: 7- 13. Konpo, Y., T. Nacaya, T. TakeucuHI, & T. TAKEMOTO. 1963. Constituents of Ericaceae. VIII. Constituents of Tripetaleia paniculata. Yakugaku Zasshi 83: 674-676. Kyocoku, T. 1962. oS on Tripetaleia paniculata Siebold et Zuccarini. Ecol. Rev. 15: 235, 236. LAWRENCE, W. J. C., & V. C. Sturcess. 1957. Studies on Streptocarpus III. Genetics and chemistry of flower color in the garden forms, species and hybrids. Heredity 11: 303-336. Lerns, P. 1964. Entwicklungsgeschichtliche Studien an Ericales-Bliiten. Bot. Jahrb, 83: 57-88. Lems, K. 1962. Adaptive radiation in the Ericaceae. I. Shoot development in the apes Ecology 43: 524-528. Masry, T. J., . MARKHAM, & a B, THomas. 1970. The systematic identi- fication a aes xl ae 4 pp. Springer, New York Marsu, C. A. 1960. Reactions . uridine diphosphate sugars catalysed by en- zymes of French-bean leaves. Biochim. Biophys. Acta 44: 359-361 — T. 1922. Trees and shrubs indigenous in Japan proper. Vol. 1. 511 + 8 pp. Shoten, Tokyo. onoein F. 1959. The pollen morphology of some west European Ericales. llen Spores 1: 19-48. mee a H. W. 1944. The classification of inflorescences. Bot. Rev. 10: 187- SAKAKIBARA, J.. H. Ina, & M. Yasue. 1974. Studies on the constituents of Pripetalers paniedlate. IV. Yakugaku Zasshi 94: 1377-1383. SANTAMOUR, F. S, 1967. Cytology and sterility in Elliottia racemosa. Morris Arb. Bull. 18: 60-63. STAPF, O. 1934. Botryostege, a new genus of Ericaceae. Bull, Misc. Inf. Kew 1934: 191-195. STEBBINS, G. L. 1974. Flowering plants: evolution above the species level. xvill + 399 pp. Belknap, Cambridge (Massachusetts). STEVENS, P. F. 1971. A oo of the Ericaceae: subfamilies and tribes. Jour. Linn. Soc. Bot. 64: Sutter, A., & H. GRISEBACH. ee UDP-glucose: flavonol 3-O-glucosyltrans- ferase fom cell suspension cultures of parsley. Biochim. Biophys. Acta 309: 289-295, , R. OrtMANN, & H. GRISEBACH. 1972. Purification and properties of an enzyme from cell- -suspension cultures of parsley catalyzing os transfer of UDP D-glucose to flavonoids. Biochim. Biophys. Acta 258: 71-87. SZCZAWINSKI, A. F, 1962. The heather family (Ericaceae) of a ae 205 pp. British rea Provincial Museum, Dept. Recreation and Con- servation, Handboo 19, TEMPLE, A. 1975. eee étude architecturale de quelques espéces. 95 pp. Thesis, Acad. de Montpelier, Univ. des Sciences et Techniques du Langue- doc. 1978] Watson, L. 1965. WILLIAMS, & G comparative study of some novel numerical techniques. Bot. 59: 491-501. WEBERLING, F. 1965. 215-221. WILKINS, C. K., & B. A. Boum. t YAMAZAKI, T. YASUE, M., var. Duersiola: 1961. udy of disjunctions. Typology of inflorescences. BOHM ET AL., CLADOTHAMNEAE The taxonomic significance of certain anatomical variations among Ericaceae. Jour. Linn. Soc. Bot. 59: 111-12 W. T _N 1966. Angiosperm taxonomy: a Jour. Linn. Soc. 1976. The flavonoids of Heuchera nucrantha Canad. Jour. Bot. 54: 2133-2140. The genera of Ericaceae in the southeastern United jour Arnold Arb. 42: 10-80. 972. Morphology and phytogeography: Ann. Missouri Bot. Gard. 59: 107-124 75. Floral anatomy and embryology of Tripetaleia paniculata and T. bracteata (Ericaceae). Bot. Mag. Tokyo 88: 267-280 J Tripetaleia paniculata, B. A.B DEPARTMENT OF BOTANY UNIVERSITY OF BRITISH COLUMBIA VANCOUVER, B. C. V6T IW5, CANADA DEPARTMENT OF BOTANY UNIVERSITY OF BRITISH COLUMBIA VANCOUVER, B. C. V6T IW5, CANADA Present address: BIOLOGY DEPARTMENT ai OF WATERLOO WATERLO ONTARIO | Na 3G1, CANADA SAKAKIBARA, & H. INA. I. Yakugaku Zasshi 91: 1252-1254. 1973. III. /bid. 93: 687-691. 1974. V. Ibid. 94: 1971a. 138-141. 5. W.B DEPARTMENT OF BIOLOGY HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 Present address: SECTION OF ECOLOGY AND SYSTEMATICS CORNELL UNIVERSITY IrHaAca, NEW YorK 14853 ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 Jour. Linn. Soc. Bot. 59: the classical approach to the Studies on the constituents of 1971b. II. 1634-1638. 342 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 LUMNITZERA ROSEA (COMBRETACEAE) — ITS STATUS AND FLORAL MORPHOLOGY P. B. ToMLINSon, J. 5S. Bunt, R. B. PRIMACK, AND N. C. DUKE THREE FORMS of Lumnitzera (Combretaceae) are known to us in (Jueens- land. Two are clearly the common and widespread mangrove species L littorea, with scarlet flowers, and L. racemosa, with white flowers. The third fort is intermediate (eg., it has pink flowers) and is very local in distribution. The identity of this third form is the substance of this paper and is set against the background of a description of floral morphology and pollination biology in the genus. In order to name this pink-flowered form, we have to go back to the botanical information that resulted from the French expedition to the South Pacific (1817-20) under the command of Freycinet in the “Uranie” and ‘‘Physicienne” (Gaudichaud-Beaupré, 1826-30). The following account is an attempt to link observations made recently in Queensland (FicurE |) with the older literature, and to inter- pret some of the illustrations and names provided by Gaudichaud. The genus Lumnitzera is a common mangrove constituent of the Asian tropics, ranging from Ceylon to Fiji. From existing accounts, its taxonomy and nomenclature would appear to be well stabilized, although based pri- marily on a study of herbarium specimens (Merrill, 1909; van Slooten, 1924; Exell, 1954). Exell’s treatment recognizes two clearly contrasted species, which can be separated as follows: a. Flowers ame pedicellate; stamens twice as long as petals; inflores- L BONCES LETMNG os ey tae chbaweoda baad be de . littorea (Jack) Voigt. b. Flowers whi oe ere e; stamens equaling or only slightly exceeding the petals; Interescences BMUaIY.. 45,0431 hike pak ebs dhahbeencues . racemosa Willd. In addition, there may be differences in the anatomy of the floral cup (hy} yanthium) and fruit, but we have not examined these in detail. Exell suggests in his descuiation that Lumnitzera racemosa lacks pneu- matophores and so differs from L. littorea, but in our experience L. race- mosa can develop the characteristic looped aerial roots (pneumatophores) that are more common in L. littorea, the form and degree of development probably varying somewhat in individual trees depending on edaphic con- ditions. The two species are not readily distinguishable in the vegetative state since they have similar leaf sizes and shapes. Lummnitzera littorea has a more diffuse, sprawling habit with a tendency for lower branches to take root. Variability is, however, illustrated by populations of the same species at Jacky Jacky Creek, Cape York Peninsula (FicurE 1), which form an upper canopy of trees up to 30 meters high and 65 cm. d.b.h. The range of the two species in Malesia, as mapped by van Slooten (1924) and more completely by Exell (1954), is very similar. Van Slooten 1978] TOMLINSON ET AL., LUMNITZERA 343 29 TORRES STRAIT & CAPE YORK: t cinaee CAROWELL™ ISLAND PENINSULA Bc Eeoos ISLA —_fcooxtown ira . 148° a. ENDEAVOUR RIVER com Sette, hated 140° 1 m9) oe os Cromnes & F rownsvitce?~\ AS CELERES y > ~NEW Java a 4 —— ee oa PS = S& 20° ISLAND HINCHINBROOK J & ® inten NORTHUMBERLAND pb. ISLANDS ‘o- SHOALWATER BAY PORT \ CLINTON ROCKHAMPTON 148° r Ficure 1. Northern Queensland. Inset figures show (lower left) the Sa alized region and (upper right) the Cairns- ee ok region in which t most detailed ground observations have been made. Lumunitzera littorea and L. racemosa occur throughout the Sige cane L. X rosea occurs only in Missionary Bay, Hinchinbrook Island (arrow) suggests that they occupy different ecological sites and so exclude each other, but does not give precise details. Our own observations confirm countered sequential zones of the two species. Both are confined to the in- 344 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 ner mangrove fringe, to the banks of tidal creeks, and, on occasion, to the upper parts of sandy beaches. Their separate distribution in some instances y be a matter of chance. However, it is also likely that each species has unique tolerances that permit separate establishment under conditions that appear very similar only under casual observation. For example, we have often found Lumnitzera racemosa but not L. littorea growing suc- cessfully in association with Ceriops and Avicennia in the inner mangrove. Equally, it is normally L. racemosa that one encounters under the harsh conditions at the margins of bare salt pans. Lummnitzera littorea appears better suited to less saline, well-drained sites. Our observations suggest that within such conditions this species develops most vigorously in more highly organic substrates and assumes its well-known shrubby appearance on substrates that might be termed skeletal. A more careful analysis of distribution patterns and associated edaphic conditions will be needed to refine these preliminary impressions. NOMENCLATURE A number of other names exist, including Lumnitzera lutea (Gaud.) Presl (= L. racemosa var. lutea (Gaud.) Exell), which was applied to a plant with yellow, axillary flowers, and stamens longer than the petals, de- scribed by Gaudichaud-Beaupré as having gland-bearing calyx tips. Originally described from Timor and for a long time known only from Gaudichaud-Beaupré’s description, this plant has been more recently col- lected by Bouman-Houtman (see van Steenis-Kruseman, 1950, p. 187). This is treated by Exell (1954) as a variety of L. racemosa, which may be appropriate since gland-tipped calyx lobes occur in both varieties. other names but one have even less validity in morphological terms and have been accepted as being based on minor variants of either Lumnit- zera littorea or L. racemosa (van Slooten, 1924; Exell, 1954). The ex- ception is Lumnitzera rosea (Gaud.) Presl: we = that we have re- discovered this entity and can account for its existen Gaudichaud-Beaupré (1826) included names now aces to Lumnit- zera under Laguncularia, the appropriate combinations being made by Presl (1834). Of the four names, Laguncularia lutea is now ee to be a variety of Lumnitzera racemosa (see above), and Laguncularia coccinea is a synonym of Lumnitzera littorea. Laguncularia purpurea is a name used, apparently in error, by Gaudichaud-Beaupré in the plate and legend of the entity that he called L. coccinea. Laguncularia rosea, de- scribed from a specimen collected from Manila, in the Philippine Islands: has subsequently been considered to be a mere color form of Puma racemosa. However, the information given in the diagnosis and the plate allows the three entities to be contrasted clearly (currently accepted no- menclature is used): Lumnitzera littorea (as Laguncularia coccinea and L. purpurea) — in- florescence a terminal, elongated spike, flowers “purple,” calyx lobes without glands, anthers with or without an appendage. 1978] TOMLINSON ET AL., LUMNITZERA 345 Lumnitzera racemosa (as Laguncularia lutea) —inflorescences axillary, flowers yellow, calyx lobes uniglandular, anthers with an appendage. Lumnitzera rosea (as Laguncularia rosea) — inflorescence axillary, flow- ers roseate, calyx lobes triglandular, anthers with an appendage. We will return to these diagnoses after we have considered floral morphol- gy. Subsequent authors have used additional diagnostic characters (e.g., stamen number, as suggested by Gagnepain, 1920) in segregating Lummnit- zera littorea from L. racemosa, but they are neither useful nor consistent. Despite this, the most consistently diagnostic floral feature, that of slight zygomorphy versus actinomorphy, has been overlooked by all authors with the exception of Brandis (1893), who illustrated it clearly. Exell (1954) refers to actinomorphic flowers in his generic description of Lumnitzera, implying that all species are alike in this respect. OBSERVATIONS FLORAL MORPHOLOGY. Floral features common to Lummnitzera littorea and L. racemosa are represented diagrammatically in FiGURE 2. Stamen num- ber is variable, since members of either antesepalous or antepetalous whorls may be missing. Flowers of L. littorea (Missionary Bay, Hinchin- brook Island, Queensland, 18°15’ S, 146°15’ E) are illustrated in FIGURE 4. This shows the short, more or less erect petals, exceeded in length by the erect stamens. Appreciable zygomorphy is shown in the median longi- tudinal section of FicurE 4, B: the calyx tube is oblique, with the style inserted high on one side. This creates a deep tube on the adaxial side of the flower in which abundant nectar, secreted by the inner surface of the tube, accumulates. Zygomorphy is also evident in the slight dorsiventral curvature of the flowers, such that the flower bends upward toward the apex of the spike. The nonmedian position of the style is further shown in the Jongitudinal section in the plane of insertion of the bracteoles (FIGURE 4,C). The stamens are unappendaged in this collection (F1GuRE 4, E), although Brandis (1883) does show a stamen appendage in his illustration of this species. Other collections elsewhere in Queensland show that the stamen appendage is an inconstant character.1 Petals have a broad in- sertion (FIGURES 4, , ae Flowers in Lumnitzera racemosa (FicurE 5) are strictly actinomorphic; the petals are spreading at maturity and are scarcely exceeded by the somewhat spreading stamens. The calyx tube is shallow and symmetrical; the style is inserted in the center so that it is sectioned equally in both dorsiventral and lateral planes (cf. FicurE 5, B, C). Flowers of L. race- mosa@ are appreciably shorter than those of L. littorea, the relative lengths (including calyx tube and pedicel) being about 18 vs. 8 mm., respectively. 1The anther is dorsifixed, but the filament is bent over in bud so th he pendage projects upward. When the filament straightens at anthesis, ‘igs eee: is more or less reversed. 346 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 ee at? FIG 4 littorea PF FIG5 TAcemosa 2-5. ee morphology in Lummnitzera. 2, generalized floral diagram. 3, Lumnitzera & r a (Hinchinbrook Is and, Seen Jan a, flower, side view; tr ne; tudinally in transverse plane; d, see e, deta . of caie anther. 4, Lummitzera littorea (Hinchinbrook Island, Queensland) : flower, side view; b, flower sec- tioned longitudinally in dorsiventral plane io sectioned longitudinally in transverse plane; d, petal; e, detail of ae unappendaged anther. < gill zera racemosa (Hinchinbrook Island, Queensland): a, flower, side vie secti j udi ly in a tens plane; d, peta’ e. oe appendaged anther. DEE ees for figures 3-5: a-d, X 2.5; e, X 6 1978] TOMLINSON ET AL., LUMNITZERA 347 = petals have a narrow insertion (Ficure 5, D; 6, D, E n both species there is a variable number of anatropous ovules (from 3 to 7, with 5 the most frequent number) suspended from a slender pla- centa which itself hangs from the upper part of the ovary loculus, as shown in all illustrations of longitudinal sections of flowers in Ficures 3-5. The ovules are narrow with the micropyle projecting apically. The stamens have a distinct connective appendage (Ficure 5, E), and the ) POLLINATION BIOLOGY. Differences between the species in floral morphol- ogy and biology relate to differing methods of pollination. Lumnitzera littorea is pollinated predominantly by honeyeaters (Meliphagidae) such as the graceful honeyeater (Meliphaga gracilis), as well as by native bees and wasps. Bird pollination is indicated in this species by the red flower color and the abundant nectar that accumulates to one side of the calyx tube. The flowers are longer and the petals are directed forward, orienting the bill of the bird and protecting the nectar from most insects. The sta- mens are directed forward to touch the bill of the bird. The terminal in- florescences accommodate large pollinators. Lumnitzera racemosa is frequently visited by a variety of day-active insects such as wasps, solitary - social bees, butterflies, and day-active moths. Wasps, the most common flower visitors observed, aggressively chased the other flower- ae insects. No insects were observed on the flowers at night. Characters related to pollination by generalized flower- visiting insects are the white flower color and the small amounts of nec- tar secreted into the shallow calyx tube. The stamens are directed to the side to contact insects that have landed on the reflexed petals. The axillary flowers accommodate small pollinators. Pollen is present in the anthers only on the day the flower opens, while the stigma appears to be- come receptive on the second and subsequent days.” As a result of this marked protandry, individual flowers are probably not capable of self- pollination. Fruit-set is high, probably over 50 percent in many trees, and even on isolated individuals, suggesting that the species is self-compatible. In Queensland, flowering times differ somewhat in the two species, but there is appreciable overlap: Lummnitzera littorea flowers from August, when insects are not yet active, to December; L. racemosa from October to March. jar LUMNITZERA IN QUEENSLAND. In addition to the two common species of Lumnitzera in Queensland, which form a characteristic element of the back mangroves, populations of a pink-flowered form have been located at Missionary Bay on Hinchinbrook Island (Ficurr 1). The inflores- cences are terminal, but the flowers are sufficiently distinct to make it dif- ficult to equate this form with ZL. littorea. This population occurs as a pocket in the stand of LZ. littorea that supplied the material for FicurE 4. As shown in Ficures 3 and 6, flowers of this form approach the size of On the basis a a positive test for esterase activity with a-naphthyl acetate (Hes- lop- ee et al., 348 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 lillorea Sg x Yosca OT FIG 6 Pi FicuRE 6. Flowers from above (stamens removed), showing Lae ene and nectary (arrow) in relation to style: a, b, L. littorea; c, L. K r - d,e, L. racemosa; all X 2.5. those of Lumnitzera littorea, but are distinguished by the spreading, pink petals. In addition, they are scarcely zygomorphic; although the style is slightly to one side of the calyx tube (FicureE 3, B), it is still cut when the flower is sectioned in both dorsiventral and lateral planes (FicuRE 3, C). The stamens are shorter than in L. littorea and usually have a minute con- nective appendage, much less prominent than in L. racemosa (FIGURE 3, E). The petals have a broad insertion (FicurE 3, D). Nectar-feeding birds, bees, and wasps have been observed visiting the flowers of this spe- cies. In order to provide an identity for this species, we have adopted the name Lumnitzera X rosea Gaud., pro sp., making the assumption that the taxon Gaudichaud- is illustrated corresponds to the form we our- selves have more recently discovered. We provide evidence that the form is a hybrid, L. littorea x racemosa. This is a considerable assumption, first because L. rosea probably originated in the Philippines, and sec- ond because Gaudichaud-Beaupré illustrated and described a plant with axillary inflorescences. Third, the type locality for L. rosea is believed to be Manila Bay, but, of the two putative parents, only L. racemosa has been collected here. However, using this name, we can summarize our field observations of Lummnitzera in Queensland in TaBLe 1. We have found, as did Exell, that the number of sepal glands is not very constant and does not provide a reliable means of distinguishing Lumunit- zera species. In addition, Lummnitzera x rosea is closer to L. racemosa in 1978] TOMLINSON ET AL., LUMNITZERA 349 TABLE 1. Floral features of Lumnitzera species in Queensland. CHARACTER SPECIES littorea x rosea racemosa Inflorescence terminal terminal axillary Petal color scarlet pink white Flower length (from insertion to style tip) 16-18 mm. 13-14 mm. 7-8 mm. Petal position + erect + reflexed + reflexed Symmetry slightly zygo- almost actino- strictly actino- morphic morphic morphic Stamen appendage sometimes absent minute present Flower visitors birds, insects birds, insects insects its more nearly sessile flowers and in the anatomy of the floral cup since it has few rather than many vascular bundles in the cortical region. It resembles L. littorea more in its early flowering time, but L. racemosa most in habit. However, natural layering of branches occurs as it does in L. littorea. STATUS OF LUMNITZERA & ROSEA. The intermediate status of this pink- flowered form strongly suggests its hybrid origin. This is further supported by pollen sterility of about 40 percent, a figure representing those grains that are collapsed and shriveled and that remain unstained in iodine-po- tassium iodide solution. Pollen sterility in both putative parents is less than | percent. DISCUSSION Current evidence from the collections and observations made in Queens- land ape strongly suggests that hybrid Lummnitzera littorea racemosa occurs as occasional individuals when the parental species grow together. The in that this taxon may have been recognized earlier by Gaudichaud-Beaupré and called L. rosea is a reasonable hypothesis, but it rests entirely on the circumstantial evidence of petal color as he re- corded it, since floral morphology is not illustrated in the necessary detail. The zy gomorphic vs. actinomorphic condition, which would permit a clear- r assessment of the specimen to which Gaudichaud-Beaupré referred, i not indicated in his illustrations. Indeed, the existence of zygomorphy in L. littorea, recognized and clearly illustrated by Brandis (1883), has been obscured in subsequent accounts. Zygomorphy is not a peculiarity of Queensland populations since it has also been seen by one of us (P. B. T in both Fiji and New Guinea. It is one part of the syndrome of characters that relates to the probable bird-pollination mechanism of this species, and 350 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 that contrasts it with the insect-pollination of L. racemosa. This differ- ence in pollination mechanism may be sufficient to provide a specific steril- ity barrier that normally operates despite the fact that the species are often associated. Occasionally this barrier is broken, and hybrids may ap- pear, although they are likely to be regarded as a pink-flowered form of either Lumnitzera littorea or L. racemosa, depending on the position of the inflorescence. A problem in our use of this name is that Gaudichaud-Beaupré described and illustrated a plant with axillary inflorescences.* This could be ex- plained if it is accepted that inflorescence position is not always dominant in the cross. A test of our hypothesis would be an artificial cross, which will be attempted. Breakdown of the mechanism of isolation could be the result of bees and wasps visiting the two species successively, as our field observations show is possible. Persistence of hybrids in Lummnitzera is facilitated by vegetative spread that results from rooting of pendulous lower branches. A fairly extensive survey of mangrove populations in northern Queensland has not revealed any further populations of L. « rosea (although the two parent species are common), but does show minor variants in floral morphol- ogy. Populations of L. littorea at Jacky Jacky Creek have flowers with somewhat spreading petals. This might be the result of introgression from L. racemosa. Future research is designed to substantiate our very incomplete obser- vations on the floral biology and ecology of the putative parents and to provide cytological documentation that will advance our statements be- yond the hypothetical. ACKNOWLEDGMENTS Support for the field work of P. B. Tomlinson in Queensland (and on mangroves elsewhere in the South Pacific) has come via grants from the National Geographic Society (1973, 1976) and in 1977 from Grant INT 76-24479 of the Office of International Programs (Australia-U.S. Co- operative Program in Science), National Science Foundation. The floral illustrations were made by Priscilla Fawcett, Botanical Illustrator, Fair- child Tropical Garden, Miami, Florida. The study was carried out with field, logistic, and laboratory support from the Australian Institute of Marine Science. LITERATURE CITED BRANDIS, D. 1893. Combretaceae. Jz: A. ENGLER & K. PRANTL, Nat. Pflanzen- fam, ed. 1. III(7): 106-130. * More recent examination of the type specimen, Perottet 6 from Manilla (Pp), shows that the inflorescence can be interpreted as terminal, and the above problem may not exist. Unfortunately, the specimen lacks flowers, and the more precise diagnostic fea- tures cannot be applied. 1978] TOMLINSON ET AL., LUMNITZERA 351 EXELL, A. W. 1954. Combretaceae. Jn: C. G. G. J. VAN STEENIS, ed., Flora Malesiana I(4): 533-589 GAGNEPAIN, F. 1920. Combretaceae. Jz: H. Lecomte, ed., Flore générale de l’Indo-Chine 2(6): 734-777. GAUDICHAUD-BEAUPRE, C. 1826-30. Botanique. J: L. DE FREYCINET, Voyage autour du monde fait par ordre du roi, sur les corvettes de 5. M. l’Uranie a Physicienne. 493 pp. Paris, Pillet. ea ee J., R. B. Knox, Y. Hestop-Harrison, & O. Mattson. 1975. Pollen-wall proteins: emissions and role in incompatibility responses. Pp. 182-202 in J. G. Duckett & P. A. Racey, eds., The biology of the male gamete. Academic Press, London. Jounston, I. M. 1944. Publication-dates of eras Botany of the Voy- age of the Bonite. Jour. Arnold Arb. 25: 48 MERRILL, E. D. 1909. A aera revision - Philippine Combretaceae. Philip. Jour. Sci. Bot. 4: 641- Presi, Kk. B. 1834. Repertorium ete systematicae. Vol. 1, pp. 155, 156. Prague, Filiorum Theophili Haase. SLoorEN, D. F. van. 1924. Contributions a l’étude de la flore des Indes néer- landaises II], The Combretaceae of the Dutch East Indies. Bull. Jard. Bot. Buitenzorg III, 6(1): 11-64. STEENIS-KRUSEMAN, M. J. vAN. 1950. Malaysian plant collectors and collec- tions. Iz: C. G. G. J. vAN STEENIS, ed., Flora Malesiana I(1): 3-639. PB. T, J.S.B.anp N.C. D HARVARD FOREST INSHORE PRODUCTIVITY GROUP HARVARD UNIVERSITY AUSTRALIAN INSTITUTE OF PETERSHAM, MASSACHUSETTS 01366 MARINE SCIENCE CAPE FERGUSON R. B. P. PMB No. DEPARTMENT OF BIOLOGY TOWNSVILLE, M.S. O Boston UNIVERSITY QUEENSLAND 4810, AUSTRALIA Boston, MASSACHUSETTS 02215 352 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 TRANSFER OF THE BRAZILIAN TRIXIS ERYNGIOIDES TO PEREZIA (COMPOSITAE, MUTISIEAE) JORGE Victor CRISCI AND CLODOMIRO MARTICORENA IN THE COURSE of a revisionary study of the South American species of the genus Trixis Browne (Compositae tribe Mutisieae subtribe Nassau- viinae), Trixis eryngioides Cabrera, a species occurring in southeastern Brazil, seemed to be out of place. A careful study of this species and of all the genera of this subtribe convinces us that Trixis eryngioides must be transferred to the genus Perezia Lagasca. A description of the plant, illustrations, pollen analysis, and discussion of the relationships of this taxon are presented here. The abbreviations for herbaria listed in the citations of specimens are taken from the fifth edition of the /ndex Herbariorum (Lanjouw & Stafleu, 1964), MATERIAL AND METHODS Some of the pollen grains were acetolyzed according to the method out- lined by Erdtman (1960); others were placed in 95 percent ethanol, stained with basic fuchsin, and mounted in glycerine jelly. For scanning electron microscopy (SEM) the acetolyzed pollen grains were used. After acetolysis the pollen samples were washed several times in glass-distilled water to remove traces of acids. Each sample was dispersed in a drop of ethanol (95 percent) and placed on a specimen holder. The samples were then air dried, coated with 50 A to 100 A carbon and about 200 A to 300 A gold-palladium alloy for conductivity. Finally, the specimens were photo- graphed using a high-resolution SEM (JEOL JK IT) at the Servicio de Microscopia Electrénica de Barrido del Consejo Nacional de Investiga- ciones Cientificas y Técnicas, Buenos Aires, Argentina. The preparation of the material for sections was made according to the Leins technique (Leins, 1968) with the modifications made by Parra and Marticorena (1972). The sections of ca. 0.5 pm. were made with an AO Spencer microtome adapted for ultrathin sections and using glass blades. Perezia eryngioides (Cabrera) Crisci & Marticorena, comb. nov. FicurEs 1-6, Trixis eryngioides Cabrera, Bol. Soc. Argent. Bot. 7: 196. PRS. 1959; Robust rosette plants, 60-70 cm. long, erect, from a taproot. Flowering stems terete, 5—7 mm. in diameter, slightly striated, glabrate, bearing 10 to 30 cauline leaves. Cauline leaves alternate, sessile, amplexicaul, ovate- lanceolate, acute, dentate, glabrate, 6-12 cm. long, 2—2.5 cm. wide. Basal leaves usually numerous, lanceolate, acute, amplexicaul, spinulose-dentate, 1978] CRISCI & MARTICORENA, PEREZIA o08 m4 7 SF i Hh i} iy ES Zs EES ae. ge = og Pog ee SKS eS {LEE FicuRE 1. Perezia eryngioides Sioa eige é Klein 7777): A, om : ee < 4; B, capitula, & 1; C, corolla, x 2 ; D, achene and pappus, 354 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 glabrate, 20-34 cm. long, 3-4 cm. wide. Inflorescence a branched terminal cyme; pedicels 1-2 cm. long, surface with glandular trichomes. Individual capitula campanulate, 10-12 mm. long, 12-15 mm. wide. Involucre hemi- spheric, formed by 2 rows of bracts; involucral bracts lanceolate, acute at apex, 8-12 mm. long, 1.5-2 mm. wide, abaxial surface pubescent, adaxial surface glabrous, inner bracts scarious at the base and along the margins. Receptacle slightly concave, pubescent. Flowers about 20; corolla white, bilabiate, 8-15 mm. long; tube broader toward the apex, 5—7 mm. long, 0.3 mm. in diameter at the base and 1 mm. at the apex; outer lip ceo: late, liguliform, with 4 veins, apex 3-toothed, 4-8 mm. long, 1.5-2 mm. wide; inner lip bifid, acute at apex, 2.5-3 mm. long; stamens 5, glabrous, the filaments inserted at a point halfway from the base of the tube; an- thers connate, 4-4.5 mm. long, with oblong apical appendage, tailed at base. Style bifid, its branches 1-1.3 mm. long, truncate, with a crown of elongate collecting hairs at the apex, papillose with shorter hairs on the adaxial surface, glabrous on the abaxial surface. Achenes turbinate, 2.5—3 / ] eet () afb h FicureE 2. Pollen. A-D, Perezia eryngioides (from Smith & Klein (iif)? A, equatorial view, 4 1250: B, polar view, X 1250; E-I, Trixis alata (from Fréderstrom & Hultén 202): o Rater view, x 1250: , membrane of colpi with spherical processes, X 2 G, cross section show- ing exine stratification, X 2000; H, pattern seen by pwr of tectum, * 2000; I, pattern formed by bacula of infratectum, x2 ales 1978 | CRISCI & MARTICORENA, PEREZIA 355 Scanning electron micrographs of pollen of Perezia al cars rae Smith & Klein ees 3, several acetolyzed grains, arrow indicates the : wm. ; and portion of surface of membrane of colpi (showin 2 spherical processes), scale bar = ca. 2 um.; 6, surface, showing conical spinules, scale bar = 2.5 em mm. long, 1-1.5 mm. wide, covered with double fe Pappus of numerous dark brown setose hairs, 7-10 mm. long, in 2 serie Pubescence of two kinds of trichomes: (1) ie biseriate (25-50 m.) —multicelled (borne on leaves, stem, receptacle, involucre, and sparsely on the achenes); and (2) double hairs or “Zwillingshaare” (Hess, 1938) (120-150 »m.) — two cells joined along their inner walls but not meeting at the apices, copper colored (on the achenes). PoLLEN. Grains 3-colporate, exceptionally parasyncolpate, spheroidal- oblate to spheroidal-prolate, 28-34 *« 29-35 ym. Apocolpium 2.5-3.5 pm. in diameter, slightly prominent to plane, rarely slightly depressed. Colpi 356 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 59 3—4.5 um. wide, with margins entire or slightly curled, extremes obtuse to acute; membrane of the colpi with notable spherical processes, 0.5-1.2 um. in diameter. Lalongate ora 2.3—5 ym. in external polar diameter, 4.7—5 pm. in internal polar diameter, and 10-16 um. in equatorial diameter; equator- ial extremes acute. Amb rounded. Exine crassisexinous, pertectate, spinu- lose, 2.5—3 ym. thick at the pole, 3-5 ym. at the equator. Spinules conical, ca. 0.5 ym. high. Terminal membrane ca. 0.5 wm. thick. Bacula usually laterally fused, forming curled internal walls and leaving spaces of labyrin- thic shape. Tectum and infractectum separated from each other by a thin layer not parallel to the nexine (zigzag). Infratectum 1-1.3 pm. thick, baculate in a similar fashion to the tectum. Nexine ca. 0.7 ,m. thick. DISTRIBUTION. Southeastern Brazil, province of Santa Catarina, in Bom Retiro, Irani, and Lajes municipios, from 700 to 1700 meters altitude. Flow- ering from October to December. SPECIMENS EXAMINED. Brazil. SANTA CATARINA: Bom Retiro, Fazenda Campo dos Padres, dwarf forest, 1650 m., L. B. Smith & R. M. Klein 7777 (xp, holo- type of Trixis eryngioides Cabrera); Lajes, Painel banhado, 950 m., Klein 4566 (LP); Irani, Campo de Irani, bog, 700-900 m., Smith & Klein 13031 (1p). Perezia eryngioides has been collected very few times and seems to be rare. It is found in humid places in the western zone of the ‘“‘campo” or prairie flora (Smith, 1962) of southern Brazil. RELATIONSHIPS OF PEREZIA ERYNGIOIDES Perezia eryngioides is a member of tribe Mutisieae subtribe Nassau- viinae, which has characteristic bilabiate corollas with the outer lip three- toothed and the inner lip bifid; anthers long tailed at the base; and style bifid with branches truncate at the apex (Crisci, 1974). Perezia eryngioides was described as a species of Trixis; it is, however, very different from the other species of this genus, as is indicated in TABLE | In addition to the criteria shown in TaBLe 1, there are a number of other features that may be labeled “tendencies”? common to some species of Trixis but from which there are exceptions: pubescence including woolly trichomes, capitula with less than 15 flowers each, leaves very short-petiolate or subsessile. Trixis P. Browne?! is a genus of about 50 species of shrubs occurring from Mexico to central Argentina and in the West Indies. There are two centers of diversity, one in southwestern Mexico, the other in southern Brazil, northern Argentina, and Uruguay (Anderson, 1972; Cabrera, 1936: Loja, 1969). Perezia eryngioides is better placed in Perezia since it shares with the species of this genus an herbaceous habit, basal rosettes, achenial tri- chomes, white corollas (in Perezia the color of corollas is variable but * Hist. Jamaica, 312.1756. Type species: Trixis inula Crantz. 1978 | CRISCI & MARTICORENA, PEREZIA Sot TABLE 1. A comparison of the genus Trixis and Perezia eryngioides. Trixis Perezia eryngioides Hasit Shrubs; not rosette plants Perennial herbs; rosette plants COROLLA Bright yellow White Colorless, ca, 150-200 um. Copper colored, ca. 120-150 DOUBLE long, oblong in shape, obtuse um. long, linear in shape, short- HAIRS ON at apex, less than 5 times bifid at apex, at least 20 times ACHENES longer than wide, releasing longer than wide, not releasing mucilage when we mucilage when we Prolate (rarely subprolate) : Oblate-spheroidal to prolate- exine with infratectum thicker spheroidal; exine with tectum POLLEN than tectum and separated thicker than infratectum and from it by a thick layer separated from it by a thin parallel to nexine (see layer not parallel to the FIGURE 2) nexine (zigzag ) never bright yellow), and pollen morphology, including exine stratifica- tion. Also, both have more than 15 flowers per capitulum Perezia Lagasca” is a genus of 32 species of herbs, 27 occurring in the Andean region of South America and 5 (including P. eryngioides) in low- land open woods of Paraguay, Uruguay, Brazil, and Argentina (Vuil- leumier, 1969; Cabrera & Klein, 1973; Crisci, 1974). Vuilleumier (1969; 1973 as B. Secon) clustered the species of Perezia into six well-defined groups: Prenanthoides, Coerulescens, Magellanica, Recurvata, Pungens, and Multiflora. One of the most distinctive and well-defined groups centers around Perezia multiflora and consists of four species: P. multiflora (H. & B.) Less. (with 2 subspecies, multiflora and sonchifolia); P. squarrosa (Vahl) Less. (with 2 subspecies, sqguarrosa and cubataensis); P. kingit Baker, and a species described since Vuilleumier’s paper, P. catharinensis Cabrera. All of the species of this assemblage are found exclusively or have populations in the Paraguay-southern Brazil-Uruguay basin, although P. multiflora itself has its main center of distribution in the high, dry puna of Peru and Bolivia. Within the genus, the four species of this group share the unusual characters of silky, copper-colored achenial trichomes, reduced number of involucral bracts, pubescence on the receptacle, and small pollen size (24-30 pm.). In addition, there are a number of features that are not exclusive to the Multiflora group such as small head size, compacted head clusters, taproots rather than rhizomatous rootstocks, and spinifer- ous foliage Members of the Multiflora group possess the lowest chromosome num- bers in the genus, including the probable base number x = 4 (Vuilleu- 7 Am. Nat. 311. 1811. Type species: Perezia magellanica (L. f.) Lagasca (= Perdicium Sie ee L. 358 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 mier, 1969), suggesting that this group is a primitive one in the genus. However, as was pointed out by Vuilleumier (op. cit.), it does not appear to represent the basal complex, but seems rather to be a closely knit cluster of species that radiates in southeastern Brazil from a very early offshoot of the main ancestral stoc Perezia eryngioides belongs to the Multiflora group since it possesses all the characters that define the assemblage. Although the Multiflora group is a very tightly knit assemblage, Perezia eryngloides seems to be more closely related to P. SATO: (especially subsp. cubataensis) than either of these two species is to P. kingii, P. catharinensis, or P. multiflora. The two taxa are sympatric in the P. eryn- gioides range and can be distinguished on the bases of corolla color and basal leaf size. Perezia squarrosa is found in humid places in southern Brazil, Paraguay, and Uruguay, at low elevations and even at sea level. To summarize, we can establish that Perezia eryngioides is easily dis- tinguished from all the species of Trixis by its herbaceous habit, white corollas, characteristic achenial trichomes, and pollen morphology, includ- ing exine stratification. It shares these characters with the species of the genus Perezia, From most of the species of Perezia, it can be distinguished by the color of its achenial trichomes, the reduced number of involucral bracts, the pubescence of the receptacle, and the small size of its pollen grains, characters it shares with P. multiflora, P. kingii, P. catharinensis, and P. squarrosa. It can be distinguished from these four species as shown in the following key. A. Capitula less than 10 mm. long. B. Involucral bracts with small white spines near the base; pappus pure VANE aad Gand d od & 2h ve he ee he et Be a P. kingit. B. Involucral bracts without spines; pappus brown. ..... P. cotharinensis. A. Capitula longer than 10 mm. C. Capitula hemispheric; florets bicolored, with yellow on the inner part of the corolla tube and the inner petals; heads often congested. be eae Seed foes ta ees ye andes eee P. nile fore. C. Capitula campanulate: florets monochromatic: heads in a 2 loose cyme. D. Basal leaves 4-20 cm. long, 0.5-2.5 cm. wide: corollas blue, violet, or POGUE nde teat ese ee de oa aiceck bad Raabe be eek P. squarrosa. D. Basal leaves 20-34 cm. long, 3-4 cm. wide: corollas white. ba onto oa eS Utcuae osetia ee 6a ee ot ees dee P: eryngivides. LITERATURE CITED ANDERSON, C. 1972. A monograph of the Mexican and Central American species of Tris (Compositae). Mem. N. Y. Bot. Gard. 22(3): 1-68 CABRERA, A. L. 1936. Las especies argentinas y uruguayas dal género Trixis. Revicha Mus. La Plata Bot. (n. s.) 1: 31-86. a R. M. Kien. 1973. Compostas. Tribo Mutisieae. ae P. R. RE1rTz, , Flora ilustrada catarinense. I Parte, Fasciculo Com 2 ie 7. V. 1974. A numerical-taxonomic study of the mui Nassauviinae (Compositae, Mutisieae). Jour. Arnold Arb. 55: 568-610 1978] CRISCI & MARTICORENA, PEREZIA 359 ErptTMAN, G. 1960. The acetolysis method. A revised description. Sv. Bot. Tidskr. 54: 561-564. Hess, R. 1938. Vergleichende Untersuchungen uber die Zwillingshaare der Compositen. Bot. Jahrb. 68: 435-496 see Ts CoE AG ee | Index herbariorum. Part 1. The herbaria world. ed. 5. Reg. Ve LEINS, . 1968. Eine einfache vee zur Herstellung von Schnitten durch azetolysierte Pollenkérner. Grana Palyn. 8: 252-254. Loya, B. Herrera ALARCON DE. 1969. Revision de las especies peruanas del género Trixis. Publ. Mus. Hist. Nat. Lima Bot. 25: 1-16. Parra, O., & C. Marticorena. 1972. Granos de polen de plantas chilenas. II. Compositae - Mutisieae. Gayana Bot. 21: 1-10 Simpson, B. B. 1973. Contrasting modes of evolution in two groups of Perezia (Mutisieae; Compositae) of southern South America. Taxon 22: 525-536. SmitH, L. B. 1962. Origins of the flora of southern Brazil. Contr. U. S. Natl. . 35: 215-249. VUILLEUMIER, B.S. 1969. The systematics and ev She of Perezia sect. Perezia (Compositae). Contr. Gray Herb. Harvard Univ. 199: 1-163. Jove, C. M. DIVISION PLANTAS VASCULARES INSTITUTO CENTRAL DE BIOLOGIA Museo DE La PLATA UNIVERSIDAD DE CONCEPCION 1900 La PLATA CONCEPCION ARGENTINA CHILE 360 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 A REVISION OF THE WEST INDIAN VERNONIAS (COMPOSITAE) STERLING C, KEELEY VERNONIA IS A widely distributed genus with nearly 1000 species, half of which are found in tropical areas of the New World. The diversity inherent in such a wide-ranging group has caused numerous problems in its systematics and taxonomy. The West Indian vernonias have been particularly troublesome in this regard due to their variable vegetative features and the apparent close relationships among species. Gleason pro- vided the first comprehensive treatment of taxa in the West Indies in his 1906 revision of the North American Vernonieae. Ekman (1914) also studied and revised the West Indian vernonias, but disagreed with many of Gleason’s specific and subsectional assignments. Since Gleason up- dated his earlier revision in 1922, several new species have been de- scribed in isolated publications, but no effort has been made to evaluate the merit of overall taxonomic relationships or treatments. Prior to the treatments of Gleason and Ekman, major contributions to Vernonia taxonomy in the West Indies were made by Grisebach (1861, 1862, 1866, 1879), Lessing (1829, 1831), de la Sagra (1850), Schultz-Bi- pontinus (1847, 1863), and Urban (1899, 1903, 1911, 1912). A complete history of the tribe Vernonieae and its subsections is given in Jones (1977). The present study has encompassed a period of several years during which observations and collections were made in the field, greenhouse ac- cessions were grown and studied, and loans from major herbaria were examined. The main descriptive and revisionary aspects presented here are drawn from more than 2500 specimens from eight herbaria, but species concepts were strongly molded by field observations. Forty-two species are treated here; keys and synonymies are provided. Evolutionary and biogeographic comparisons are not treated because relationships with main- land taxa are presently unknown. MATERIALS AND METHODS Loans of herbarium material were obtained from F, GH, Mo, NY, 5, TEX, and us. Additional type material was generously provided by B, BM, GOET, K, L, M, and p. Population samples were collected during December, 1974 (Jamaica, Hispaniola, Puerto Rico), and December, 1975 (Jamaica, the Lesser Antilles, Trinidad). Greenhouse accessions were established from achenes and/or rootstocks collected during these trips and were used to obtain material for chromosome counts and morphological comparisons. Twenty-five to one hundred herbarium specimens were selected and scored for each putative taxon examined, except where fewer than twenty- five specimens were available, in which case all specimens were scored. In 1978 | KEELEY, VERNONIA 361 scoring those taxa with large available collections, the number studied depended on the degree of taxonomic confusion and on the variability of the material. Local populations collected in the field were also scored to obtain a measure of variation due to environmental factors. The number of individuals available for this type of comparison varied widely and was determined by distribution and accessibility. Herbarium and field-collected specimens were measured or scored for the following characters: plant height, leaf blade length and width and ratio of these two measurements, petiole length, involucre height and width, inner phyllary length and width, number of flowers per head, achene length and surface, inner and outer pappus length, corolla length, and anther length. Midcauline leaves were selected for measurement on each specimen. Only fully matured heads were used for measurements of length and width of involucres and phyllaries. Anther length and corolla length were measured on soaked, fully mature flowers. Means and ranges were determined for the numerical values of each character. An attempt was made to use the descriptive terms of Featherly (1954) and the shapes outlined by Lindley (1848, pp. 345-383). RESULTS AND DISCUSSION Characters scored and ranges of numerical values are given in the spe- cies descriptions. Distributions were determined from herbarium speci- mens and field collections. Representative specimens are cited for each species, but the lists have been kept brief. Distributions are shown on maps. A complete list of exsiccatae is available from the author on request; all specimens examined have been annotated. Chromosome counts were determined using the standard aceto-carmine squash technique with pollen mother cells (PMC’s). Most taxa are diploid n = 17, but one hexaploid species, Vernonia trinitatis, was discovered. Chromosome numbers, when available, follow the species descriptions. Species treatments and subsectional assignments differ in many cases from those of Gleason (1906, 1922) and Ekman (1914). The number of species recognized has been reduced from 60 (Gleason, 1922) and 61 (Ek- man, 1914) to 42 in this treatment. Subsectional differences have been detailed, and a comparison is presented elsewhere (Keeley & Jones, 1977a). The range of major West Indian subsections and the general morphology of the inflorescence of taxa in each subsection are shown in Ficure 1. Keys to sections and subsections are presented first as an aid to identi- fication. This key is followed by keys to the species within each subsection. Key To SECTIONS AND SUBSECTIONS OF WEST INDIAN VERNONIA A. Plants annual; sect. TEPHRODES: one widely distributed, pantropical weed. Banded cee 8a 4: etna ok Sk B cata gee ee RE sa 42. V. cinerea. A. Plants perennial; sect. LEPIDAPLOA. B. Herbaceous perennial. VII. subsect. PANICULATAE: one species, Florida and the: -Babaias: (0 4 60.5 3s barren ea eee ea e204 41. V. blodgettit. 362 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 B. Suffruticose perennials, restricted to the West Indies, . Achenes glabrous or resinous, ca. 4 mm. long. . Achenes 3- or 4-angled or with faint ribs; involucres campanulate. C. Achenes pubescent, 2-3 mm. long. E. Leaves membranous, glabrous or minutely pubescent on both sur- faces, margins serrate. V. subsect. PALLESCENTES: one species, St. WUEMLs erence ceca ee Lab oe Das 39. V. pallescens. E. Leaves more or less coriaceous, pubescent especially on lower sur- =] ob js) cad oO wm ra) @ i=) a fae) p o io] wn oO wn a ° an Tc mm. ie) =. a. ra) < = ie mn = a Seg =e =) o mn i) se, se) oO rat) od. =] ga G. Corollas with prominent hairs in clefts of corolla lobes. VI. subsect. SCORPIOIDES: one species, Trinidad and Haiti. eee ean bani «duh bea eon eee 40. V. scorpioides. F. Inflorescence an open scorpioid cyme, or if condensed then not FicuRE 1. Distribution and general morphology of subsections of Vernonia section LEPIDAPLOA in the West Indies: ARBORESCENTES (a), PALLESCENTES (b), SCORPIOIDES (Cc), SAGRAEANAE (d), PoLYANTHES (e), and BUxIFOLIAE (f). 1978] KEELEY, VERNONIA 363 KEY TO SUBSECTION J, ARBORESCENTES A. Pappus white B. Leaves linear or narrowly linear-lanceolate c. aa shiny-pustulate above, slandulat with pronounced pubescence on veins of lower surface. ..................... 25. V. commutata. c one eat white-strigose above, sericeous with inconspicuous pubescence on veins of lower surface. .......... 24. . steno phylla. B. Leaves orbicular, ovate, lanceolate, obovate to oblanceolat D. Heads solitary; leaf margins conspicuously repand; ener cliffs, OAs ct asee dk eee es edo eae eee aed 22. V. complicata. D. Heads in cymes or clusters; leaf margins various but not conspicuous- ly repand; various habitats ee the West Indies. Ey lowers per head 5 (07. .vcus sks u Sy eeak hese ns 5. V. segregata. E. Flowers per head oe orm Leaves ovate, pees obovate to obcordate, less than 3(5) cm. long. G. Plants low-growing shrubs, sprawling; petioles less than 1 mm. long; Hispaniola, ................ 21. V. fruticosa. G. Plants upright shrubs; petioles more than 1 cm. long; vari- ous locations in the West Indies. H. Leaves crowded on terminal branches, gray-green, white- tomentose below; Bahamas. ........ 23. V. arbuscula. H. Leaves evenly spaced along stems, yellow-green, glandu- lar, closely pubescent below; Cuba. Involucres 11-12 mm. high? phyllaries purple, nar- rowly and stiffly lanceolate. ..... 18. V. pineticola. I. Involucres 3-6 mm. high; phyllaries gold to Cale deltoid to ovate or lancsolate. but not stiffl PT Dee ny Te eee ee See eee i. V. pee F. Leaves ovate, oo to lanceolate or oblanceolate, more than 5 c ong. J. Stems ea faintly ribbed, closely pubescent, and con- spicuously green even when old; Trinidad. ............. Sage e SARS ea ela om wk ou ee CA eS 12. V. trinitatis. J. Stems distinctly ribbed, variously pubescent, becoming glabrate with age, gray to brown; various locations through- out the West Indies K. Phyllaries loosely imbricate in few series; achenes par- tially exposed even in immature heads. .............. ee ee ee ere eee een ee ee 20. V. gnaphaliifolia. K. Phyllaries firmly imbricate in at least five series; achenes concealed except in fully matured, reflexed eads. L. Involucres 6-11 mm. high. M. P i aaa in more than six series; involucres long-campanulate or turbi- Sey aE eee Reb eees ae ee dee eee eee eased es 17. V. desiliens. ee in five or fewer series; involucres broadly campanulate. N. Corollas 5.5-9 mm. long; anther less than 2 mm. long. an JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 O. Phyllaries pale yellow-green, tips recurved. ..... 15. V. jenssenii. OQ. Phyllaries gold to brown, tips erect or if curved not uniformly so. bene nn ea oe ee ee ee V. leptoclada. anthers more than 2 mm. long. |. 19. V. yunquensis. N. Corollas 9 mm. or longer; Le Cenee ies than 5(6) mm. high. eaves glabrate, resinous above, with prominently pubescent veins on low- Oe UAC, oak vicky een vee een eae eeetowawnes: 16. V. urbaniana. P. Leaves pubescent, resinous, with caehonnee on lamina and veins of lower surface. Q. Anthers 2.5-3 mm. long; phyllaries uniformly tawny, often with a strong central nerve extending into the tip; Martinique, St. Lucia, St. WIQCEO Genco eeteeeeyes dbase ent epee 1. V. arborescens. Q. Anthers ca. 2 mm, ones phyllaries gold to brown, with dark tips, lack- ing a strong central nerve; various locations throughout the West Indies. R. Leaves oblong to lanceolate, TOW f CUAUCLOUD Es ao nee eG ne oOUd bade g asa R. Leaves ovate to elliptic, densely pubescent at least on younger ba. 10. V. divaricata. sparsely pubescent above and_ be- 7. V. albicaulis. leaves; Jamaica, Cu sity aay A RARE eae A. Pappus brown or tawny. S. Leaves whorled; plants sprawling. ................. 5. Leaves alternate; plants upright. T. Leaves broadly ovate to orbicular, the apex blunt, rarely acute. U. Phyllaries imbricate in ten or more spiral series; leaf margins densely white-pubescent. ...0..0..0............. 4. V. rigida. U. Phyllaries imbricate in five or fewer series, not conspicuously spiraled; leaf margins glabrous, not white. |... _V. harrisiz. T. Leaves elliptic to oblong or long-ovate, the apex acute to acuminate, 3. V. verticillata. — obcordate. . Flowers per head less than 8; Blue Mts., Jamaica. be eeu V. pluvialis. > V. Flowers per head more than 10; various locations. W. Leaves glabrous or sparsely pubescent, not rugose, the apex acuminate, rarely acute; Jamaica. .......... 1. V. acuminata. . Leaves pubescent, at least on lower surface, often rugose, apex acute to obcordate; not in Jamaica. X. Flowers per head ca. 18; plants sprawling ig vinelike; florescence of zigzag cymes. ......... . borinquensis. X. Flowers per head ca. 12(15); plant sigeunestene inflores- cence of smooth cymes. Y. Anthers less than 1.5 mm. long; leaves dark green and sparsely pubescent above, white-sericeous below, linear 8. V. sericea. = n- Y. Anthers more than 2 mm. long; leaves green on both surfaces, minutely pubescent to glabrous above and be- low, ovate to lanceolate, occasionally obcordate. ...._- 7. V. albicaulis. 1978 | KEELEY, VERNONIA 365 > >> > > KEY TO SUBSECTION II. BUXIFOLIAE . Pappus of two equal series of bristles, appearing as one. .... 26. V. barkeri. . Pappus of two series of bristles, the outer series ca. 1/10 as long as the inner. B. Corollas more than 9-11 mm. long; leaves sessile or with resin-dotted petioles to 0.5 mm. long; pappus yellow-orange. . . es Oe Vm oe B. Corollas 6-8 mm. long; petioles 1-2 mm. long, not conspicuously esin- ous; pappus yellow to pink. ..................... a Vs Her iennd KEY TO SUBSECTION III. POLYANTHES Corollas resinous; petioles 8-15 mm. long. ........... 31. V. menthaefolia. . Corollas glabrous; petioles 1-6(-8) mm. long. B. Flowers per head 15 or more; pappus white or very pale tawny. ....... 4d Doc sues afl: obi eh ucaigte an a ge Se eee a 29. V. a Pa reen B. Flowers per head less than 10; pappus yellow-brown. ....... PO Cee Seen eae ee re ee ree re ree eer ee 30. V. hieracioides. KEY TO SUBSECTION IV. SAGRAEANAE ) Blowers per head,-35 (0 O0< ..2iviw Gea esee eSeeteee eE PA 32. V. aronifolia. . Flowers per head 30 or les B. Involucres narrowly cylindrical: flowers per head 4 to 8. ............ Pay ete not hd BS a eee Ee ae ee ee 33 _ pur purata. B. Involucres campanulate; flowers per head 9 to 1 C. Leaves rugose, shining, glabrate and scabrous above, hirsute to to- mentose below D. Leaves with prominent, pubescent veins on lower surface, sparse- ly pubescent to glabrate on lower surface of lamina, oblanceolate to lanceolate, widest above the middle, length to width ratio ca. 3. OT eee Tee Tee ee et eye ree ae 34. V. a D. Leaves without prominent veins, densely wubeccent < on lower su face, lanceolate, widest at or below the middle, a to width ra- 3 TOCA. O. 2s ens eich dee a eee eee ee ae 5. V. sprengeliana. C. Leaves coriaceous, not rugose, dull to shiny, generally pubescent above and below. E. Leaves with ad long, white hairs above and below; corol- CO rie OE 1ONCCE: 25.461 unm mosey aang ea wok S02: viminalis. E. Leaves glabrous i. minutely a cag above, closely resinous-pu- bescent below; corolla 6-9 m F. Phyllary tips recurved, slightly pubescent: pappus whit Sheth ett ods Oh HtSee AROSE URED BAAS or. Vs wrightii. F. Phyllary tips straight, villous; pappus brown to pink. ...... 38. V VERNONIA Section LEPIDAPLOA Subsection I. ARBORESCENTES Vernonia acuminata Lessing, Linnaea 6: 663, 664. 1831. TYPE: Ind. Occ. [Jamaica], Swartz s.n. (holotype, s!). Vernonia divaricata de Candolle, Prodr. 5: 48, 49. 1836, p.p., non V. divaricata 366 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Swartz, 1806. Type: see Occ. [Jamaica], Swartz s.n. (holotype, G-pc (IDC B-800-2. 773: II. 5!)). Cacalia acuminata (Lessing) Vue Rev. Gen. Pl. 2: 969, 1891. Vernonia expansa Gleason, Bull. N. Y. Bot. Gard. 4: 186. 1906. TypE: Ja- maica., a Parish, near Troy, 6 Dec. 1940, Harris 8796 (holotype, NY!; isotypes, F!, us!). Shrub, 1-2 meters tall; stems pilose-hispid to sparsely hispid, ribbed. Leaves evenly spaced, at least along the main stem; petioles 2-5 mm. long, strigose to a ie hispid; blades 4-7.5 cm. long, 1.5-2.8 cm. wide, length/ width ratio ca. 2, widest at or slightly above the middle, elliptic to oc- casionally oval, gland dotted and sparsely pubescent to glabrate above, minutely pubescent and conspicuously gland dotted below, apically acumi- nate to acute, basally cuneate to rounded, margins entire. Inflorescences of condensed cymes, divaricate and often with elongated pedicels, indeter- minate; heads often 9 to 12 per cyme, 10- to 18-flowered, sessile or with strigose to tomentose peduncles to 1 mm. long; involucres campanulate, 4—5 mm. high, 4-7 mm. wide, the phyllaries pubescent, not tightly ap- pressed, golden brown to brown, the inner phyllaries lanceolate to rhom- boid, 3-5.4 mm. long, 0.7-1.3 mm. wide, tips acute, the outer phyllaries ovate, 0.9-1.6 mm. long, 0.3-0.7 mm. wide; pappus brown, the inner bristles 3-5 mm. long, the outer pappus of fimbriate scales, 0.6-1.3 mm. long; corollas 4.5-6.5(-—7.5) mm. long, white to pink, gland dotted in some, sometimes with a strong odor; anthers 1.5—-2 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed. Chromosome number x = 17. Flower- ing and fruiting throughout the year, but mainly between January and March. v pe 1. Distribution of oe acuminata a circles), V. divaricata (circles with stars in centers), V. harristi (squares V. pluvialis (triangles), V. rigida (stars), and V. verticillata (diamond) in Jamaica, and Oriente Province, Cuba. 1978] KEELEY, VERNONIA 367 DISTRIBUTION. Open, wooded limestone hills, along roadsides, and on bauxite-containing soils, between 150 and 1650 meters, in Jamaica (Map REPRESENTATIVE SPECIMENS. Jamaica. CLARENDON ParisH: Peckham Woods, Proctor 8253 (NY). MANCHESTER PaRIsH: Marshall’s Pen, Proctor 29396 (TEX). PorTLAND ParisH: Seamen’s Valley, Maxon & Killip 66 (F, GH, NY, US). ST. ANpREW ParisH: mile post 12-12% on Rock Hall Rd., Adams 7016 (Mo). ST. ANN ParisH: Reynolds Mine near Lydford P.O., Howard, Proctor, & Stern 14593 (GH). St. THomAs ParisH: slopes of Curva Curva Gap, Britton 4055 (NY). This species is one of the two most common vernonias in Jamaica; it can easily be distinguished from the other common species, Vernonia di- varicata, by its dark brown pappus and its glabrate leaves. Vernonia acu- minata is similar to V. Aarrisii and is probably very closely related to it. 2. Vernonia harrisii S. Moore, Jour. Bot. 66: 164. 1928. Type: Ja- maica. Clarendon Parish, Peckham Woods, 27 Dec. 1917, Harrts 12776 (holotype, BM!; isotypes, F!, GH!, K!, Mo!, Ny!,s!). Shrub, 2—3 meters high; stems ribbed, covered with dense down. Leaves evenly spaced along main stem, more crowded terminally; petioles 3-6 mm. long, downy; blades 3.2—5 cm. long, 1.4-2.7 cm. wide, length/width ratio ca. 2, widest at or below the middle, broadly ovate, shiny, gland dotted, with short, stiff, sparse hairs above, hairs more numerous and prominent on veins below, apex rounded to acute, base rounded to cuneate, margins entire. Inflorescences of scorpioid cymes, variably extended; cere 9- to 12-flowered, sessile or with pubescent peduncles to 2 mm. long; 1 volucres campanulate, 5-7 mm. high, 4-5 mm. wide, the ohyllaries resin dotted on tips and near bases, variably pubescent, loosely appressed, reflexed at maturity, gold-brown, the inner phyllaries awl shaped to lan- ceolate, much narrower at the tip than at the base, 4.1-6 mm. long, to 1.2 mm. wide, with tips tapering, acute to acuminate, the outer phyllaries ovate to deltoid, 1-1.4 mm. long, 0.5-0.8 mm. wide; pappus brown, the inner bristles 4—5.5 mm. long, the outer pappus of fimbriate scales, 0.6— 1.5 mm. long; corollas 5.5—7 mm. long, occasionally resin dotted near tips, lavender to purple, with no detectable odor; anthers 1.9-2.3 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed, some with resinous dots. Flow- ering and fruiting from December to March. DISTRIBUTION. Steeply eroded limestone pinnacles, amid dense vege- tation, from 500 to 800 meters, in two locations in Jamaica (Map 2). REPRESENTATIVE SPECIMENS. Jamaica. CLARENDON PARISH: wooded lime- stone hill, Peckham Woods, type locality, Proctor 9766 (GH, US). TRELAWNY ParisH: Mango Tree Hill, along road between Burnt Hill and Spring Garden, Proctor 31177 (MO, TEX). This species is similar in leaf shape to some specimens of Vernonia ri- gida, with which it has been confused in the past. Vernonia harrisit has 368 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 many fewer rows of phyllaries, possesses smaller heads, and lacks the conspicuous white leaf edges of V. rigida. Vernonia harrisii appears to be most closely related to V. see and may be a recent derivative of this more widespread species. 3. Vernonia verticillata Proctor ex Adams, Phytologia 21: 409. 1971. Type: Jamaica. St. James Parish, on wooded limestone hilltop, White Rock Hill, ca. 1 mi. S. of Sweetwater, small, spreading shrub, Howers bright purple, 2 Dec. 1962, Proctor 22987 (holotype, 1J!; isotype, UCWI! ). Shrub, erect or sprawling, 2.5-3 meters tall; stems ribbed, with oc- casional resin glands. Leaves whorled, at least the cauline, usually with three leaves per node; petioles 1-3 mm. long, with short hairs and fre- quently with resin dots; blades 3.8-7 cm. long, 2-3 cm. wide, length/ width ratio ca. 1.8, widest at or below the middle, ovate to elliptical, shiny, gland dotted, with few short hairs above and below, apex acute to acuminate, base rounded, margins entire. Inflorescences greatly com- pacted clusters of 3 to 10 heads, sometimes along a cyme; heads (4- to) S- to 9-flowered, sessile or with gland-dotted, hispid to hirsute peduncles to 2 mm. long; involucres narrowly campanulate, 5-7 mm. high, (3—)4- BH” oat Y 2 ee > ot : & i eo FEE on of Vernonia fasten = albicaulis (solid circles) Map 2. ributio ie subsp. ioe istylis Ce with s in centers), V. arborescens (squares), and V. pallescens Ga: ae Lower ‘eft inset, Cadac and Marie-Galante; lower right inset, St. Vin 1978 | KEELEY, VERNONIA 369 7 mm. wide, the phyllaries resin dotted at tips and bases, loosely ap- pressed, brown to gold, some with dark tips, the inner phyllaries lanceo- late, 4.5—6.2 mm. long, 0.8-1.4 mm. wide, tips recurving at maturity, in- ner surface resinous, edges occasionally fimbriate, the outer phyllaries deltoid, 0.6-1.1 mm. long, 0.3-0.8 mm. wide; pappus light brown, the in- ner bristles 4.5-5.5 mm. long, the outer pappus of fimbriate scales, 1—1.2 mm. long; corollas 6-7.2 mm. long, bright purple, some with resin dots at tips of petals; anthers 2.3-2.9 mm. long. Achenes ca. 2 mm. long, seri- ceous, ribbed. Chromosome number » = 17. Flowering and fruiting from December to March. DISTRIBUTION. Steep, eroded limestone hills in the Cockpit Country of Jamaica, between 300 and 900 meters, restricted to two locations (Map 1 REPRESENTATIVE SPECIMENS. Jamaica. St. JAMES ParisH: Glencoe district, 1 mi. SE. of Stonehenge, Proctor 24717 (GH); type locality, White Rock Hill, ca. 1 mi. S. of Sweetwater, Keeley & Keeley 1247-1267 (GA) This species has a distinctive habit, with an almost vinelike appearance and whorled leaves. Its restricted distribution and unusual morphology suggest that it has diverged from the main line of evolution of ARBORES- ceNtEs in the West Indies. Vernonia verticillata also has an unusual spineless pollen grain (Keeley & Jones, 1977a 4. Vernonia rigida (Swartz) Swartz, Fl. Ind. Occ. 3: 1322. 1806. — rigida Swartz, Prodr. 113. 1788. TyPe: ou Swartz s.n. (lecto- pe (here designated), s!; isotypes, BM!, s! Shrub, 1.5—2 meters high; stems ribbed and covered with downy, gold- brown tomentum. Leaves sparse except at tips of branches; petioles 4— 7 mm. long, with dense, golden brown pubescence; blades 3—5 cm. long, 2— 3.5 cm. wide, length/width ratio ca. 1.5, widest at or below the middle, ovate to round-deltoid, coriaceous, shiny, gland dotted and with occasional short hairs above, resinous and more pubescent below, apex blunt to ob- cordate, base rounded to truncate, margins densely white-tomentose, en- tire. Inflorescences of rigid, extended cymes; heads (13- to) 20- to 25- flowered, sessile; involucres turbinate to deeply campanulate, 10-15 mm. high, 6-7 mm. wide, the phyllaries shiny, resinous, and with scattered pu- bescence, tightly appressed, imbricate in many spiral series, golden brown, often with dark tips, the inner phyllaries lanceolate, 5.7—7 mm. long, 1.3— 1.5 mm. wide, with tips acuminate, recurved on some, the outer phyllaries deltoid, 0.9— 1.5 mm. long, 0.4-1 mm. wide; pappus brown, the inner bristles 4—6.5 mm. long, the outer pappus of fimbriate scales, 0.8—1.5 mm. long; corollas ca. 9 mm. long, bright rose-purple, glabrous or rarely resin dotted at base; anthers ca. 3 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed. Chromosome number » = 17. Flowering and fruiting from De- cember to March. 370 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 DisTRIBUTION. Eroded limestone cliffs and hillsides near the Cockpit Country, Jamaica, between 400 and 600 meters (Map 2). REPRESENTATIVE SPECIMEN. es CLARENDON PARISH: Quaco Rock, above Ritches, Keeley & Keeley 1186 ( This species, with its thick stems and large heads, is distinctive. It is included here in subsection ARBORESCENTES because of the characteristics of its flowers and achenes; however, it is possible that it may actually represent a separate subsection. The shape of the involucres and the number of phyllaries is unusual, but several Cuban species share these traits and are otherwise typical ARBORESCENTES taxa. Vernonia rigida also has an atypical spineless pollen grain; its nearest analogue is found in V. pineticola of Cuba (Keeley & Jones, 1977a). Swartz’s description of Conyza rigida refers to Plumier’s Plantarum Americanarum 4: t. 14, fig. 1. There is no illustration of a Conyza or even of a Composite in this work, however. Fournier (1932) reported that Plumier’s drawings, recopied and edited in this volume by J. Burmann, contained many errors and misplaced parts, and the protologue to which Swartz referred is no longer available. Since Swartz collected genuine Vernonia rigida in Jamaica and described it accurately, it seems reasonable to choose a lectotype from among his collections. 5. Vernonia segregata Gleason, Bull. Torrey Bot. Club 40: 327, 328. 1913. Type: Cuba. Oriente Prov., vicinity of Camp San Benito, 24 Feb. 1910, Shafer 4050 (holotype, Ny!) Shrub, ca. 1 meter high; stems closely pubescent, ribbed. Leaves even- ly distributed along main stems; petioles 1-2 mm. long, densely pubescent; blades 1.7-4.9 mm. long, 1-2.4 mm. wide, length/width ratio ca. 1.5, widest at or slightly above the middle, oblong to oblong-elliptic, resinous, punctate and with occasional hairs above, pubescent and with prominent veins below, the apex acute, more rarely acuminate, the base rounded to cuneate to cordate, the margins entire or remotely denticulate on upper third, occasionally revolute. Inflorescences of condensed divaricate cymes, appearing as capitate eee heads 5- to 7-flowered, sessile; involucres campanulate, 3.5-5 mm. high, 3.5-6 mm. wide, the phyllaries thinly cov- ered with inne hairs, Sea at the margins, loosely appressed, golden, the inner phyllaries linear to oblong, (3.5—)4—-4.8 mm. long, 1.3-2 mm. wide, with tips tapering, but sharp, the outer phyllaries triangular to lanceolate, 1.4-2 mm. long, 0.4-0.8 mm. wide; pappus white to pale straw colored, the inner bristles 4 mm. long, the outer pappus of fimbriate scales, ca. 1 mm. long; corollas 5-6 mm. long, white to purple, glabrous; anthers 2-3 mm. long. Achenes 1.5-—2.5 mm. long, sericeous to sparsely hirsute, ribbed. Flowering and fruiting throughout the year. DISTRIBUTION. Between 500 and 600 meters, in Oriente Province, Cuba (Map 4) 1978 | KEELEY, VERNONIA vel REPRESENTATIVE SPECIMENS. Cuba. vive Prov.: on the road from Navas to Sierra de Buena Vista, Ekman 3876 (s); Camp La Gloria, S. of Sierra Moa, Shafer 8216 (¥, NY); Taco Bay, on the ae between Santa Maria and Jiquani rivers, Ekman 3776 (GH, MO, NY). This species is very poorly represented in collections. It recalls Vernonia pluvialis of Jamaica, to which it may be related. Field work in Cuba with this species would be helpful in resolving relationships with other species, as well as in affirming the validity of this taxon. 6. Vernonia pluvialis Gleason, Bull. Torrey Bot. Club 40: 312. 1913. TYPE: Jamaica. St. Thomas Parish, Blue Mt. Peak, 14 May 1906, Shreve sn. (holotype, Ny! ). Vernonia proclivis Gleason, ibid. 312, 313. Type: Jamaica. St. Andrew Parish, vicinity of Cinchona, 2-10 Sept. 1906, Britton 102 (holotype, Ny!). es ee Gleason, ibid. 313, 314. Type: Jamaica. St. Andrew Parish, f Cinchona, 2-10 Sept. 1906, Britton 203 (holotype, NyY!). eee rere Gleason var. reducta Fawcett & Rendle, Fl. Jam. 72361 Shrub, 1-2 meters high; stems sparsely pubescent, ribbed. Leaves evenly spread along main stems, generally all of similar size, sessile or with strigose to villous petioles to 5 mm. long; blades 3.5—7.5 cm. long, 0.9- 5 cm. wide, length/width ratio ca. 3, widest at or just above the middle, elliptic to lanceolate, rugose, sparsely gland dotted and with scattered thin pubescence above, gland dotted and with prominent veins below, apex acute to acuminate, base cuneate, the margins entire to serrate, occasional- ly revolute. Inflorescences of compact divaricate cymes, often much con- densed on terminal branches; heads 5- to 7-flowered, sessile, variably pubescent; involucres campanulate, sometimes appearing constricted at the top, 4-6 mm. high, 3-4 mm. wide, the phyllaries tomentose, tightly ap- pressed, gold to brown, the inner phyllaries linear to oblong, sides often rolled inward creating a triangular phyllary, 4-5 mm. long, 0.8-1.4 mm wide, tips acute, the outer phyllaries elliptic, 0.9-1.5 mm. long, 0.3-0.7 mm. wide; pappus brown, the inner bristles 4-5.5 mm. long, the outer pappus of fimbriate scales, 0.5-1 mm. long; corollas 5.5—7 mm. long, pink to white, with resinous dots; anthers 1.5—2 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed. Chromosome number 2 = 17. Flowering and fruiting from October to July. DISTRIBUTION. Moist woods, along paths through rain forest areas and on mountain tops, in the Blue Mountains, Jamaica, between 1500 and 2500 meters (Map 1). REPRESENTATIVE SPECIMENS. Jamaica. PORTLAND ae upper E. ridge and summit, Mossman’s Pk., Maxon 9690 (GH, US). St. THOMAS ParRISH: Blue Mts., montane rainforest between Portland Gap and Blue Mt. Pk., Webster 5446 (cH, us); head of Mabess River, Blue Mts., Philipson 865 (f, NY) O72 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 This species is characteristic of the highest peaks in Jamaica and can be found in abundance along trails and in clearings. It is extremely poly- morphic, but collections from within a small area encompass the variations recognized as distinct species by Gleason (1906). It is low growing and often forms small thickets and hedges. 7. Vernonia albicaulis Persoon, Syn. Pl. 2: 404. 1807. Synonymies and typifications are given under the subspecific headings. Shrub, 0.5—2.5 meters high; stems generally woody, pilose to hispid, gland dotted, occasionally gray- or black-speckled, usually ribbed. Leaves evenly dispersed along main stems and branches, typically more crowded on terminal branchlets; petioles 1-11 mm. long, pubescent; blades 2-15 cm. long, 0.8—5.6 cm. wide, length/width ratio ca. 2, widest at or slightly above the midale oblong to elliptic, sparsely pilose-hispid to glabrate and gland dotted above, sparsely strigose below, the apex acute to obtuse, oc- casionally obcordate, the base cuneate to tapering, the margins entire to repand. Inflorescences cymose to rarely subcorymbose and much con- densed; heads (10- to) 11- to 20- (to 23-)flowered, sessile or with pedun- cles to 1 mm. long and pubescent as on leaf surfaces; involucre cam- panulate, 4-6 mm. high, 4-7 mm. wide, the phyllaries pilose to hispid, pu- bescence decreasing with age, occasionally gland dotted, imbricate in few series, firmly appressed until maturity, golden brown with brown tips, the inner phyllaries linear to elliptic, 3.5-5 mm. long, 0.6-1 mm. wide, with tips acute, sometimes recurved, the outer phyllaries awl shaped to deltoid, 0.8-1.3 mm. long, 0.3-0.8 mm. wide; pappus pale brown to rarely white, the inner bristles 4-6 mm. long, the outer pappus of fimbriate scales, 0.5— 1 mm. long: corollas 5-10 mm. long, lavender, pink, or white, glabrous; anthers 2—3.5 mm. long. Achenes ca. 3 mm. long, ribbed, hispid to seri- ceous, occasionally resin dotted. Chromosome number x = 17. Flowering and fruiting throughout the year. Two subspecies of Vernonia albicaulis are recognized. Their distribu- tions are shown in Map 2. They are characterized and distinguished by the following key: Pappus brown; styles 6-10 mm. long. ............... 7a. subsp. albicaulis. Pappus white or slightly dirty white; styles 12-25 mm. long. 7b. subsp. longistylis. 7a. Vernonia albicaulis Persoon subsp. albicaulis. Typr: St. Crucis [St. Croix], Vaki 737 (holotype, p-Ju (IDC B—6206-2. 621: I. 4!)). Conyza glabra Willdenow, Sp. Pl. 3: 1940, 1941. 1803. Type: America Meridionali meen America], Vahl s.n. (Herb. Willd. #15622) (lecto- ype (here designated) B, as photo B!). Eupatorium beer ie Willdenow. ibid. 1768. Type: America Calidiore 1978] KEELEY, VERNONIA O73 [Central America], it (Herb, Willd. #15144) (lectotype (here designated) B, as phot Vernonia longifolia pee, Syn, Pl. 2: 404. 1807. Type: St. Crucis (St. Croix), Vahl 737 (holotype, p-Ju (IDC B-6206-2. 621: I. 4!)). Lepidaploa albicaulis (Persoon) Cassini, Dict. Sci. Nat. 26: 17, 18. 1823. Ledidaploa lanceolata Cassini, ibid. 18. Type: Desfontaines s.n. (holotype, EI). ee Ee Swartz ex Wikstrom, Sv. Vet.-Akad. Handl. 1827: 72, 828 YPE: Guadeloupe, Forsstrom s.n. (lectotype (here a, i ais emarginata Wikstrom, ea 73. Type: Guadeloupe, Forsstrom sn. (lectotype (here designated), s Vernonia vahliana Lessing, ieee 4: 306-308. 1829. Type: America Meri- dionali [Central America ], sie ah (Herb. Willdenow #15622) (lectotype (here designated) B, as phot ; Eupatorium secundiflorum ne ex de Candolle, Prodr. 5: 48. 1836, p.p. TypPE: Guadeloupe, 1821, Balbis 44 (holotype, c-pc (IDC B-800-2. 773: 0 Ra) Vernonia arborescens 8 ovatifolia de Candolle, zbid., p.p. Type: Cuba. La Ha- vane, 1825, de la Ossa s.n. (holotype, G-DC (IDC B-800-2. 773: II. 3!)). Vernonia thomae Bentham, Vidensk. Medd. Naturh. For. Kjgbenhavn 5-7: 66. 1852. Type: St. Thomas. Hillside thicket, Water Island, 31 Jan.—4 Feb. 1913, Britton, Britton, & Shafer 142 (isoneotypes (here designated), Ny!, us!). No specimen cited as type or found at K or BM; it is presumed to have been lost or destroyed. Cacalia punctata (Swartz ex Wikstré6m) ee Rev. Gen. Pl. 2: 970. 1891. Cacalia thomae (Bentham) Kuntze, zbid. 1: Vernonia longifolia a genuina Urban. Sy ven 1: 456. 1899. Type: Guadeloupe. Basse Terre, Vieux Fort, 1893, Duss 2489 (isolectotypes (here designated), GH!, NY!). Vernonia denote B sintenisii a ibid. Type: Puerto Rico. Mayaguez aoe n declivibus montis i, inter filices, 8 July 1886, Sintenis 4371 (lec- tot — ee designated), V ernonia longifolia y vahliana en Urban, ibid. 456, 457. Type: America eae [Central America], Vahl s.n. (Herb. Willdenow #15622) (lec- e (here designated), B, as photo B!). Las sintenisii (Urban) Gleason, Bull. N. Y. Bot. Gard. 4: 187. 1906. Vernonia Gleasonii Ekman, Ark. Bot. 13: 54, 55. 1914. Type: Puerto Rico. ayaguez Prov., in declivibus onus Llano, 6 July 1886, Simtenis 4749 (isolectotypes (here designated), F!, GH!, MO!, NY Vernonia ee Gleason, Bull. ee Bot. Club 46: 238, 239. 1919. TYPE: Montserrat. Gages, below Soufriére, 20 Jan. 1907, Shafer 172 (holotype, NY!; oo cH! us!). DISTRIBUTION. Limestone cliffs and hillsides, 150 to 250 (to 600) meters, Puerto Rico and the Lesser Antilles, as shown in Map 2 REPRESENTATIVE SPECIMENS. Anguilla: N. of the Valley, Proctor 18543 (us). Antigua: Rose, Fitch, & Russell 3291 (¥, US). Dominica: ‘Sylvania,’ Cooper 88 F, GH, NY, US). Great Camanoe: near top of South Hill, D’Arcy 799 (cH). Guadeloupe: Vieux Habitantes, Questal 4529 (us). Montserrat: gorge of river leading from the Soufriére, Shafer 589 (Ny, US). Puerto Rico: Guayanilla to 374 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Tallahoa, rocky ee thicket, Shafer 1999 (¥, Ny, US). St. Barthélemy: Morne Grand Ford, Questel 112 (us). St. Croix: roadside near East End, Ricksecker 220 (F, GH, MO, NY, a Santa Cruz: Camby s.n. (NY). St. Kitts: banks of gulch r Old Road, Britton & Cowell 144 (Ny). St. John: vicinity of Caneel Bay, Woodworth 78 (Ff). St. Maarten: hills along northern side of Cul de Sac, Stoffers 2635 (cH). St. Thomas: Water Is., Britton, Britton, & Shafer 142 (¥, Ny, US). Virgin Gorda: summit & eastern slope of Virgin Pk., Smith 10566 (GH, NY, S, US). This is the most common species of coastal areas of Puerto Rico and the Lesser Antilles. It is often found on rocky cliffs and is visible at a distance due to the dense clusters of flowers. It is occasionally cultivated. There was some confusion in the Willdenow iene surrounding the types of Conyza glabra and Eupatorium obtusifolium (H. W. Lack, pers. comm.). Labels on the covers of the folders disagree in part with the de- scription of these two species. It is not clear which, if either, of these sheets represents the holotype of C. glabra or EF. obtusifolium. Sheet #15144, labeled Eupatorium obtusifolium, was chosen as the lectotype for this species; sheet 4£15622, labeled Conyza glabra, was selected for Ekman (1914) described a hybrid between Vernonia albicaulis and V. sericea and published it above a new name, V. gleasonii, in his revision. Since he did not place Vernonia gleasonti in a prominent position, but rather published the two species names first, with an * between them to indicate the hybrid nature of this “taxon,” I do not believe that he actual- ly wished to recognize a new species. I have placed the name in synonymy on this basis, and also because the specimens cited are well within the normal range of variation in Vernonia albicaulis and are not indisputably hybrid in origin. Vernonia albicaulis typically grows near the coast and at relatively low elevations in Puerto Rico, whereas V. sericea is found in the center of the island, above 300 meters; the hybrids between these species must be a great rarity if they occur at all in nature. 7b. Vernonia albicaulis subsp. longistylis Keeley, subsp. nov. 5, Corolla 9-10 mm. longa; antherae 3-3.5 mm. longae; stylus exertus 10- 20 mm. longus. Pappi setae albae, eis interioribus 5.5-6 mm. longis, eis externis I-1.2 mm. longis. Type: Guadeloupe. Basse Terre, limestone cliffs along roadside, facing inland, 2 km. W. of Pt. de la Saline, along hwy. to St. Ann, 20 Dec. 1975, Keeley & Keeley 1948a (holotype, Ga!) DIsTRIBUTION. This subspecies is known only from Guadeloupe and the small islands near it, Isles des Saintes and Marie-Galante. It is some- what taller and more robust than the typical subspecies and can easily be recognized by its white pappus and prominent styles. It is found on lime- stone and volcanic soils between 150 and 240 meters. REPRESENTATIVE SPECIMENS. Guadeloupe. BAsseE TERRE: Road to Vieux Fort 1978] KEELEY, VERNONIA on past Grande Anse beach, on steep volcanic banks, Keeley 1975 (GA). GRANDE TreRRE: vicinity of Anse Patate, Proctor 19926 (GH, US); vertical limestone cliffs, 1 km. W. of Fonds-Thézan, Keeley & Keeley 1993 (GA). 8. Vernonia sericea Richard, Actes Soc. Hist. Nat. Paris 1: 112. 1792. Type: Puerto Rico, Baudin 292 (lectotype (here designated), P-JU (IDC B-6206—2. 620: HI. 7!)) Lepidaploa phyllostachya Cassini, Dict. Sci. Nat. 26: 16, 17. 1823. TYPE: Puerto Rico, Baudin 292 (lectotype (here designated), p-Ju (IDC B-6206- PO 20e Ider ))s), Vernonia Berteriana de Candolle, Prodr. 5: 52. 1836. Type: Puerto Rico, 1820, Bertero & Balbis s.n. (holotype, c-pc (IDC B-800-2. 774: I. cane Conyza portoriccensis Bertero ex de Candolle, ibid. Type: Puerto Rico, 1820, Bertero & Balbis s.n. (holotype, c-pc (IDC B-800-2. 744: I. 3!)). Vernonia racemosa Delponte, Mem. Reale Accad. Sci. Torino, ser. 2. 14: 396— 398. 1854. Type: op. cit. pl. IT, p. 398, with analyses. Vernonia arborescens 8 Lessingiana Grisebach, Fl. Brit. W. Indian Is. 353. p.p. Type: Puerto Rico. Based on C. Plumier (ed. J. Burmann), Pl. 57 Vernonia a remotifiora Richard forma angustifolia Grisebach, Cat. Pl. Cubens. 14 66, p._p. Type: Cuba, 1860, Wright 2786 (holotype, K!; isotype, GH!). Vernonia stenophylla Gleason, Bull. N. Y. Bot. Gard. 4: 181. 1906. TYPE: Cuba, 1860, Wright 2786 (holotype, GH!; isotype, K!) Vernonia phyllostachya (Cassini) Gleason, ibid. 176, 177. riers venusta Gleason, ibid. 177. Type: Puerto Rico. Ponce Prov., on the s rd., 7 mi. from Ponce, 27 Nov. 1902, Heller 6136 (holotype, NY!; es i ‘us!). ? d / Vernonia araripensis sensu Gleason, ibid. 181, non V. araripensis Gardner, 1846 Vernonia a Wright ex Ekman, Ark. Bot. 13: 78, 80. 1914. TyPE: uba , Wright 2786 (holotype, K!: isotype, GH!). es sericea Richard subsp. racemosa ' (Delponte) Ekman, ibid. 85, 86. Vernonia racemosa sensu Gleason, N. Am. FI. 33: 64, 65. 1922, non V. race- nosa Delp d leans maestralis Ekman ex Urban, Repert. Sp. Nov. 26: 99. 1929. TYPE: Cuba. Oriente Prov., Sierra Maestra, La Bayamesa, on the ridge between Gro and Rio Yao, 5 May 1916, Ekman 7211 (holotype, s!; isotypes, F!, s!). Vernonia sericea Richard subsp. racemosa var. angustifolia (Delponte) Ek- man, Ark. Bot. 23A: 48, 49. 1930. Type: Haiti. Départ de Sud-Ouest, Mas- de la Selle, Pétionville, Chapi Faure, 27 Aug. 1924, Ekman 1653 (holo- type, s!; isotype, us! Shrub, 1—-1.5 meters high; stems pubescent, becoming glabrate with age, gray to brown, ribbed. Leaves often crowded at nodes, sometimes fascicu- late; petioles 1-5 mm. long, pubescent; blades 3.5-10 cm. long, 0.2—2. 5 cm. wide, length/width ratio variable, widest at or slightly below the mid- dle, alliptic. lanceolate to linear, pilose-hispid to strigose and with few glands above, sericeous and with brownish, protruding veins below, apex 376 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 a ee Ma Distribution of pee ee (triangles), V. sericea (circles), V~. stenophylla (squares), and V. fruticosa (stars) in Oriente Province, Cuba, Haiti and the Dominican eae oe Rico, and the Virgin Islands. acute to rarely acuminate, base tapering to cuneate, the margins entire, often revolute. Inflorescences of scorpioid cymes, variably extended; heads I1- to 15- (to 18-)flowered, sessile; involucres campanulate, 5—6 mm. high, 5-8 mm. wide, the phyllaries strigose to villous, pubescence thin- ning at maturity, loosely appressed, tan to golden colored, some with dark tips, the inner phyllaries subulate to lanceolate, 4.6—-6 mm. long, 0.8-1.1 mm. wide, tips acuminate to acute, the outer phyllaries lanceolate to el- liptic, 1.5—-2 mm. long, 0.3-0.5 mm. wide; pappus brown, the inner bristles (3.5—-)4-4.5 mm. long, the outer pappus of fimbriate scales, 0.3-0.7 mm. long; corollas (4.5—)5-6 mm. long, lavender to white, glabrous; anthers 0.8-1.5 mm. long. Achenes ca. 2 mm. long, sericeous, ribbed. Chromosome number x = 17. Flowering and fruiting throughout the year. DISTRIBUTION. Mountains, pine forests, on limestone and in moist areas, between 325 and 1600 meters, in Puerto Rico, the Virgin Islands, His- paniola, and in Oriente Province, Cuba (Map 3). PRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov.: Sierra de Nipe, near Woodired. Shafer 3552 (¥, NY, S). Dominican Republic. Dristriro NATIONAL: Llano Costero at San Carlos, Ekman 11393 (F, GH, NY, US). La VEGA PRov.: slopes along hwy. between eae and Valle Nuevo, Allard 17433 (NY, S, US). PEDERNALES Prov.: Agua Negra, Sierra del Bahoruco, Liogier 13955 (GH, NY). PUERTA PLATA PRoy.: El Choco, Jiménez 4374 (us). SAMANA Prov.: Peninsula de Samana, Boca Rio San Juan, Ekman 15006 (us). SANTIAGO PRov.: Pastor, Jiménez 1794 (GH, US). Haiti. DEPART DE L’ARTIBONITE: SE. of Ennery, Leonard 8990 (us). DEPART pU Norp-OUEST: Montagnes de Terre Neuve, Les Gonaiives Méme, Ekman 9464 (F, NY, TEX, US). DEPART pU Sup: prope Civette, Ekman 222 (us). Départ pu Sup-OUEsT: along rd. E. of Plateau Pistache, Massif de La Selle, Proctor 10768 (us). ILE DE La Torture: E. of La Vallée, Leonard & Leonard 11382 (us); Morne des Commissaires, Holdridge 1310 (GH, NY). Puerto Rico. AGUADILLA Prov.: Lares, Britton, Britton, é& Hess 2747 (Ny, US). SAN 1978] KEELEY, VERNONIA yi Juan Prov.: Rio Piedra, Bo. Guadalcanal, Distr. Bayamon, Otero 417 (GH, MO). Humacao Prov.: Luquillo, Bo. Pitajaya, Distr. Bayamon, Otero 519 (F, GH, MO). MayYAGuEz Prov.: 14 mi. NE. of Mayaguez, Heller 4472 (F, Ny, US). PONCE v.: Ponce to Pefiuelas, Britton, Britton, & Marble 1781 (Ff, NY). VIEQUES Is.: Cerro Ventana, Shafer 2867 (GH, NY). Virgin Islands. V1iRGIN GorRDA: thickets on mt., Fishlock 102 (Ny). St. THomas: Charlotte Amalie, Armour 551 (F, NY), TorToLa: hillside, high bush, Britton & Shafer 818 (Ny). ST. JAN [St. John]: Roseberg, Britton & Shafer 303 (Ny, Us). The variation in leaf size has been reflected in several infraspecific designations in past treatments. There is no justification for the main- tenance of these infraspecific taxa since both broad- and narrow-leaved forms occur in the same population and, on some occasions, on the same plant. This is an easily recognized species and can be found in forests and wet places in the mountains. The Cuban collections all appear to be narrow leaved, but collections are meager. 9. Vernonia borinquensis Urban, Symb. Antill. 3: 390, 391. 1903. TyPE: Puerto Rico. Arecibo Prov., Maricao in sylvis, 12 Nov. 1884, Sintenis 388 (isolectotypes (here designated), Ny!, s!). Vernonia hie ale var. stahlii Urban, ibid. 391. Type: Puerto Rico. San , Bayamon, Stahl 667 (lectotype (here designated), s!). Vernonia Sen ieaes var. resinosa Gleason, Bull. Torrey Bot. Club 46: 236. 1919, Type: Puerto Rico. Guayama Prov. Cayey, mt. slopes above the town, Jan. 1911, Holway s.n. (holotype, Ny!). Vernonia boringuensis var. hirsuta Gleason, ibid. Type: Puerto Rico. Maya- o de Maricao, wooded valley, 2 April 1913, Britton, Stevens, & Hess a ‘Gree ny!; isotype, Us!) Shrub, often sprawling, 0.5-3 meters tall; stems densely gray-white- pubescent, ribbed. Leaves evenly spaced along main stems and below heads; petioles 1-4 mm. long, densely pubescent; blades 2—4.7 cm. long, 0.7-2.5 cm. wide, length /width ratio ca. 2, widest at or just below the middle, ineeaiate to ovate, gland dotted, shiny, often rugose or pustulate- pubescent above, often similar below, but may be covered with velutinous tomentum, most pubescence along veins, the apex acute, tapering to ar- cuate, the base tapering, occasionally oblique, the margins entire or mi- nutely denticulate, often revolute. Inflorescences of conspicuous stiff, zig- zag cymes; flower heads large and showy, (13- to) 16- to 22-flowered, ses- sile; involucres campanulate, 5.5—8 mm. high, 7-11 mm. wide, the phyllaries densely pubescent to barely so, loosely appressed in mature heads, green to golden brown, the inner phyllaries awl shaped to lanceolate, (5. 2—)6- 7.5 mm. long, 0.8—-1.4 mm. wide, with tips long-tapering, acute, the outer phyllaries lanceolate, tapering, 1-2.4 mm. long, 0.4—0.7 mm. wide; pappus brown, the inner bristles 5-6.5 mm. long, the outer pappus of fimbriate scales, 0.5-1 mm. long; corollas 6-8 mm. long, white or lavender, glabrous, sometimes with resinous dots at tips of petals, no conspicuous odor; 378 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 anthers 2.5-3.3 mm. long. Achenes 1-2 mm. long, ribbed, sericeous. Chromosome number 7 = 17. Flowering and fruiting throughout the year. DISTRIBUTION. Steep, moist banks of clay, limestone, or serpentine, be- tween 300 and 1200 meters, in Puerto Rico (Map 3 REPRESENTATIVE SPECIMENS, Puerto Rico. AGUADILLA PRov.: Guavate State Forest, Liogier 10354 (Ny, US). AREcIBO Prov.: roadside bank, vicinity of Utua- do, Britton 5202 (Ny, US). GAYAMA Prov.: 2 km. W. of Aibonito, Howard & pel 15336 (GH). Mayacurez Proy.: in woods, Las Mesas, Liogier 10528 GH, NY, US). PONCE PRov.: moist bank, vicinity of Ala de la Piedra, above Villalba Britton & Earl 6091 (Ny, Us). This species can be easily recognized by its large heads and zigzag cymes. It is rather variable with respect to pubescence, accounting in part for the several varieties that have previously been recognized. Green- house-grown accessions show variable amounts of pubescence with age, and there seems little reason for maintaining subspecific status based on this trait. 10. Vernonia divaricata Swartz, Fl. Ind. Occ. 3: 1319. 1806. TypE: Jamaica, Swartz s.n. (holotype, s!; isotype, s! Vernonia arborescens Swartz y divaricata Grisebach, Fl. Brit. W. Indian Is. Vernonia acuminata sensu Grisebach, ibid., non V. acuminata Lessing, 1829. Vernonia gnaphalifolia sensu Grisebach, Cat. Pl. Cubens. 144. 1886, p.p., non V. gnaphaliifolia Richard, 1850. —— albicoma Gleason, Bull. N. Y. Bot. Gard. 4: 185, 186. 1906. TyPE: maica. West of Long Mt., 13 Jan. 1896, Campbell 6152 (holotype, Ny!). pe: mtonsa Gleason, ibid. 182. Type: Jamaica. Arnold Rd., 25 Nov. 1895, Campbell 6091 (holotype, Ny!). Cane sega Gleason, Bull. Torrey Bot. Club 40: 307, 308. 1913. TYPE Hanover Parish, “Greenland” (Green Island), 13 March 1908, Gea bo (holotype, nv!). Vernonia — Gleason, Bull. N. Y. Bot. Gard. 4: 181, 182. 1906. Type: Jamate St. Andrew Parish, Cinchona, May 1903, Shreve s.n. (holotype, nNyY!). Vernoma parvuliceps Ekman, Ark. Bot. 13: 71, 72. 1914. Type: Cuba. Orien- te Prov., Pinar de S. J. de Buenamito, 18 Nov. 1860-64, Wright 2788 (holo- type, K!: isotype, GH!). Shrub, 1-3 meters high; stems tomentose, ribbed. Leaves uncrowded, subbracteal leaves resembling the cauline, but smaller; petioles 4-8 mm long, tomentose; blades (3.8-)4-9 cm. long, 2-4.2 cm. wide, leneth/ width ratio ca. 2, widest at or below the middle, ovate to broadly lan- ceolate, rugose, occasionally bullate, shiny dark green surface, covered by numerous hairs above, densely tomentose, often velvety below, apex acute to acuminate, with tip slightly curved, base rounded to truncate, margins entire. Inflorescences of scorpioid cymes, subbracteal leaves re- duced or occasionally lacking; heads 12- to 20-flowered, sessile or with 1978] KEELEY, VERNONIA 379 hirsute to woolly-pubescent peduncles to 3 mm. long; involucres campanu- late, 4-6 mm. high, 4-6 mm. wide, the phyllaries hirsute to villous, loose- ly appressed at maturity, often with recurving tips, straw colored to greenish, tips often purple, the inner phyllaries oblong to near top, then sharply acute, 4-5 mm. long, 0.8-1.4 mm. wide, with tips acute, often re- curved, the outer phyllaries ovate, 0.8-1.6 mm. long, 0.3-0.7 mm. wide; pappus straw colored to clear white, the inner bristles 4-5 mm. long, the outer pappus of fimbriate scales, 0.5—-1 mm. long; corollas 5.5—7 mm. long, lavender, white, or rarely pink, occasionally resin dotted at tips of petals; anthers 1.5-2.2 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed. Chromosome number » = 17. Flowering and fruiting throughout the year, but most conspicuously from December to March. DisTRIBUTION. Moist slopes, in forests along roadsides and on lime- stone, serpentine, and clay soils, between 250 and 1800 meters, in Jamaica and Cuba (Map 1). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov.: Sierra Maestra, Loma Barbi, Ekman 15651 (s); Sierra de Nipe, Loma la Cintura, near Piloto del oa dio, Ekman 10135 (GH, s, US). Jamaica. HANOVER PARISH: Quashiba Mt., mi. W. of George’s Plain, Webster & Wilson 5099 (GH). MANCHESTER Pan. IsH: Martin’s Hill, Britton 3762 (Ny). PoRTLAND ParisH: Blue Mts., near Re- source, Perkins 1265 (GH). ST. ANDREW ParisH: Liguanea Ridge, Harris 12207 (F, GH, MO, US). St. ANN ParIsH: on roadside below Union Hill, 2.5 mi. SSW. of MMoneante: Howard & Proctor 13605 (GH). ST. CATHERINE PARISH: ca. 1 mi. SW. of Parks Rd., Webster & Wilson 5023 (cH). St. ELIZABETH PARISH: gravelly hillside, Great Pedro Bay, Britton 1259 (NY). ST. THoMAS PaRISH: bushy hillside near Whitfield, Proctor 6605 (Ny, US), TRELAWNY ParRIsH: Miss Laura’s Hill, Wilson Valley District, ca. 1 mi. N. of Warsop, Proctor 31180 (TEX). This species is the most commonly encountered Vernonia in Jamaica, but in Cuba it is known only from Oriente Province. Ekman (1914) de- scribed the Cuban population as Vernonia parvuliceps; however, there are no consistent differences to delimit the Cuban from the more abundant Ja- maican plants. Vernonia divaricata has often been confused with V. ar- borescens in the past, largely due to Linnaeus’s description that was based primarily on Plumier’s figure, but secondarily modified using a Jamaican plant collected by Browne. For a more detailed discussion of this prob- lem, see Ekman (1914). This species appears most similar to Vernonia arborescens, with which it has been confused, but is also similar to V. lepto- clada, V. sericea, and V. gnaphaliifolia. The extent of relationships of these species to each other remains to be investigated. 11. Vernonia arborescens (Linnaeus) Swartz, Fl. Ind. Occ. 3: 1320. 1806 Conyza arborescens Linnaeus, Syst. Nat. ed. 10. 2: 1213. 1759. Type: Mar- tinique. Based on C. Plumier (ed. J. Burmann), Pl. Amer. 122. ¢. 130, fig. Bef Se 380 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 Map 4, Distribution - Vernonia species in Cuba: a, V. complicata (square), V. desiliens (triangle | . divaricata (circle with star in center . gnapha- liifolia (solid circles) ; Ea commutata (square), V. hieracioides (circle with star in center), V. ae (solid circle), V. leptoclada (star), V. sagraeana (triangles); c, V. havanensis (squares), V. Meee (triang le), V. pineticola (star), V. purpurata (circle with star in center), V’. segregata (solid circle); d, 1978] KEELEY, VERNONIA 381 Lepidaploa arborescens Cassini ex Lessing, Linnaea 4: 302. 1829. Vernonia divaricata Lessing, ibid. 306. Type: Martinique, cee (Fl. Mar- tinique #190) (holotype, L!; isotype, L! Vernonia arborescens 8 ovatifolia de Candolle e, Prodr. 5: 48. 1836, p.p. TYPE: Cuba(?). La Habana, 1825, Ossa s.n. (holotype, c-pc (IDC B-800-2. 773: II. 4!)). Eupatorium secundiflorum Bertero ex de Candolle, zbid., pro. syn. Vernonia icosantha de Candolle, ibid. Type: Martinique, ee sn. (Fl. Mar- tinique #190) (holotype, L!; isotype. L!). Vernonia arborescens Swartz B Lessingiana Grisebach, Fl. Brit. W. Indian Is. 353. 1864. TYPE: aa Based on C. Plumier (ed. J. Burmann), PI. Amer. 122. t. 130, fig. 2.17 Cacalia arborescens aunsaeus) Kuntze, Rev. Gen. Pl. 1: 323, 324. 1891. Vernonia ventosa Gleason, Bull. N. Y. Bot. Gard. 4: 179. 1906. Type: Mar- tinique, “Duss 303, 304, 4069” (holotype, Ny!). Shrub, 1-2 meters high; stems velvety-pubescent, ribbed beneath pu- Beeccnse. and becoming glabrate with age, green to brown in color. Leaves generally spaced evenly, at least along the main stem, sometimes crowded on terminal branches and near heads; petioles 2-11 mm. long, pubescent; blades 4.5-11.3 cm. long, 1.3-4.5 cm. wide, length/width ae ca. 3, wid- est below the middle, ovate-elliptic, gland dotted, villous to strigose, pu- bescence lessening above with age and expansion of the blade, hirsute- villous to sparsely hirsute and gland dotted, with prominent protruding veins below, the apex acuminate, sometimes acute, the base cuneate to rounded, the margins entire to repand. Inflorescences of scorpioid cymes, variable in length, occasionally condensed; heads 15- to 25- (to 29- ) flow- ered, sessile; involucres campanulate, 4.5— 6 mm. high, 5-9 mm. wide, the phyllaries hirsute, villous to nearly glabrous, imbricate, loosely appressed, reflexed in maturity, golden brown, the inner phyllaries linear-lanceolate, 4-6 mm. long, (0.8—)1—-1.3 mm. wide, with tips acuminate, rarely mu- cronate, often with one strong central rib continuing into tip, the outer phyllaries subulate to linear, 1.3-2.5 mm. long, (0.3—)0.5-0.9 mm. wide; pappus white, the inner bristles 4-5 mm. long, the outer pappus of fim- briate scales, 0.7-1.2 mm. long; corollas 6-8 mm. long, lavender to white, resinous at tips of petals, rarely with faint sweet odor; anthers (2.2— )2.5- ong. Achenes 1.5—2 mm. long, sericeous, ribbed. ie number 2 = 17. Flowering and fruiting throughout the year, but chief- ly from December to May. DISTRIBUTION. Volcanic soils along roadsides or rarely in thickets, from 50 to 300 meters, in Martinique and its immediate satellite islands, and St. Lucia and St. Vincent (Map 2). REPRESENTATIVE SPECIMENS. Martinique: Hauteurs de Case-Pilote, Duss 984 (ny); Morne Vert region, Webster, Ellis, & Miller 9206 (¥, GH, US). St. Lucia: V. aronifolia (solid circle), V. menthaefolia (circles with stars in centers), V. ur- baniana (square), V. wright (star), V. yu nquensis (triangle). Vernonia viminalis is not included since no distribution is known other than in Cuba. 382 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 Anse-la-Raye, Beard 1099 (GH, MO, S, US). St. Vincent: Montrose Hills, Eggers 6545 (GH, US). Vernonia arborescens has been among the most confused and poorly understood of all West Indian species. It is typical of the ARBORES- CENTES subsection and is undoubtedly closely related to several of the other species in this subsection, particularly to V. divaricata. Ekman (1914) discussed the difficulties inherent in the correct application of the name V. arborescens. The present treatment is in accord with Ekman’s detailed history of the correct name for V. arborescens. The type of Vernonia arborescens 8 ovatifolia does not represent a Cuban distribution for this species, but rather a description of a plant collected probably in Martinique and grown in the Botanical Garden in Havana. De la Ossa was the director of the Botanical Garden and presumably cultivated plants from various parts of the West Indies. Several species are included in the c-pe collection under this trinomial. Only the de la Ossa collection represents V. arborescens. 12. Vernonia trinitatis Ekman, Ark. Bot. 13: 39. 1914. Type: Trini- ) dad, 1888, Hart 2036 (holotype, s!; isotype, K!). Shrub, 1-2 meters high; stems green, ribbed to terete, often with white to brown villous pubescence. Leaves large and not crowded along main stems; petioles 2—5(—10) mm. long, densely pubescent; blades 4.7-14.4 cm. long, 2.5—6.8 cm. wide, length/width ratio ca. 2, widest at or just be- low the middle, elliptic, shiny, pilose-hispid with distinct pubescence on veins above, hirsute to sericeous below, apex acuminate to aristate, base rounded to cuneate, the margins entire, repand, occasionally revolute, some remotely denticulate. Inflorescences of scorpioid cymes; heads (16- to) 19- to 25-flowered, sessile; involucres campanulate, 5-7 mm. high, 6— 10 mm. wide, the phyllaries glandular-pubescent, becoming less so with age, loosely appressed, green to brown, some with dark tips, the inner phyllaries lanceolate to elliptic, 5-7 mm. long, 1-2 mm. wide, with tips acuminate to aristate, generally less than 1 mm. long, the outer phyllaries lanceolate to subulate (1.6—)2.5-4 mm. long, 0.5—0.8(—1.1) mm. wide; pappus white, occasionally appearing dirty, the inner bristles 4.5—6(-8) mm. long, the outer pappus of fimbriate scales, (0.8—)1—-1.8 mm. long; corollas 5.5—7 mm. long, lavender to white, with occasional resin dots on tips of petals; anthers 2.5—3 mm. long. Achenes ca. 2 mm. long, sericeous, ribbed. Chromosome number x = 51,;. Flowering and fruiting from De- cember to March. DISTRIBUTION. Found only in northwestern Trinidad in and around Teteron Bay, on rocky cliffs in red soil near the sea, between 25 and 100 meters (Map 5) REPRESENTATIVE SPECIMENS. Trinidad: Teteron Bay, near Chaguaramas, on overgrown roadside banks along the closed section of Western Main Rd. to Scotland Bay, Keeley & Keeley 1933-1948 (GA). 1978] KEELEY, VERNONIA 383 Map 5. Distribution of Vernonia gee (triangles) and V. scorpioides Ges) in Trinidad and Tobago, and Hai This species is unusual in that it is the first polyploid Vernonia found in the West Indies, and is photosynthetic even in older stems. Ekman (1914) postulated that this was the ancestral species for the ARBORESCENTES. This seems unlikely since it is a polyploid and all of the other members of this subsection are diploid. Vernonia trinitatis exists in natural populations only with protection afforded it by its location within the military base at Teteron Bay. The coastal islands, which reportedly had this species at one time, are now cleared of nearly all vegetation. The species is pre- served in the yards of some local people for use in preparing cough medi- cines and a syrup used in treating colds. The common name is “man better man.” 13. Vernonia orbicularis Alain, Contr. Ocas. Mus. Colegio “De La Salle” 18: 15, 16. 1960. Type: Cuba. Oriente Prov., Pinares, Mina Johnson, Moa, 20 July 1947, Leon & Clemente 23210 (holotype, LS; isotype, NY!). Vernonia leonis Alain, ibid. 14. Type: Cuba. Oriente Prov., sur de la region de Baracoa, 17 July-4 Aug. 1924, Leon 12088 (holotype, Ls; isotype, NY!). Shrub, 0.7-1 meter high; stems densely pubescent. Leaves crowded near heads, cauline leaves evenly distributed; petioles 3-6 mm. long, densely pubescent: blades 1-3 cm. long, 1-2 cm. wide, length/width ratio ca. 1, 384 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 widest at or below the middle, broadly ovate to orbicular, dark green, shiny in some, sparsely pustulate: pubescent and with eel long trichomes ees light yellow-green and densely villous to villous-hirsute below, apex obtuse to blunt, base rounded, the margins entire, sometimes revolute. Inflorescences of scorpioid cymes often greatly compacted; heads 9- to 12-flowered, sessile; involucres campanulate, 3-6 mm. high, 4—6 mm. wide, the phyllaries hirsute to villous, loosely appressed at ma- turity, gold to brown with occasional purple tips, the inner phyllaries lanceolate, 3.5—5.5 mm. long, 0.5-1 mm. wide, with tips acute, sometimes recurved, the outer phyllaries lanceolate, 0.5-2 mm. long, wide; pappus straw colored to clear ae. the inner briciles 3-5 mm. ee the outer pappus of fimbriate scales, 0.9-1.6 mm. long; corollas 6-7.5 mm, long, purple to lavender, sparsely dotted with white trichomes; anthers 1.5—2.5 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed. Flowering and fruiting throughout the year. DISTRIBUTION. Serpentine barrens and pinelands, from sea level to 900 meters, in Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov.: Région de Moa, sur la charrascal serpentineux de la Playa de la Vaca, Victorin, Clemente, & Alain 21466 (NY, US); Guantanamo, Monte Libanon, San Fernandez, Ekman 10260 (F, NY). This species is easily recognized by its orbicular leaves that are pe- culiarly yellow-green beneath, and by the dense tomentum on the stems and leaves. It is similar to Vernonia leptoclada, however, and may be a variant. The leaf shape and inflorescence of Vernonia orbicularis recalls the Jamaican Permolles form of V. divaricata which was first described as a distinct species (Gleason, 1906) until field work proved this distinc- tion to be artificial. Observations of populations in the field would be very helpful here. 14. Vernonia leptoclada Schultz-Bipontinus, Jour. Bot. London 1: 233, 234, 1863. Type: Cuba. In Cuba orientali, 1860, Wright 1309 (holotype, P; isotypes, GH!, GoET!, Mo!) cea, wrightii sensu Grisebach, Cat. Pl. Cubens. 144. 1866, non V. wrightii Schultz-Bipontinus, 1863. Vernonia ss een sensu Gleason, ea N. Y. Bot. Gard. 4: 178. 1906, p.p.. cae Richard, 18 to e. alae Gleason, Bull. Torrey ne Club 40: 318, 319. 1913. Type: uba. Oriente ie Sabanilla to Yamuri Arriba, 30 Jan.-1 Feb. 1911, Shafer 8408 (holotype, NY!; isotypes, F!, cael Vernonia calophylla Gleason, bid. 317. Type: Cuba. Oriente Prov., Camp loria, Sierra Moa, 24-30 Dec. 1910, cam 8102 (holotype, ar iso- type, U is!), Vernonia semitalis Gleason, ibid. 319. Type: Cuba. Oriente Prov., trail Rio amaniguey to Camp Toa, 22-26 Feb. 1910, Shafer 4176 (holotype, Ny!)., Vernonia vicina Gleason, ‘bid. 317, 318. Type: Cuba. Oriente Prov., Camp 1978] KEELEY, VERNONIA 385 La Gloria, S$. of Sierra Moa, 24-30 Dec. 1910, Shafer 8202 (holotype, NY!). Vernonia neglecta Gleason, ibid. 318. Type: Cuba. In Cuba Orientali, 1859- 1860, Wright 1309 (holotype, GH!). Vernonia cristalensis Alain, Contr. Ocas. Mus. Colegio “De la Salle” 18: 13, 14. 19 PE: Cuba, Oriente Prov., Mayari on the crest of El Cristal, 2-7 April 1956, Alain, Acuna, & Figueras 5609 (holotype, Ls; isotype, Ny!). Vernonia acuiae Alain, ibid. 13-16. Type: Cuba. Oriente Prov., Pinares, Mina Johnson, 20 July 1947, Leon & Clemente 23182 (isotype, N¥!). Vernonia moaensis Alain, ibid. 14. Type: Cuba. Oriente Prov., Pinelands, Moa, Jan. 1943, Leon 21289 (holotype, LS; isotype, NY!). Shrub, ca. 2 meters high; stems densely pubescent, yellow to brown. Leaves not crowded along main stems, subbracteal leaves similar to cauline but smaller, highly variable; petioles 1-2.5 cm. long, densely tomentose, blades 2-4 cm. long, 1.5-3.5 cm. wide, length/width ratio ca. 2, widest at or below the middle, ovate to elliptic, occasionally lanceolate, rugose to smooth, occasionally pustulate, with sparse white trichomes above, dense- ly and closely glandular-pubescent, appearing grainy, with prominently pubescent veins below, the apex obtuse to acute, the base rounded to sub- cordate, rarely cuneate, the margins subcrenate to entire, often revolute. Inflorescence cymose, cymes variably scorpioid to relaxed divaricate; heads (11- to) 15- to 25-flowered, sessile; involucres campanulate, 5-8 mm. high, 6-8 mm. wide, the phyllaries often tomentose, loosely appressed at maturity, gold to brown, with prominent nerve and dark tips in some, the inner phyllaries linear to linear-lanceolate, 4—6.5 mm. long, 0.5-1.5 mm wide, with tips acute, occasionally recurved; the outer phyllaries awl shaped to ovate, 1.5—2 mm. long, 0.5-0.8 mm. wide, with prominent cen- tral nerve; pappus white to dirty white, the inner bristles 3-5 mm. long, the outer pappus of fimbriate scales, 1-1.5 mm. long; corollas 5.5—9 mm. long, lavender to purple, occasionally gland dotted; anthers 1.5-2 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed. Flowering and fruiting throughout the year. DISTRIBUTION. Serpentine barrens, pinelands, and areas of limestone soils, between 400 and 900 meters, in Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov.: Moa, plateau entre le Rio Cananas et le Rio Moa, Victorin & Clemente 21744 (us); Baracoa, on the path to Florida, Ekman 3504 (¥F, GH, NY, S$). There was apparently an interchange of Wright’s numbers 284 and 1309 at the Cosson-Durand Herbarium in Paris. Grisebach misidentified Ver- nonia leptoclada as V. Wrightii. This error was further elaborated upon by Gleason (1906) in his treatment of Vernonia gnaphaliifolia. Ekman (1914; pages 16, 17) gave a more detailed account of the circumstances of this interchange. This species, Vernonia leptoclada, constitutes one of the most variable assemblages of specimens yet encountered in the West Indies. It is found in an area of serpentine soils and appears to be part of a complex that includes Vernonia desiliens, V. orbicularis, and V. pineticola, These species 386 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 may all be endemic to this area, but they appear to hybridize to some de- gree as well and may represent only two rather polymorphic species. Ver- nonia pineticola is quite distinct, and its pollen is different from that of the others. Ekman 3714, 3715, 3543, and 6797 appear to represent vari- ous hybrids. Vernonia yunquensis occurs in this area, too, and may also be a hybrid; if this is so, the parental types are not clearly identifiable. Vernonia leptoclada and its allies in Oriente Province should definitely be reexamined in the field, since the treatment presented here is not definitive. 15. Vernonia jenssenii Ekman ex Urban, Repert. Sp. Nov. 26: 98, 99. 1929. Type: Cuba. Oriente Prov., in hills near Rio Sagua, 7 Nov. 1916, Ekman 8261 (holotype, s!; isotypes, BM!, F!, GH!, K!, MO!, NY!) Shrub, ca. 1 meter high; stems ribbed, with dense tomentum. Leaves evenly spaced along main stems, sessile or with densely tomentose petioles to 2 cm. long; blades 2.8-4.8 cm. long, 1.4-1.6 cm. wide, length/width ratio ca. 2, widest at or slightly below the middle, lanceolate to ovate, hir- sute, rugose and slightly bullate above, densely white-tomentose below, apex acute to acuminate, base rounded, margins entire. Inflorescences of loose, terminally branched scorpioid cymes; heads 21- to 25-flowered, sessile; involucres campanulate, 6-7 mm. high, 7-10 mm. wide, the phyl- laries recurved, loosely appressed, pale yellow-green, the inner phyllaries ovate to ovate-lanceolate, 5.5-6 mm. long, 1—-1.4 mm. wide, tips acute, the outer phyllaries awl shaped, recurved, 3.2—4.7 mm. long, 0.4—0.5 mm. wide, with tips 1-3 mm. long, sharply pointed; pappus pale yellow to white, the inner bristles 4-5 mm. long, the outer pappus of broad, fimbriate scales, 0.4—0.6 mm. long; corollas 6.3-6.5 mm. long, purple, glabrous; anthers 1.7 mm. long. Achenes 1-1.3 mm. long, sericeous, ribbed. Flowering and fruiting in November and December. DISTRIBUTION. Pastures and steep hillsides, Oriente Province, Cuba (Map 4) RESENTATIVE SPECIMEN. Cuba. ORIENTE Proy.: Sierra Maestra, Daiquiri, in pastures, Ekman 8340 (ny) Few specimens of this species were available for examination. It is very similar to Vernonia leptoclada except for the recurved phyllaries. It may, in fact, be but an odd variant of this latter species; for the moment it can be separated from Vernonia leptociada in the herbarium and will there- fore be maintained as a species; its status should, however, be reevaluated in light of field work when this is possible. 16. Vernonia urbaniana Ekman ex Urban, Repert. Sp. Nov. 26: 99. 1929. Typr: Cuba. Oriente Proy., Sierra de Nipe, in thickets around Rio Piloto, 10 Nov. 1919, Ekman 10083 (holotype, s!; isotype, US!). 1978 | KEELEY, VERNONIA 357 b, ca. 1 meter tall; stems densely pubescent, becoming glabrate with age, ribbed. Leaves sparse along main stem, smaller and more nu- merous on terminal branches; petioles 2-3 mm. long, variously pubescent: blades 4.2—6.8 cm. long, 0.7-2 cm. wide, length/width ratio ca. 5, widest at or slightly below the middle, lanceolate to oblong-elliptic, rugose, glan- dular-punctate and shiny above, glabrous below except on veins, which are densely pubescent and resin dotted, apex acuminate, base cuneate to rounded, the margins entire, often reqolute. Inflorescences of divaricate, scorpioid cymes; heads 15- to 20- (to 22-)flowered, sessile; involucres campanulate, 5-6 mm. high, 7-8.5 mm. wide, the phyllaries closely pu- bescent, firmly appressed, golden to brown, the inner phyllaries broadly oblong to rhomboid, fimbriate, 4.8-5.8 mm. long, 0.9-1.7 mm. wide, tips blunt to gradually acute, the outer phyllaries deltoid, 1-1.4 mm. long, 0.5— 0.7 mm. wide; pappus white to straw colored, inner bristles 4—4.8 mm. long, outer bristles 0.6-1 mm. long; corollas 7—-7.5 mm. long, lavender to purple, glabrous; anthers 1.8-2.5 mm. long. Achenes 1-3 mm. long, seri- ceous, ribbed. Flowering and fruiting throughout the year. DIsTRIBUTION. Steep hillsides and banks in Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Proy.: Sierra de Nipe, on a hillside at Rio Piloto, Ekman 5018 (¥, 8); Rio Piloto, Ekman 3338 (GH, NY, S, US). There are few specimens of Vernonia urbaniana extant, and its distribu- tion appears to be quite local. Since this species is sympatric with several other species in the ARBORESCENTES, it is possible that it is either a hybrid or a more variable series of collections that should be included under Vernonia leptoclada. This should be reexamined with field work in Cuba. The name Vernonia urbaniana was reported to have been published in 1909 by Glaziou (Bull. Soc. Bot. France, Mém. 3). He did not treat Compositae in this publication, however, and nowhere mentioned Vernonia. Apparently the publication of this name was in error as was the proper date for the third mémoire in this series, which is 1905 rather than 1909, In the series in the University of Georgia, Athens, library, the third mémoire is bound in volumes 52 and 53, not 56 as previously cited. 17. Vernonia desiliens Gleason, Bull. Torrey Bot. Club 40: 316. 1913. Type: Cuba. Oriente Prov., among rocks near water, Arroyo del Medio, above the falls, 20 Jan. 1910, Shafer 3232 (holotype, NY! ; isotypes, F!, GH!, Mo!, s!, us!). Shrub, 0.5-1.5 meters high; stems closely tomentose, sparsely so in places. Leaves decreasing in size near terminal heads, not crowded along main stem; petioles 2-5 mm. long, closely and sparsely tomentose; blades 2-8 cm. long, 1.5-3.5 cm. wide, length/width ratio ca. 3, but variable, widest below the middle, elliptic to lanceolate, dark green, sparsely pubes- cent above, light yellow-green and tomentose below, apex acute to oc- 388 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 casionally blunt, base rounded to cuneate, the margins entire to minutely crenate, revolute in some. Inflorescences of broad scorpioid cymes, some- times sharply angled at point of flower head attachment; heads 20- to 30- (to 35-) flowered, sessile; involucres campanulate, 9-11 mm. high, (8—)9- 11 mm. wide, the phyllaries sparsely pubescent to villous, resinous, loose- ly imbricated at maturity, dark gold-brown, often with darker tips, the inner phyllaries linear to ligulate, 7-9 mm. long, 1-1.5 mm. wide, with tips obtuse to acute, often recurved, the outer phyllaries deltoid, 1.5—2 mm. long, 0.6-1.2 mm. wide; pappus white, the inner bristles 5—6.5 mm. long, the outer pappus of fimbriate scales, 1-1.5 mm. long; corollas 7.5— 8.5 mm. long, purple to lavender, glabrous; anthers 2.5 mm. long. Achenes 1.5—3 mm. long, sericeous, ribbed. Flowering and fruiting throughout the year. DISTRIBUTION. Serpentine and limestone soils, pine barrens, and rocky areas, from 400 to 1200 meters, Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov. Sierra de Nipe: Loma Mensura, Ekman 3178 (¥, Ny, $s); headwaters of Brazo Dolores, Ekman 3121 (GH, MO, US) This species, Vernonia desiliens, is one of a group that occurs on ser- pentine soils in the mountains of Oriente Province. It has thick and rigid stems in comparison with the other species, and heads that appear heavy. It may hybridize with Vernonia orbicularis, V. leptoclada, and/or V. pine- ticola: field studies would be helpiul in understanding the appearance of individuals with intermediate characteristics 18. Vernonia pineticola Gleason, Bull. N. Y. Bot. Gard. 4: 176. 1906. yPE: Cuba. Oriente Prov., Baracoa, in pine woods, in clumps, March 1903, Underwood & Earle 1341 (holotype, Ny! ) Shrub, 1-2 meters high; stems densely pubescent, decreasing with age, ribbed, yellow-green. Leaves evenly spaced along main stems, more crowded on terminal branches; petioles 1-2 mm. long, densely pubescent; blades 1.5—4 cm. long, 1—2.5 cm. wide, length/width ratio ca. 1.5, widest below the middle, ovate to elliptic, rugose, pustulate with sparse pubescence, shiny above, woolly-tomentose and light yellow-green below, apex obtuse to blunt, base cuneate to subcordate, the margins entire, often revolute. Inflorescences of scorpioid cymes, often greatly extended: heads 15- to 20- flowered, sessile; involucres campanulate, 11—-12.5 mm. high, 10-13 mm. wide, the phyllaries stiff, glabrous to woolly-pubescent, loosely appressed at maturity, rose-purple, the inner phyllaries awl shaped, 9.5—-12.0 mm. long, 1-1.5 mm. wide, with tips acute, much tapered, the outer phyllaries awl eae 2.5-5 mm. long; pappus brown, the inner bristles 5-7 mm. long, 0.5-1 mm. wide, the outer pappus of fimbriate scales, 0.5-1 mm. long; corollas 7.5-10 mm. long, rose-purple, occasionally resin dotted; anthers 2.5-3 mm. long. Achenes 1.5—3 mm. long, sericeous, ribbed. Flowering and fruiting from March to December. 1978 | KEELEY, VERNONIA 389 DistRIBUTION. Pine barrens, serpentine and limestone soils, between 600 and 700 meters, in Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov.: Baracoa, pineland hill N. of El Yunque, Ekman 3542 (¥, MoO, NY, S, TEX); Guantanamo, Monte Liba- non, San Fernandez, Ekman 10268 (NY, S). The distinctive spikelike phyllaries and their bright purple coloration set this species o rom other sympatric taxa. ernonia pineticola appears to hybridize with Vernonia leptoclada and its allies, however. Field observations would be helpful in clarifying the relationships of V. pineticola and others found in nearby areas. Vernonia pineticola has unusual spineless pollen grains and represents an offshoot from the main line of ARBORESCENTES (Keeley & Jones, 1977a). 19. Vernonia yunquensis Gleason, Bull. N. Y. Bot. Gard. 4: 191. 1906. Type: Cuba. Oriente Prov., El Yunque, Mt. Baracoa, March 1903, Underwood & Earle 661 (holotype, Ny!; isotype, GH!). Shrub, 1-2 meters tall; stems densely tomentose to villous, light yel- low-green. Leaves crowded near terminal heads, less so along main stem; petioles 0.2-0.5 mm. long, densely pubescent; blades 3.5—-5 cm. long, 2— 2.5 cm. wide, length/width ratio ca. 2, widest at or slightly below the mid- dle, elliptic to ovate, dark green, pustulate, shiny, sparsely pubescent above, light yellow-green, tomentose to villous below, apex acute to obtuse, base rounded to cuneate, margins entire. Inflorescences of divaricate scorpioid cymes, often much condensed; heads ca. 25-flowered, sessile; involucres campanulate, 8-10 mm. high, 6-8 mm. wide, the phyllaries white-tomen- tose to villous, loosely appressed, brown to purple with darker tips, the inner phyllaries rhomboid to linear-lanceolate, 6-7 mm. long, 1.5—2 mm. wide, with tips cuspidate to acute, recurved, the outer phyllaries lanceolate to elliptic, 1-1.5 mm. long, 0.5-0.8 mm. wide; pappus straw colored to white, the inner bristles ca. 4 mm. long, the outer pappus of fimbriate scales, 1.5 mm. long; corollas ca. 9 mm. long, rose-purple, gland dotted, with few sparse hairs; anthers 2—2.5 mm. long. Achenes ca. 3 mm. long, sericeous, ribbed. Flowering and fruiting from December to March. DiIsTRIBUTION. E] Yunque, Mt. Baracoa, near top of peak, Oriente Province, Cuba (Map 4). EPRESENTATIVE SPECIMEN. Cuba. ORIENTE Prov.: Baracoa, top of El Yunque, Ekman 3962 (F, GH, NY). Because of its greatly condensed inflorescences that form a clustered group of nearly sessile heads at the tops of the branches, this species ap- pears distinct from others occurring in the area of Oriente Province. In addition, the heads are longer and the phyllaries more densely imbricate than in other species. There are very few collections of Vernonia yun- 390 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 quensis, which may actually represent a hybrid. This should be investi- gated when field work is possible. 20. Vernonia gnaphaliifolia Richard in Sagra, Hist. Fis. Pol. Nat. Cuba 2: 33, 34. 1850. Type: Cuba. Crescit prope Canasi, mense julio florens, Sagra s.n. (holotype, P!; isotype, F!) Vernonia arborescens (Linnaeus) Swartz var. divaricata Grisebach, Cat. PI. Cubens. 144. 1866. Type: Cuba. Pine woods, Los Remates, La Cipa, Dec. 1860-64, Wright 2787 (holotype, GoET!; isotypes, GH!, MO!). Vernonia membranacea Grisebach, ibid. 144, 145. Type: Cuba Occ. [Cuba], 1863, Wright s.n. (#490 was apparently added later to this label) (holo- type, GOET!; isotype, GH!). Vernonia crassinervia Wright ex Gleason, Bull. N. Y. Bot. Gard. 4: 180. 1906. Type: Cuba. Pine woods, Los Remates, La Cipa, Dec. 1860-64, Wright 2787 (holotype, GOET!; isotypes, GH!, MO!). aes sublanata Gleason, ibid. 177. TYPE: Cuba. ee ns Madruga, 28 N 1903, Britton, Britton, & Shafer 784 roma ae nes sublanata var. angustata Gleason, ibid. 177, 178. Type: Cuba. riente Prov., dry hill, Santiago, March cn Underwood & Earle 202 Vernonia angustata Gleason, Bull. Torrey Bot. Club 40: 309. 1913. Vernonia gnaphaliifolia Richard var. ea Nes Gleason, zbid. 46: 238. 1919. TYPE: Oriente Prov., Ensenada de Mora, 26-29 March 1912, Britton, & Shafer 12933 eae NY!). ie. platyphylla (Gleason) Ekman ex Urban, Repert. Sp. Nov. 26: 100. 1929, Vere nervosa Alain, Contr. Ocas. Mus. Colegio “De la Salle” 18: 150. 9 1960. Type: Cuba. Oriente Prov., “El Cerrito,’ an eruptive hill between Zarzal and hae 12 April 1915, Ebman 5378 (holotype, LS; isotypes, GH!, Ny!, us!) Shrub, 1-2 meters high; stems thinly pubescent to tomentose on young- est branches. Leaves sparse along main stems, more numerous on ter- minal branches; petioles 1-4 mm. long, pubescent: blades 2.9—9.2 cm. long, .8-2 cm. wide, length/width ratio ca. 3, widest at or just below the middle, oblong, ovate to linear-lanceolate, dark green, glandular-punctate, sparse- ly tomentose and rugose above, closely sericeous to tomentose below, veins prominent, apex acute to obtuse, base tapering to cuneate, the margins undulate or repand, often revolute. Inflorescences of branched cymes; heads 16- to 25- (to 33-)flowered, sessile or with peduncles to 2 mm. long; involucres campanulate, 6-7 mm. high, 7-12 mm. wide, the phyllaries ae to villous, resin dotted, loosely appressed, reflexed at maturity, golden to golden- brow, the inner phyllaries ovate to lanceolate, rarely rhomboid, 5—6.9 mm. long, 0.8-1.3 mm. wide, tips acuminate, the outer phyllaries awl shaped to subulate, 1.3-3.2 mm. long, 0.3-0.6 mm. wide; pappus white to slightly dirty white, inner bristles 4—6 mm. long, outer bristles 0.5-1 mm. long; corollas 6-7.5 mm. long, lavender to purple, resin dotted at tips of petals; anthers 1.5—-2 mm. long. Achenes ca. 3 mm. 1978] KEELEY, VERNONIA 391 long, with numerous short, stiff hairs, resinous, ribbed. Flowering and fruiting throughout the year. DisTRIBUTION, Serpentine soils, near roadsides in the mountains, be- tween 300 and 600 meters, in Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. Camaciey Prov.: Sierra Cubitas to Santa Rosa, Shafer 541 (ny). LA Hasana Prov.: Tetas de Camarioca, Matan- zas. Britton, Britton, & Wilson 14004 (F, GH, MO, NY, us). ORIENTE PROV.: Holguin, Cerro de Fraile, Ekman 3224 (F. GH, Mo, NY, $, US). PINAR DEL Rio Prov.: Bahia Honda, S. of town, Ekman 10421 (F, GH, NY, US). SANTA CLARA Prov.: Pozo de la Bermuda, W. of Manacas, Leon & Cazanas 5847 (Ny). ISLA pe Pinos: Lomas de Cafiada, Ekman 12133 (NY). This is a highly variable species found throughout Cuba. The nu- merous taxa that have been described previously and placed into synonymy with Vernonia gnaphaliifolia attest to its variability. Field collections would greatly improve our understanding of this polymorphic species. 21. Vernonia fruticosa (Linnaeus) Swartz, Fl. Ind. Occ. 3: 1323. 1806. Conyza fruticosa Linnaeus, Sp. Pl. ed. 2. 1209. 1763. Type: based on C. Plumier (ed. J. Burmann), Pl. Amer. 83, 84. ¢. 95, fig. 1s ist. Cacalia fruticosa (Linnaeus) Kuntze, Rev. Gen. PI. 2: 968. 1891. Vernonia buchii Urban, Repert. Sp. Nov. 16: 146. 1919. Type: Hait. Plaine Cul de Sac, March 1918, Buch 1506 (isotype, GH!). Shrub, 0.4—-1.2 meters high; stems densely tomentose to hirsute, ribbed, resinous, light gray to brown. Leaves evenly distributed along main stems, often sparse; petioles 0.1-0.5 mm. long, gray to brown, tomentose; blades 0.8-2.5 cm. long, 0.4-1.8 cm. wide, length/width ratio ca. 2, widest below the middle, ovate to parabolic, rugose, covered with pustulate pubescence, often velvety above, velvety tomentose and with few glands below, the apex blunt, obtuse to retuse, the base rounded to subcordate, the margins entire to crenate or undulate, sometimes revolute. Inflorescences of ex- tended cymes; heads 14- to 25-flowered, sessile or with densely tomentose to hirsute peduncles to 2 mm. long; involucres campanulate, 4.5—7 mm. high, 4-7 mm. wide, the phyllaries pubescent, spreading to reflexed at maturity, golden brown, the inner phyllaries lanceolate to rhomboid, 4.5- 6.5 mm. long, 1-1.5 mm. wide, with tips acute, often recurved, the outer phyllaries awl to needle shaped, 0.8-2.9 mm. long, 0.1-0.5 mm. wide; pappus white, the inner bristles 3.5-5 mm. long, the outer pappus of fimbriate scales, 0.3-1 mm. long; corollas 5-7 mm. long, white to laven- der, glabrous; anthers ca. 2 mm. long. Achenes ca. 3 mm. long, sericeous ribbed, Chromosome number » = 17. Flowering and fruiting throughout the year. DISTRIBUTION. Pinelands or open hillsides, on limestone, bauxite, and serpentine soils, between 1200 and 1300 meters, in the mountains of Haiti and the Dominican Republic (Map 4). 392 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 REPRESENTATIVE SPECIMENS. Dominican Republic. AzuA Proy.: Sierra de Ocoa, San José de Ocoa, Bejucal, Ekman 12047 (F, GH, NY, S). BARAHONA Prov.: Fuertes 655 (®, GH, MO, NY, S, US). LA VEGA PrRov.: serpentine barrens, Barrancon, from Banao to Hato Viejo, Liogier 14838 (Ny). Monre CRrist1 Prov.: Cordillera Central Moncién, Ekman 13011 (Ss, US, TEX). PEDERNALES Proy.: near El Pon on the way to Aceitillar, Sierra del Bahoruco near Peder- nales, Liogier 13791 (GH, NY). SANTIAGO PRov.: Guatara de Chino, Monte Ji- comé, Valeur 298 (s). Haiti. D&part pu Norp-Ouest: presqu‘ile du Nord- Ouest, montagnes de Terre Nueve, top of Morne La Fortune, Ekman 4979 (s, us). Départ pu Sup-Ouest: Port-au-Prince, Ekman 3039 (s). Vernonia fruticosa has a characteristic appearance and can be easily recognized by its sprawling habit and extremely rugose leaves. It is lo- cally abundant, but otherwise rare in the highest mountains of Hispaniola. Some sheets of Vernonia gnaphaliifolia, a Cuban species, bear a strong re- semblance to V. fruticosa, and relationships between these two taxa should be investigated when living material is obtainable. 22. Vernonia complicata Wright ex Grisebach, Cat. Pl. Cubens. 143. Type: Cuba, 1860-64, Wright 2790 (holotype, K!; iso- types, GH!, Mo!, NY!, s!). Shrub, 5-10 dm. high; stems closely tomentose. Leaves crowded, es- pecially on terminal branches; petioles 0.2-0.5 mm. long, closely tomen- tose; blades 0.5-1.5 cm. long, 0.6-1.5 cm. wide, length/width ratio ca. 1, widest at or above the middle, obovate, spatulate, or obcordate, softly tomentose and yellow-green above, softly tomentose and glandular, white to yellow-green below, the apex obtuse, emarginate to blunt, the base taper- ing to cuneate, the margins entire, typically very repand. Inflorescences ter- minal, usually solitary, rarely with two heads, in a divaricate cyme; heads 8- to 12-flowered, sessile; involucres campanulate, 4-5 mm. high, 4-5 mm. wide, the phyllaries softly tomentose to villous, firmly appressed until emp golden brown, the inner phyllaries awl shaped, tapering, (3—) 3.5—4.5 mm. long, 1-1.1 mm. wide, tips acute, the outer phyllaries broad- ly deltoid to awl shaped, pubescent, 1-2 mm. long, 0.3-0.5 mm. wide; pappus white to straw colored, the inner bristles (4—)5—6 mm. long, the outer pappus of fimbriate scales, 0.5-1 mm. long; corollas 6—9.6 mm. long, lavender to purple, glabrous; anthers 2—2.5 mm. long. Achenes 1.5—3 mm. long, ribbed, softly sericeous, hairs shiny. Flowering and fruiting from November to March DIsTRIBUTION. Cliffs facing the ocean, in Oriente Province, Cuba (Map REPRESENTATIVE SPECIMENS. Cuba. ORIENTE PRov.: Guantanamo, Caimanera W. of Emerald Pt. in wild cliffs facing the sea, Ekman 15765 (r, GH, MO, NY); coastal cliffs, Leeward Pt., Britton 2225 (cu, us). This species appears to be most closely related to Vernonia arbuscula of the Bahamas. Vernonia complicata can easily be recognized and dis- 1978] KEELEY, VERNONIA 393 tinguished from all other Caribbean vernonias by its low growth habit and its conspicuously repand leaves. 23. Vernonia arbuscula Lessing, Linnaea 6: 664. 1831. Synonymies, typifications, and representative specimens are given in Keeley and Jones (1977b). Shrub up to 3 meters tall; stems gray- to brown-tomentose, gland dotted, often much branched at least terminally. Leaves crowded, especially on terminal branches; petioles 6-8 mm. long, gray-brown-tomentose and glandular; blades 1.5-2(-4) cm. long, 0.6-1.5(—2.2) cm. wide, length/ width ratio ca. 1.5, widest above the middle, spatulate, obovate, obcordate, to lanceolate, surface gray-green-tomentose, gland dotted, upper surface frequently darker and greener in appearance than lower surface, which occasionally appears whitish, the apex acute to blunt, retuse, emarginate to obcordate, the base tapering, the margins entire to repand. Inflorescences much condensed on terminal cymes, usually with 3 to 5 heads; heads 8- to 12-flowered, sessile; involucres campanulate, 4-5 mm. high, 4-5 mm. wide, the phyllaries tomentose to villous, closely appressed when imma- ture, golden brown, the inner phyllaries awl shaped, 3.5-4 mm. long, 1- 1.5 mm. wide, tomentose to villous, with conspicuous resinous dots, with tips tapering, acute, the outer phyllaries broadly awl shaped, densely pu- bescent, size variable with age; pappus whitish to straw colored, the inner bristles 5-6 mm. long, the outer pappus of fimbriate scales, 0.5-1 mm. long, with jagged, often deeply cleft edges; corollas 6-9 mm. long, violet to rose-purple, with a smooth surface; anthers (1.5—)2.5-3.2 mm. long. Achenes 2—3.5 mm. long, with 5 to 10 ribs (5 larger or 10 smaller), seri- ceous to hirsute, hairs shiny. Chromosome number 2 = 17. Flowering and fruiting throughout the year. DISTRIBUTION. Limestone areas in pinewoods and clearings through- out the Bahamas. This species is highly variable, especially with regard to leaf shape. A full treatment of the differences is given in Keeley and Jones (1977b). 24. Vernonia stenophylla Lessing, Linnaea 6: 667, 668. 1831. TYPE: Dominican Republic. Barahona Prov., prope Las Salinas ad ri- pam fluminis, Oct. 1911, Puertes 1392 (isoneotypes (here desig- nated), F!, GH!, Mo!, Ny!, s!, us!). Vernonia remotiflora forma angustifolia Grisebach, Cat. Pl. Cubens. 144. Type: Cuba, 1860, Wright 2786 (holotype, ee a TEX!), Garda seenepagilla (Lessing) Kuntze, Rev. Gen. Pl. 2: 971. Vernonia corallophila Gleason, Bull. Torrey Bot. Club 40: ce ie TYPE: a. Oriente Prov., U. S. Naval Station, Guantanamo Bay, coral lime- stone cliffs, 17-30 March 1901, Britton 1939 (holotype, Ny!; isotypes, GH!, MO!, US!). Vernonia nematophylla Ekman & Urban, Repert. Sp. Nov. 26: 100, 101. 1929. Q = oa 394 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Type: Cuba. Oriente Prov., Sierra Maestra, near Alto Gracia, on the trail to Bacano, 4 Nov. 1916, Ekman 8189 (holotype, s!; isotype, F!, Ny!). Shrub, ca. 1 meter high; stems strigose to sericeous, white to gray, be- coming glabrate with age, barely ribbed to terete. Leaves evenly spaced, sometimes crowded at a node in a fascicular arrangement; petioles less than 1 mm. long, pubescent; blades 3-10 cm. long, 0.1-0.8 cm. wide, length/width ratio ca. 30, widest slightly below the middle, linear-lan- ceolate, shiny, with strigose hairs with persistent pustulate bases, hirsute above with main vein sunken, white sericeous below, the apex acute to mucronate, with main vein extending into point, the base cuneate to taper- ing, with vein widest at the base, frequently making up much of the basal width, the margins conspicuously revolute. Inflorescences racemose, heads on much-reduced cymes; heads (11- to) 17- to 22-flowered, sessile; in- volucres campanulate, 5-6 mm. high, 7-8 mm. wide, the phyllaries seri- ceous, loosely appressed, pale green to gold at maturity, both series of equal length, the inner phyllaries lanceolate to elliptic, 4.7-5.3 mm. long, 0.6-1 mm. wide, tips narrowly acute, the outer phyllaries lanceolate to linear-lanceolate, 2-4 mm. long, 0.4-0.6 mm. wide; pappus brown, inner bristles 4—5 mm. long, outer bristles 0.5-0.7 mm. long; corollas 5.5-6 mm. long, lavender to white, glabrous; anthers ca. 1.5 mm. long. Achenes ca. 2 mm. long, sericeous, occasionally glandular, obscurely ribbed. Flowering and fruiting throughout the year, but chiefly between September and De- cember. DISTRIBUTION. Dry limestone hills and banks, from 25 to 900 meters, in Cuba and central Hispaniola (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov.: U.S. Naval Station, Guantanamo Bay, Britton 1939 (F, GH, MO, NY, US). Dominican Republic. Ba- HORUCO Prov.: El Cercado, serpentine barrens S. of town, Liogier 12476 (GH NY, US). La VEGA PRov.: Cordillera Central, La Vega W. of Rio Yaque, Ekman 14162 (Ss, TEX). Monte Cristi Prov.: El Morro, Liogier 15630 (Ny, US). SAN- TIAGO Prov.: Pinar de Caimito, ca. 5 mi. NW. of San José de las Matas, Liogier 16277 (GH, NY). Haiti. DEPART DE L’ARTIBONITE: vicinity of Ennery, Leonard 9011 (NY, US). Depart pu Norp: vicinity of St. Michel de l’Atalaye, Leonard 7168 (GH, NY). DEPART DU Norp-OUEST: vicinity of Jean Rabel, Leonard & Leonard 12926 (MO, Us). DEPART pU SuUp-OuEstT: Port-au-Prince, Monfleurs, Ekman 1993 (¥, GH, NY, S) The loosely imbricate, subequal phyllaries distinguish Vernonia steno- phylla from other narrow-leaved taxa in Cuba and the Dominican Re- public. It is predominantly in Hispaniola and was previously found in Cuba at Guantanamo Bay, where it may now be extinct. 25. Vernonia commutata Ekman, Ark. Bot. 13: 77, 78. 1914. Type: Cuba. In Cuba Seat 1856- 57, Wright 286 (helene: K!; iso- types, F!, GH!, v!). Vernonia araripensis sensu ord Bull. N. Y. Bot. Gard. 4: . 1906, non V. araripensis Gardner, 1846. 1978] KEELEY, VERNONIA 395 Vernonia remotiflora forma angustifolia Grisebach, Cat. Pl. Cubens. 144. 66, p.p., non V. remotiflora Richard, 1792. Type: Cuba. In Cuba Orientali, 1856-57, Wright 286 (holotype, K!). Shrub, 1-1.5 meters tall; stems ribbed, often with appressed gray pube- scence. Leaves uncrowded at least along the main stem, subbracteate leaves similar to but smaller than the cauline; petioles 1-3 mm. long, finely strigose- or hispid-pubescent; blades 1-8 cm. long, 0.2-0.8 cm, wide, length/width ratio ca. 2, widest at or slightly below the middle, linear to linear-lanceolate, shiny, pustulate-pubescent to glabrate in older leaves, rugose and dark green above, glandular and resinous with pro- nounced pubescent veins below, apex acute to obtuse, base cuneate to tapering, the margins revolute, repand, entire. Inflorescences of scorpioid cymes; heads 17- to 21-flowered, sessile or with peduncles to 1 mm. long; involucres campanulate, 4-5 mm. high, 5-7 mm. wide, the phyllaries finely pubescent, becoming glabrate with age, loosely appressed, golden to light brown, often with purple tips, the inner phyllaries lanceolate to elliptic, 3.8-5 mm. long, 0.8-1 mm. wide, with tips acute to acuminate, often continued from central nerve, the outer phyllaries deltoid to lanceo- late, 1-1.5 mm. long, 0.3-0.5 mm. wide; pappus white, the inner bristles ca. 4 mm. long, the outer pappus of fimbriate scales, 0.5—0.7 mm. long; corollas 5.5—6 mm. long, lavender to purple, glabrous; anthers 1.2—-1.5 mm. long. Achenes ca. 2 mm. long, sericeous, faintly ribbed. Flowering and fruiting from November to March. DISTRIBUTION. Open pinelands and forests, from 400 to 800 meters, in Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba, ORIENTE Prov.: Bayamo, Corojo, El Pinar, Ekman 5066 (GH, MO, NY, S); pinelands vicinity of El Cuero, Britton & Cowell 12758 (F, NY, US). Vernonia commutata is very similar to Vernonia sericea and V. steno- phylla and almost appears to be a glabrate variety of V. sericea. It can, however, be separated from this species by the lack of leaf pubescence. Subsection II. BuXIFOLIAE 26. Vernonia barkeri Ekman ex Urban, Ark. Bot. 23A:; 49. 1931. TYPE: Haiti. Plaine Cul-de-Sac, Ganthier, riverbed of Riviére Blanche, 12 Dec. 1926, Ekman 7351 (holotype, s!; isotypes, F!, GH!, NY!, s!, us!). Shrub, 1-4 meters high; stems densely and closely pubescent, be- coming glabrate with age, gray to brown, ribbed. Leaves not crowded along main stems, most leaves on terminal branches; petioles 1-5 mm. long, with lanate to glandular tomentum; blades 1.4-3 cm. long, 0.91.6 cm. wide, length/width ratio ca. 1.5, widest at or above the middle, broadly spatulate, shining, with numerous sunken glands above, pilose to hispid, with many resinous glands below, apex rounded to acute, base 396 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 fap 6. Distribution of Vernonia barkeri (squares), V. buxifolia (circles), V. tuerckheimii (triangles), V. ekmanii (star), and V. sprengeliana (diamonds) in Haiti and the Dominican Republic. tapering, margins entire. Inflorescences in terminal cymes, often with clusters of 3 to 5 heads; heads 10- to 14-flowered, the peduncles 3-10 mm. long, covered with dense, glandular pubescence; involucres camp- anulate to turbinate, 3-5 mm. high, 4-6 mm. wide, the phyllaries various- ly pubescent, with long hairs prominent along margins, some glandular and sticky, tightly appressed, green to brown, the inner phyllaries linear to linear-lanceolate, 3—4.1 mm. long, 0.8—1.2 mm. wide, with tips usually acuminate, but occasionally cuspidate; the outer phyllaries deltoid, 0.5-0.8 mm. long, 0.4-0.6 mm. wide; pappus yellow to white, inner brist- les (4—)5—6 mm. long; corollas 6-7 mm. long, white to lavender, with resinous dots on petal tips; anthers 3-4 mm. long. Achenes 1.8—2.5 mm. long, glabrous, with resinous glands, ribbed. Flowering and fruiting throughout the year. DISTRIBUTION. Dry, rocky soils, 100 to 250 meters, in Haiti and the Dominican Republic (Map 6). REPRESENTATIVE SPECIMENS. Dominican a BARAHONA PROV.: prope Las Salinas et Ferrusco, Fuertes 1394 (us). LA VEGA Prov.: in thickets, El Numero, near Azua, Liogier r 16604 (GH, NY, Us). Haiti. DEPART DU SuUD-OQUEST: vicinity of Fond Parisien, Etang Saunas. Leonard 10142 (GH, US). Vernonia barkeri can easily be recognized by the lack of a distinct outer series of pappus bristles, as well as by some vegetative features. It is one of the few West Indian vernonias growing in an arid environment. 27. Vernonia buxifolia (Cassini) Lessing, Linnaea 4: 313. 1829. Lepidaploa buxifolia Cassini, Dict. Sci. Nat. 26: 18, 19. 1823. Type: Do- minican Republic, Desfontaines s.m. (holotype, FI). Vernonia dommieones de Candolle, Prodr. 5: 30. 1836. Type: Dominican Republic, 1821, Balbis s.n. nee c-pc (IDC B-800-2. 767: I. 4!)). Eupatorium domingense Sprengel ex de Candolle, ibid. Proustia domingensis Sprengel ex de Candolle, ibid. Cacalia buxifolia (Cassini) Kuntze, Rev. Gen. Pl. 2: 969. 1891. 1978] KEELEY, VERNONIA oo7 Vernonia montana Gleason, Bull. N. Y. Bot. Gard. 4: 191. 1906. Type: Haiti. Le Brande to Mt. Balance, 15 Aug. 1905, Nash & Taylor 1756 (holotype, NY!) Shrub, 2-5 meters high; stems finely gray- or brown-pubescent, resin dotted. Leaves crowded in young shoots, often at terminal ends of bran- ches: sessile or with resin-dotted, finely pubescent petioles to 0.5 mm. long; blades 0.5-2.5 cm. long, 0.5-1.8 cm. wide, length/width ratio ca. 1.5, widest at or above the middle, rarely wider below, broadly ovate to rhomboid to elliptic, shiny, glandular-punctate, occasionally with sparse pubescence above, finely tomentose, occasionally resinous, punctate be- low, the apex acute to mucronate, the base constricted, cuneate, the mar- gins entire or with 3 to 5 small teeth. Inflorescences subumbellate; heads 5- to 14-flowered, the peduncles 2-4 mm. long, resinous, sparsely pube- scent; involucres turbinate, 3-10 mm. high, 5-8 mm. wide; the phyllaries finely pubescent, resin dotted, loosely appressed at top, persistent at base, golden to brown, the inner phyllaries linear to narrowly awl shaped, 2—7 mm, long, 0.6—-1.3 mm. wide, tips acute, the outer phyllaries orbicular to roundly deltoid, 0.6-1.5 mm. long, 0.4-1.2 mm. wide; pappus yellow to orange, inner bristles 4-7 mm. long, outer bristles 0.50.7 mm. long; corollas 8-11 mm. long, white to purple, glabrous; anthers ca. 3 mm. long. Achenes ca. 4 mm. long, ribbed, resin dotted. Flowering and fruiting throughout the year. DISTRIBUTION. Limestone soils in wet areas of mountain forests and pine barrens, from 150 to 900 meters, throughout the Cordillera Central of Hispaniola (Map 6). REPRESENTATIVE SPECIMENS. Dominican Republic. BARAHONA PrRov.: La Hotte, between La Cueva and Placer Bonita, Howard 12257 (GH, US). LA VEGA PRov.: El Montazo, from Constanza to Valle Nuevo, Liogier 15432 (GH, NY, US). Monte Cristr Prov.: Moncion, Mera, Herb. Jiménez 1898 (Us). PUERTA Piata Prov.: Sierra de Yaroa, Liogier 13559 (GH). SAMANA PRov.: vicinity of Laguna, Samana Peninsula on the Pilar de Azucar, Abbott 2388 (Ny, Us). Haiti. Depart pe L’ARTIBONITE: Massif de Cahors, Petite Riviere de 1’Arti- bonite, Rd. to Medor at Sferlin, Ekman 3391 (s, us). Depart pU Norp: Bay- eux Morne Brigand, in naked limestone rocks, Ekman 2659 (s); vicinity of Dondon, Leonard 8594 (F, MO). DEpart pu Sup: Morne de la Hotte, in decliv. austral. Mont Ma Blanche, prope Dauyette, Ekman 466 (GH, S$, US). The large, turbinate heads and orange pappus of Vernonia buxifolia distinguish it from the closely related V. tuerckheimii. Both species were apparently common in the past, but lumbering and slash and burn agri- cultural practices have reduced their ranges; Vernonia buxifolia is now restricted to high-elevation areas not accessible by roa 28. Vernonia tuerckheimii Urban, Symb. Antill. 7: 421, 422. 1912. Type: Dominican Republic. La Vega Prov., prope Constanza, 1250 m. in declivibus pineti, Feb. 1910, Tuerckheim 2959 (holo- type, S; isotype, NY!). 398 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Vernonia microphylla Alain, Mem. N. Y. Bot. Gard. 21: 156. 1971. Type: ominican Republic. Bahoruco Prov., El Cercado, serpentine barrens on hillside S. of town, 4 Sept. 1968, Liogier 12477 (holotype, Ny!; isotypes, H!, us!). Vernonia ela Alain, Phytologia 25: 279. 1973. Type: Dominican Re- public. Pedernales Prov., Aceitillar, Cayo, Bahoruco Mts., 26 Feb. 1971, Liogier 17916 ( oe Ny!). Shrub, 1-2 meters high; stems ribbed, gray to brown, finely tomentose, occasionally resinous. Leaves sparse on main ae crowded on lateral branches and below Mao iemagoe petioles 1-2 mm. long, densely white- tomentose; blades 0.5—1.2 cm. long, 0.5—0.9 cm. wide, length/width ratio ca. 12, widest at or above the middle. obovate to broadly elliptic, shiny and glandular-punctate above, densely tomentose below, apex acute to mucronate, base tapering, the margins entire, occasionally revolute. In- florescences subumbellate to corymbose, with clusters of 5 to 10 heads; heads (7- to) 10- to 14-flowered, the peduncles 1-3 mm. long, gray- to brown-pubescent; involucres campanulate, 5.5—7.5 mm. high, 4—7.5 mm. wide, the phyllaries variously pubescent, tomentose on edges, frequently resinous, firmly appressed, gold to brown, frequently with purple tips and more rarely entirely purple, the inner phyllaries narrowly ovate to angu- lar ovate, (4.1—) 4.5-5.1 mm. long, 0.8-1.6 mm. wide, tips acute to blunt, the outer phyllaries ovate to orbicular, 0.6-1.2 mm. long, 0.6-1.2 mm. wide; pappus yellow to purple, inner bristles 5-6 mm. long, outer bristles 0.5-1 mm. long; corollas 6—8 mm. long, white or purple, glabrous, sweet scented; anthers ca. 3 mm. long. Achenes 3—3.5 mm. long, weakly ribbed, glandular. Chromosome number 7 = 17. Flowering and fruiting through- out the year. DistRIBUTION. Throughout the higher mountains of Hispaniola, on steep, rocky hillsides and banks, sterile pinelands, and openings in forests, from 600 to 1200 meters (Map 6). REPRESENTATIVE SPECIMENS, Dominican toes Azua Prov.: Sierra oc Ocoa, S.J. de Oca, Los Coroza, Ekman 11732 (F, TEX, US). BAHORUCO PRovy, El] Cercado, S. of town, Keeley & Keeley are, (GA). LA VEGA Prove: Cordillera Central Constanza, near El Chorro, Ekman 13966 (F, GH, NY, S, TEX, us). SANTIAGO PRov.: between La Guacara and La Guacarita rivers, La Guacara Arriba, tributary of Boa River, Liogier 13434 (NY) Vernonia tuerckheimii was first described by Urban in 1912, on the basis of a specimen collected by Tuerckheim near Constanza in the Do- minican Republic. This specimen had distinctly purple-tipped phyllaries and reddish purple pappus bristles. Ekman knew of this species and con- curred with Urban on its delimitation. In 1929 Ekman collected several sheets of similar-appearing plants; these, however, had neither the purple- tipped phyllaries nor the reddish pappus of the typical form. From notes attached to these sheets (e.g., Ekman 11732 S), it appears that he at first felt this was a new species and noted another name (V. ocana) on the label; this name was never published. Sometime later when he returned 1978 | KEELEY, VERNONIA 399 to Stockholm, Ekman must have discussed these specimens with Urban, as the specimens have annotations by both men, and the name V. ocana was replaced by V. tuerckheimii. Ekman made notes about this species on the back of several collection labels; for example, on the back of his number 11732, he says “P.S. V. tuerckheimii does not always have purple pappus and is moreover as variable as all Vernonias.”” From this informa- tion it seems fairly clear that Vernonia tuerckheimii is the same as, and takes precedence over, the more recently published name for the non- purple plant, Vernonia microphylla Alain, which is in every other respect identical to it. Subsection II]. PoLYANTHES 29. Vernonia havanensis de Candolle, Prodr. 5: 37. 1836. Type: Cuba. Juxta Havanam, Sagra s.n. (holotype, G-pc (IDC B-800-2. 769: Tit, 31) ); Vernonia cubensis Grisebach, Cat. Specs 144. 1866. Type: Cuba occ. [Cuba], 1863, Wright 2791 ce GOE Vernonia hieracoides 8 cubensis (Grisebach) an Anal. Soc. Esp. Hist. Nat. 19: 268. Vernonia ee Wright, Sauv. Anal. Cien. Havana 6: 176. 1894. TyPE: Cuba, Wright 3596 (holotype, GoeT!; isotypes, GH!, Ny!, Us!). Vernonia cubensis var. cajalbanensis Ekman ex tiban. Repert. Sp. Nov. 26: 101. Type: Cuba. Pinar del Rio Prov., Pinar del Cajalbana, on the very top of the mountain, 28 Aug. 1923, Bemaon 17344 (holotype, $!; iso- , NY!), Cuchi havanensis (de Candolle) Kuntze, Rev. Gen. Pl. 2: 970. 1899. Shrub, 1-2 meters high; stems ribbed, villous to sparsely pubescent, becoming glabrate with age. Leaves crowded along main stems; petioles 1-8 mm. long, pubescent as stems; blades 5—13 cm. long, 1—5.6 cm. wide, length/width ratio ca. 3, widest at or just above the middle, ovate to lan- ceolate, rarely obovate, shiny, with many glands and short pilose-hispid pubescence, scabrous above, gland dotted and with scattered white hairs below, the apex acute to acuminate, occasionally obtuse, the base rounded to cuneate, the margins entire to denticulate. Inflorescences highly branched cymes, often divaricate with few condensed heads; heads 11- to 24-flow- ered, the peduncles to 2 cm. long, pubescent; involucres campanulate, 4-7 mm. high, 5-8 mm. wide, the phyllaries pe occasionally with scat- tered hairs on the edges, loosely appressed, 1 or 6 series, golden to brown, typically with purple tips; the inner Beales linear to linear- lanceolate, rarely oblanceolate, 3-5 mm. long, 0.8—1.1(-1.4) mm. wide, tips acute to acuminate with strong central nerve extending into them, the outer phyllaries lanceolate to ovate, 0.8-1.3 mm. long, 1.4-1.8 mm. wide; pappus white to dirty white, inner bristles 4-6 mm. long, outer bristles 0.5-1.2 mm. long; corollas 6-8(-9) mm. long, purple to lavender, resin dotted along the tube; anthers 2.2-3 mm. long. Achenes ca. 2.5 mm. long, 400 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 ribbed, resin dotted, with short hairs prominent between the ribs. Flower- ing and fruiting throughout the year. DISTRIBUTION. Serpentine ridges, ravines, fields, and open places near rivers and in pine woods, from 300 to 800 meters, in various provinces of Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ee eee Proy.: thickets, Cumbre, Leon, Victorin, é& Alain 19726 (NY). ORIENTE Prov.: Baracoa region, Altas de Farola, 35 km. (airline) S. of Baracoa, Webster 4023 ce PINAR DEL Rio PRov.: Bahia Honda in “cuabales” S. of town, Ekman 10419 (F, MO, NY, US). SANTA CxaRA Prov.: dry field, vicinity of Sopapo, Buenos Aires, Trinidad Mts., L. B. Smith, Hodgdon, & Gonzales 3357 (GH, Us). ISLA DE PINos: vicinity of San Pedro, Britton, Wilson, & Selby 14467 (Ny, Us). There has been much confusion surrounding this species in the past. Ekman (1914) described Vernonia cubensis and V. havanensis, here syno- nymized, but apparently doubted this distinction, saying that “the three species of V. kavanensis are by no means very good ones. V. cubensis Griseb. in many respects is intermediate between V. havanensis DC. an V. hieracioides Griseb.”’ After examining the available material, I cannot find any consistent differences between the first two species, but I have maintained Vernonia hieracioides on the basis of flower number 30. Vernonia hieracioides Grisebach, Mem. Am. Acad. 8: 511. 1861. YPE: Cuba. Prope villam Monte Verde, dictam Cuba Orientali, Jan.—July 1859, Wright 1306 (holotype, K!; isotypes, F!, GH!, MO!, S§!, TEX!) Cacalia hieracioides edie Kuntze, Rev. Gen. Pl. 2: 970. 1891. Vernonia orientis Gleason, Bull. Torrey Bot. Club 40: 330. 1913. Tyee: Cuba. Oriente Prov., Sierra de Nipe, near Woodfred, 10 Jan. 1910, Shafer 3509 (holotype, Ny!; isotype, us!). Shrub, 1-3 meters high; stems ribbed, lanate to pilose-hispid, usually brown. Leaves clustered along stems, especially so below inflorescences; petioles 3-6 mm. long, cinereous-pubescent; blades 3—8.2 cm. long, 0.8—3 cm. wide, length/width ratio ca. 2, widest above the middle, obovate to elliptic-obovate, shiny, gland dotted, variably rugose and sometimes spar- ingly pilose-hispid above, sparsely pilose-hispid and glandular, with prominent veins below, apex acuminate to occasionally acute or obtuse, base rounded to cuneate, the margins denticulate, with teeth remote. In- florescences branched, compressed, corymbose, usually with elongate pedicels; heads 5- to 8-flowered, the peduncles 1-10 mm. long, cinereous; involucres campanulate, 4-6 mm. high, 3-6 mm. wide, the phyllaries sparsely strigose to glabrous, conspicuously resin dotted, loosely appressed in 4 to 5 series, gold to brown, often with dark tips, the i inner phyllaries oval to oblong-ovate, 2.9-4.2 mm. long, 0.6-1.8 mm. wide, with tips acute to acuminate, ca. 1 mm. long, the outer phyllaries lanceolate to ovate, 0.9-1.8 mm. long, 0.3-1 mm. wide; pappus yellow to brown, inner bristles 1978] KEELEY, VERNONIA 401 (3.5-)4-5 mm. long, outer bristles 0.3-0.5 mm. long; corollas (4.8—) 5-6 mm. long, purple to lavender, glabrous; anthers 2.2—2.8 mm. long. Achenes ca. 2 mm. long, ribbed, with short hairs between ribs, occasionally resin dotted. Flowering and fruiting throughout the year. DISTRIBUTION. Pine barrens, on serpentine soils, in wet woods, from 400 to 800 meters, in Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov.: 5 km. S. of Woodfred, Sierra de Nipe, Howard 6189 (GH, MO, NY, US); Sierra de Nipe, wet places along Rio Piedra in the pinelands, Ekman 1764 (F, 5 This species superficially resembles Vernonia havanensis Grisebach, from which it can be distinguished by its many fewer flowers per head (ca. 6 vs. 12 to 29 in V. havanensis) and by its fewer resinous dots on the achenes. In processing, the collections of Vernonia hieracioides and V. havanensis were apparently mixed. Wright 1306 (F, GH, K, MO, S, TEX) are V. hier- acioides, while those at s, us, and ny are V. hosiunensis 31. Vernonia menthaefolia (Poeppig) Lessing, Linnaea 4: 268. 1829. Eupatorium menthaefolium Poeppig ex Sprengel, Syst. Veg. ed. 16. 3: 412. 1826. Type: Cuba, 1823, Poeppig s.n. (holotype, Ww, as photo in GH!; iso- type, BM!). Eupatorium perrinianum Sprengel, ibid. Tyee: West Indies, Perrin s.n. (iso- types, GH!, NY!). Vernonia ottonis Schultz- Bipontinus, Linnaea 20: 508. 1847. Type: Cuba. Cafetal Fundador, 1839, Otto 35 (holotype, p!; isotype, GH!). oe grisebachii Schultz-Bipontinus, Jour. Bot. by 231,232, 1863. “TYPE: . Fruticose, ascending, fl. white, prope villam Monte Verde dictam, Cu ba Orientali, 13 Jan. 1859, Wright 1305 (holotype, GOET!; isotype, s!). Vernonia menthifolia var. grisebachii (Schultz-Bipontinus) Grisebach, Cat. Pl. Cubens. 144. 1866 Shrub, 1-2 meters tall; stems ribbed, with minute pubescence. Leaves regularly spaced along main stem, often crowded on lateral branches; petioles 10-15 mm. long, closely pubescent: blades 5—9 cm. long, 1.8—3.5 cm. wide, length/width ratio ca. 3, widest at or slightly below the middle, elliptic to elliptic-ovate, scabrellate, with dense, short hairs on young leaves, their persistent bases on older leaves above, glandular and closely pubescent in younger leaves, sparsely so with age below, apex acute to occasionally acuminate, base rounded, the margins entire to revolute, oc- casionally denticulate. Inflorescences corymbose, often dense; heads 12- to 20-flowered, the peduncles 3-5 mm. long, closely pubescent; involucres campanulate, 4-6 mm. high, 4-8 mm. wide, the phyllaries resinous, with long, floccose hairs along margins, closely appressed, gold to brown, oc- casionally with dark tips, the inner phyllaries oblong-lanceolate to oblan- ceolate, 3.44.6 mm. long, 0.6-1.1 mm. wide, with tips acute to acuminate, rarely obtuse, the outer phyllaries ovate, 0.8-1.1 mm. long, 0.4-0.8 mm. wide; pappus tawny with a yellow cast, inner bristles 4.4—5.5 mm. long, 402 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 59 outer bristles 0.5—0.6 mm. long; corollas 6.5—8.5 mm. long, purple to white, resinous, especially at tips of petals, fragrant; anthers ca. 2.5 mm. long. Achenes 2—3 mm. long, ribbed, with numerous short hairs between ribs. Flowering and fruiting throughout the year. DISTRIBUTION. In thickets, on the edges of savannas near the sea, in woods, between 150 and 1500 meters, La Habana, Santa Clara, Pinar del Rio, and Oriente provinces, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba, LA HaBaNna Proy.: Playa de Marianao, Britton & Wilson 4526 (NY). ORIENTE Prov.: Bayate, Sabana Miranda, edge of the savannah, Ekman 8568 (F, GH, MO, NY, S, US). PINAR DEL Rio Prov.: serpentine barrens, Cajalbana, La Palma, Alain 2375 (NY). SANTA CLARA PRov.: Colonia Limones, Ingenio Soledad, near Cienfuegos, Pringle 92 (F, GH, MO, NY). This species is similar in overall appearance to Vernonia havanensis, but can be distinguished from it by the shape of the inflorescence and the color of the pappus. The leaves are only rarely toothed in Vernonia men- thaefolia, not typically so as in V. Aavanensis. Hybrids may occur in the field, and therefore local population studies are needed. Subsection IV. SAGRAEANAE 32, Vernonia aronifolia Gleason, Bull. Torrey Bot. Club 40: 323, 324. 1919, Type: Cuba. Pinar del Rio Prov., limestone hills, vicinity of Sumidero, 2, 4 Aug. 1912, Shafer 13514 (holotype, Ny!; isotypes, F!, GH!, us!). ) Shrub, 1-2.5 meters high; stems closely pubescent to woolly on young twigs, ribbed. Leaves abundant along stem, often crowded beneath heads; petioles 2-5 mm. long, glabrous to sparsely pubescent; blades 5.2—7.5 cm. long, 3.1-3.7 cm. wide, length/width ratio ca. 1.7, widest at or just below the middle, elliptic to ovate, glabrous above, lower surface with short, stiff hairs on lamina, veins prominent, tomentose, apex acuminate, base rounded to tapering, margins remotely spinulose-denticulate. Inflorescences of scorpioid cymes, usually branched at least twice; heads 38- to 66- flowered, sessile; involucres campanulate, 8-10 mm. high, 10-12 mm. wide, the phyllaries closely pubescent, tightly appressed, gold to — the inner phyllaries subulate to lanceolate, 8-8.8 mm. long, 1.2-1.7 m wide, with tips up to 1 mm. long, stiff, recurved, the outer phyllaries ne shaped, 3.5-5 mm. long, 0.4—-1.1 mm. wide; pappus white to dirty white, the inner bristles 7-8 mm. long, the outer pappus of fimbriate scales, 1.5—2 mm. long; corollas 12 mm. long, white, glabrous; anthers 3.5 mm. long. Achenes 3-4 mm. long, resinous, angled but not ribbed. Flowering and fruiting from March to December. DISTRIBUTION. Limestone hills and steep hillsides, Pinar del Rio Pro- vince, Cuba (Map 4). REPRESENTATIVE SPECIMEN. Cuba. PINAR DEL Rio Prov.: Vifiales, among shrubs in Sierra de Vinales, Ekman 18030 (GH, Ny). 1978] KEELEY, VERNONIA 403 The large number of flowers per head, giving the heads the appearance of bursting, is distinctive in this species. It is apparently rare, being con- fined to rocky limestone areas. 33. Vernonia purpurata Gleason, Bull. Torrey Bot. Club 40: 322, 323. 1913. Type: Cuba. Oriente Prov., Sevilla Estate, near Santiago, Jiquarito Mts., Sierra Maestra, 18 Sept. 1906, Taylor 544 (holo- type, NY! Vernonia praestans Ekman & Urban, Repert. Sp. Nov. 26: 101. 1921. Type: uba. Oriente Prov., Sierra eres io oo 18 April 1915, Ekman 5508 (holotype, s!; isotype es, s!). Vernonia praestans var. cacuminis eee et Muniz, Acta Bot. Acad. Sci. Hungar. 18: 48. 1973. Type: Cuba. Oriente Prov., Sierra Maestra, in pluvi silvis nebulosis muscosis montis Pico Real, cacumen Mt. Pico Turquino, supra pag., 7 Dec. 1969, Borhidi, Ocujal Muniz, & Vasquez sn. (holotype, SV; isotype, BP!). Shrub, 2—2.5 meters tall; stems striate, thinly tomentose. Leaves often crowded; petioles 2-5 mm. long, closely tomentose; blades 3-6.5 cm long, 1-3 cm. wide, length/width ratio ca. 3, widest at or slightly above the middle, elliptic to ovate or oblong, shiny, with rigid hairs along main veins and more rarely on blade above, the veins raised, pubescent primari- ly along those below, the apex obtuse to acuminate, the base rounded, the margins entire, repand, occasionally revolute and with spinulose teeth. Inflorescences corymbose to cymose, usually much condensed; heads 4- to 8-flowered, the peduncles up to 3 mm. long, closely and densely pubescent; involucres narrowly campanulate, (3.5—-)4-7 mm. high, 2.5-5 mm. wide, the phyllaries villous, firmly appressed, gold to brown, the inner phyllaries linear-lanceolate, 6.4—-6.5 mm. long, ca. 1.5 mm. wide, with tips acute to acuminate, pubescent, the outer phyllaries ovate, 0.5-1.8 mm. long, 0.6— 8 mm. wide; pappus yellow-brown, inner bristles 6—7.5 mm. long, outer bristles ca. 1.2 mm. long. Achenes angled, glabrous. Flowering and fruit- ing from April to August. DIsTRIBUTION. Rocky ridges, from 1200 to 2000 meters, Oriente Pro- vince, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov. Sierra Maestra: on top of unta de Palmanoche, S. of Yara, Ekman 14304 (GH, MO, s); Pico Turquino, Ekman 5508 (F); Pico Turquino, S. slopes, Sezfriz 1069 (Ny). All specimens of this species, including the type, have very poorly de- veloped heads, none with flowers. The plants appear to resemble Vernonia pluvialis of Jamaica and apparently share a similar habitat; Vernonia pluvialis, however, has sericeous achenes. 34. Vernonia ekmanii Urban, Ark. Bot. 17: 62, 63. 1921. TypE: Haiti. Départ du Sud, prope Civette in collibus siccis, 11 June 1917, Ek- man 223 (holotype, s!; isotypes, F!, Ny!, US!). 404 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Shrub, 1—2 meters tall: stems ribbed, covered with dense, white, villous pubescence. Leaves evenly spaced along main stems, not scarce, usually sessile, occasionally with tomentose petioles to 7 mm. long; blades 6.5—10.2 cm. long, 1.8—4.4 cm. wide, length/width ratio ca. 3, widest at the middle, elliptic to linear-lanceolate, rugose, glandular and shiny above, dotted with short, stiff hairs and glands on lamina below, veins on lower surface prominent and villous, the apex acute to mucronate or cuspidate, with tip up to 1 mm. long, the base truncate to oblique, the margins spinulose- denticulate and crenate. Inflorescences of numerous scorpioid cymes, some appearing corymbose; heads 9- to 11-flowered, the peduncles up to 4 mm. long, densely villous; involucres campanulate, 8.5-10 mm. high, 5-7 mm. wide, the phyllaries densely tomentose especially on tips, loosely appressed, gold to brown, the inner phyllaries subulate to linear or linear- oblong; 8-9 mm. long, 1-1.8 mm. wide, sometimes pubescent on inner surface, with tips aristate, up to 1.5 mm. long, usually villous, recurved, the outer phyllaries narrowly deltoid to lanceolate, 1.5—2.3 mm. long, 0.4-1 mm. wide; pappus white, often appearing dirty, inner bristles 6—7.5 mm. long, outer bristles 1.5—2 mm. long; corollas 9-11 mm. long, purple, glabrous; anthers 3.5 mm. long. Achenes 3—4 mm. long, glabrous, with ca. 9 ribs. Flowering and fruiting from December to July. DIsTRIBUTION. In deep ravines, 700 to 1400 meters, Massif de la Hotte, Haiti (Map 6). REPRESENTATIVE SPECIMENS. Haiti. DEPART DU Sup: Massif de la Hotte, western group, Torbec, between La Mare Proux and Laurent, Ekman 5351 (GH, Ss): in decliv. occtlental. mont Ma Blanche, Ekman 493 (s). This species appears to be similar to Vernonia sprengeliana, which it greatly resembles in overall aspect. The pappus color is highly variable in V. ekmanii, a rarer species. 35. Vernonia sprengeliana Schultz-Bipontinus, Jour. Bot. London 1: 232, 233. 1863. TyprE: Dominican Republic. Sto. Domingo, 1866, Bertero 507 (holotype, P!). Shrub, 1.5-3 meters tall; stems ribbed, covered with a dense, grayish tomentum. Leaves crowded on terminal branches; petioles 1—2 cm. long, densely pubescent; blades 4—10.5 cm. long, 0.6-2.2 cm. wide, length/ width ratio ca. 5, widest at or slightly above the middle, lanceolate to linear-lanceolate, rugose to bullate, shiny above, densely tomentose below, apex acuminate to acute, base tapering to subcordate, the margins revo- lute, repand. Inflorescences of scorpioid cymes, sometimes appearing corymbose, often crowded on terminal branches: heads 9- to 16-flowered, the peduncles 1-5 mm. long, densely pubescent; involucres campanulate, 7-11 mm. high, 5-9 mm. wide, the phyllaries thinly strigose, firmly ap- pressed, gold to brown, often with purple tips, linear to oblong, 6.5—10 mm. long, 0.9-1.5 mm. wide, with tips rounded to mucronate, 1-3 mm. long, ie: oiltee phyllaries deltoid, 1-1.5 mm. long, 0.3-0.7 mm. wide; pappus 1978] KEELEY, VERNONIA 405 white or rarely purple tinged, the inner bristles 5-9 mm. long, the outer pappus of fimbriate scales, 0.5-1(-1.5) mm. long; corollas 8-12 mm. long, purple, glabrous; anthers 3.8-4 mm. long. Achenes ca. 4 mm. long, glabrous, angled or faintly ribbed. Chromosome number x = 17. Flower- ing and fruiting throughout the year. DISTRIBUTION. Rocky hillsides, near towns and disturbed sites, often near water, between 250 and 1250 meters, throughout Hispaniola (Map REPRESENTATIVE SPECIMENS. Dominican Republic. BARAHONA PRrov.: Las Salinas a Los Churos, Fuertes 1388 (F, GH, NY, US). LA VEGA PRov.: vicinity of Piedra Blanca, along Duarte Hwy. near dam, Allard 14753 (mo, US). Haiti. DEPART DE L’ARTIBONITE: Ennery, Ekman 2460 (NY, 8, TEX). DEPART DU NorD: vicinity of St. Michel de l’Atalaye, Leonard 7601 (mo, Us). Vernonia sprengeliana is among the most easily recognized species in the West Indies. From a distance it can be identified by the bright flower clusters in a subcorymbose arrangement. This species is closely allied to V. ekmanit. 36. Vernonia viminalis Gleason, Bull. N. Y. Bot. Gard. 4: 184. 1906. Type: Cuba, Wright 285 (holotype, Ny!; isotype, GH!). Shrub, ca. 1 meter high; stems with dense gray-white tomentum. Leaves evenly spaced along main stems; petioles up to 4 mm. long, densely vil- lous: blades 8-14 cm. long, 1-3 cm. wide, length/width ratio ca. 4, widest at or slightly below the middle, elliptic to narrowly oblong, shiny, with sparse pubescence of long, white hairs above, softly white-pilose beneath, with densest tomentum on veins of lower surface, apex acuminate to acute, base rounded to nearly cordate, margins minutely denticulate. Inflorescences of extended scorpioid cymes, subbracteal leaves well de- veloped; heads 23- to 30-flowered, sessile; involucres campanulate, 10 mm. high, 10 mm. wide, the phyllaries densely villous, firmly nance gold to brown, the inner phyllaries oblong to linear lanceolate, 8.5—-9.5 m. long, 0.9—1.2 mm. wide, with tips cuspidate, often recurved, the outer See deltoid to subulate, 1.8-2 mm. long, 0.4—0.5 mm. wade pappus white, inner bristles 6.5—-7.5 mm. long, outer bristles 0.8-1 mm. long; corollas 10 mm. long, purple, glabrous; anthers 3—3.5 mm. long. Achenes 3—4.5 mm. long, glabrous, angled. Flowering and fruiting dates not re- corded on collections. DISTRIBUTION. Limestone areas, Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMEN. Only the type specimen was available for measure- ment and scoring. Wright’s collection number 285 was apparently mixed up in processing and actually contains two species, Vernonia viminalis and V. sagraeana. The same numbers are, therefore, listed under both species; Wright 285 in 406 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 59 Ny and cu is V. viminalis, while this number in GoET, F, MO, TEX, and s represents V. sagracana. Vernonia viminalis has conspicuous long, white hairs on the leaves, while V. sagraeana is glabrous. It is possible that V. viminalis is a hybrid or simply an unusual variant of a more common species; more collections are needec 37. Vernonia wrightii Schultz-Bipontinus, Jour. Bot. London 1: 234. 1863. Type: Cuba. In prope Villam Monte Verde, dictam, Cuba Orientali, Jan—July 1859, Wright 284 (holotype, P; isotypes, GH!, K!). “ nonia leptoclada sensu Gleason, Bull. N. Y. Bot. Gard. 4: 183. 1906, non . leptoclada Schultz-Bipontinus, 1863. Shrub, 1-4 meters tall; stems densely tomentose. Leaves often crowded on main stems; petioles to 5 mm. long, densely pubescent; blades (2.1—) 4-8.4 cm. long, 0.7-3.6 cm. wide, length/width ratio ca. 3, widest at or below the middle, elliptic to linear-lanceolate, firm, shining and glabrous above, finely pubescent and resinous below, hairs prominent on veins of lower surface, apex obtuse to acute or acuminate, base rounded to sub- cordate, the margins entire, repand, occasionally spinulose-denticulate. Inflorescences of scorpioid cymes; heads (15- to) 17- to 26- (to 29-) flowered, sessile; involucres campanulate, 7.5-10 mm. high, 8-12 mm. wide, the phyllaries pubescent, becoming glabrate with age, firmly ap- pressed, gold to brown, the inner phyllaries linear to subulate, (6—) 7-8.5 mm. long, 0.9-1.6 mm. wide, with tips reflexed, subulate, to 5 mm. long, the outer phyllaries ovate to lanceolate, 1-3 mm. long, 0.4-1.1 mm. wide; pappus tawny brown, often with a pinkish cast, inner bristles 5—6(—7.5) mm. long, outer bristles ca. 1 mm. long; corollas 8-9 mm. long, purple to white, glabrous; anthers 2.5-3.5 mm. long. Achenes ca. 3 mm. long, re- motely angled, with a few ribs, rarely with a few short hairs. Flowering and fruiting throughout the year. DISTRIBUTION. Savannahs, pinelands, and mountainsides, between 300 and 900 meters, in Oriente Province, Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. ORIENTE Prov. Baracoa: pineland hill N. of El Yunque, Ekman 3541 (F, GH, NY, S); in carrascales on the path to Florida, Ekman 3500 (¥, GH, MO, US). The true identity of Vernonia wrightii has had a long and rather con- fused history due to the interchange of labels on Wright’s specimens num- bered 284 and 1309 (see Ekman, 1914, pp. 16, 17, 73, for a full discus- sion). Wright’s collections numbered 284 that are dated 1859 are V. Wrightii; those dated 1856-57 are V. sagraeana. 38. Vernonia sagraeana de Candolle, Prodr. 5: 55. 1836. Type: Cuba. Circa Havanam, Sagra 663 (holotype, c-pc (IDC B-800-2. 775 I. 2!)) 1978] KEELEY, VERNONIA 407 Vernonia valenzuelana Richard in Sagra, Hist. Fis. Pol. Cuba (Paris) 11: 33, 34. 1850. a ba. Crescit in Vuelta de Abajo, Valenzuela s.n. Golo: ty Pl; isotype, S Vernonia rubricaulis sensu Grisebach, Mem. Am. Acad. 8: 511. 1861, non V. rubricaulis Humboldt & Bonpland, 1810 Vernonia inaequiserrata Schultz-Bipontinus, Jour. Bot. London 1: 232. 1863. i uba, 1860, Wright 285 (holotype, Pp; isotypes, F!, GoET!, Mo!, NY!, TEX!), Vernonia rigida Swartz var. sagraeana (de Candolle) Grisebach, Cat. PI. Cubens. 144. 1866 Vernonia rigida Swartz var. valenzuelana (Richard) Grisebach, ibid. Vernonia inaequiserrata Schultz-Bipontinus var. obtusifolium Grisebach, ibid. Type: Cuba. In Cuba Orientali, 1861, Wright 285, pro parte (holotype, GOET!). Vernonia ee var. angustifolia Grisebach, zbid. Type: Cuba. In Cub ientali, 1861, Wright 2784 (holotype, GoreT!; isotypes, GH!, MO!, Nel ae aa Vernonia fallax Gleason, Bull. Torrey Bot. Club 40: 324, 325. 1913. TypeE: e: Santa Clara Prov., Trinidad Mts., Arroyo Grande to Trinidad, 12 March 1910, Britton & W ilson 5478 (holotype, N y!), Vernonia aceratoides Gleason, ibid. Type: Cuba. In Cuba Orientali, 1861, Wright 2784 (holotype, Mo!; isotypes, GH!, Ny!, TEX!). Vernonia angusticeps Ekman, Ark. Bot. 13: 14, 15. 1914. Type: Cuba, 1856- 57, Wright 284 (holotype, c!; ‘ee F!, GH!, mMo!). Vernonia linguaefolia Ekman, ibid. 19, 20, p.p. TYPE: a In Cuba Ori- entali, 1860-64, Wright 285 (holotype, i isot — noma sagraeana de Candolle var. an dustine bs pa Gleason, Bull. Tor- y Bot. Club 46: 240. 1919. We sins reedii Ekman & Urban, Repert. Sp. Nov. 26: 97. 1929. Type: Cuba. a Clara Prov., Lomas del Banao, El Purial on Rio Banao, in rocky places on top of “El Purial,” 27 Jan. 1923, Ekman 16227 (holotype, s!; isotypes, F!, GH!, Ny! Vernonia avrordons Ekman & Urban, ibid. 98. Type: Cuba. Santa Clara rov., in montibus Liguanea Trinidad in cacumine Pico Potrerillo, Mart. fruct., Ekman 18956 (holotype, s!; isotype, NY!). Shrub, 1-2 meters high; stems ribbed, closely and densely pubescent. Leaves evenly spaced along main stem, more crowded under inflorescen- ces, sessile or with pubescent petioles to 3 mm. long; blades 5—10.5 cm. long, 0.7-3.4 cm. wide, length/width ratio ca. 3, but variable, widest at the middle, elliptic, shiny, pubescent to glabrate above, closely tomen- tose and resinous below, pubescence decreasing with age on lower surface of blade but persistent on veins, apex obtuse to acute or acuminate, base rounded to cuneate, the margins entire to remotely denticulate, revolute to repand. Inflorescences of scorpioid cymes, often extended; heads (8- to) 11- to 25-flowered, sessile or rarely with densely ee peduncles to mm. long; involucres campanulate, 6-10 mm. high, 5-11 mm. wide, the phyllaries villous, most pubescence near tips, ea appressed, gold to brown, the inner phyllaries lanceolate to linear, 5.3—-7.9 mm. long, 0.9-1.3 mm, wide, with tips acuminate, rarely recurved, the outer phyllaries del- 408 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 toid, 1-2 mm. long, 0.4-1 mm. wide; pappus white, the inner bristles 5.5-7 mm. long, the outer pappus of fimbriate scales, 1-1.5 mm. long: corollas (8—)9—9.5 mm. long, purple or white, glabrous: anthers 3—3.5 mm. long. Achenes 3-4 mm. long, glabrous, angled when mature. Flowering and fruiting throughout the year. DistriBuTION. Hillsides and disturbed areas, 500 to 1200 meters, in much of Cuba (Map 4). REPRESENTATIVE SPECIMENS. Cuba. La HasaNna Prov.: Camp Florida, in “Cuabales” at Rio Quezada, Ekman 13576 (GH, NY, $). ORIENTE Prov.: Sierra Maestra, above Daiquiri, Ekman 8052 (F, GH, S). PINAR DEL Rio Prov.: Pinar de Cajalbana on E. slope of mt., Ekman 17361 (mM, s, Us). SANTA CLARA PRov.: Lomas del Banao, Loma del Obispo, Ekman 16280 (rR, 8). Vernonia sagraeana is a highly variable species that has been separated into entities many times on the questionable basis of leaf shape, but rarely on the more reliable basis of flower number. Leaf size and shape are highly variable in all vernonias, and are not generally good key characters at the species level. Subsection V. PALLESCENTES 39. Vernonia pallescens Gleason, Bull. N. Y. Bot. Gard. 4: 192, 193. 1906. Type: St. Vincent. Mountains at 2000 ft. and above: nowhere common, Smith & Smith 992 (holotype, N¥!; isotype, K!). Shrub, 1-2 meters tall; stems ribbed, with inconspicuous, finely hispid pubescence. Leaves large, well spaced, appearing to be almost. basal: petioles 5-10 mm. long, minutely pubescent to glabrate: blades 7.8—23.5 cm. long, 2.7-6 cm. wide, length/width ratio ca. 3, widest at the middle, elliptic, minutely puberulent and resin dotted above. shiny and often sparsely pubescent and resinous below, expanded and membranous, apex long-acuminate, base cuneate, the margins serrate, particularly on upper half of leaf. Inflorescences of extended scorpioid cymes, subbracteal leaves not present; heads (9- to) 10- to 13-flowered, sessile; involucres camp- anulate, (4.5-)5-6 mm. high, 4-6 mm. wide, the phyllaries minutely strigose, loosely appressed, gold to brown, frequently with dark tips, the inner phyllaries ovate to oblong, (3.7—)4—-5 mm. long, 0.8-1.2 mm. wide, tips acuminate to obtuse, the outer phyllaries lanceolate to ovate, 1-1.9 mm. long, 0.4—-0.6 mm. wide; pappus pale brown, inner bristles 4—4.5 mm. long, outer bristles 0.5-1 mm. long; corollas 5—6 mm. long, purple to white, resin dotted; anthers 1.5(—2) mm. long. Achenes ca. 2 mm. long, obscurely ribbed, sericeous. Flowering and fruiting from March to May. DIsTRIBUTION. Mt. Grand Bonhomme and the mountains near the Chateaubelair River, between 400 and 800 meters, St. Vincent (Map 2). REPRESENTATIVE SPECIMENS. St. Vincent. Valley of S. fork of the Cumberland River, Morton 5794 (vs): mts. above the Chateaubelair River, Morton 5276 (GH, Us); ESE. ridge of Mt. Grand Bonhomme, Proctor 26060 (TEX). 1978 | KEELEY, VERNONIA 409 This species is distinct from all other West Indian vernonias because of its serrate, membranous leaves. It also has a distinctive pollen type, inter- mediate between the other major pollen types in the West Indies (Keeley & Jones, 1977a). Subsection VI. ScoRPIOIDES 40. Vernonia scorpioides (Lamarck) Persoon, Syn. Pl. 2: 404. 1807. Conyza scorpioides Lamarck, Encycl. Méthod. 2: 88. 1786. TYPE: Brazil, Commerson sn. (holotype, p-JuU (IDC B-6206-2. 621: I. 3!)) ley subrepanda Persoon, ibid. Type: Brazil, Commerson s.n. (holotype, u (IDC B-6206-2. 621: I. 3!)). Vernonia tournefortioides HBK. Nov. Gen. Sp. (quarto ed.), 34, 35. 1818. Type: Venezuela. Caracas, Humboldt s.n. (holotype, B, as photo B!). ne scorpioides Cassini, Dict. Sci. Nat. 2: 16. 1823 Staehelina solidaginoides Willdenow ex Lessing, Linnaea 4: 281, 282. 1829. Type: Venezuela. Caracas, Humboldt s.n. (holotype, B, as photo B!). Vernonia shea Lessing, ibid. 6: 657. 1831. Type: Brazil, eee sm. (lectotype (here designated), p-yu (IDC B-6206-2. 621: Vernonia scorpioides Persoon a centriflora de Candolle, Prodr. aN . 1836. yee: Brazil. Bahia, April 1831, M. d’Hostky sn. (lectotype ee des- me c-pc (IDC B-800-2. 771: I. 6!) Vernonia scorpioides B subrepanda (Persoon) de Candolle, ibid. Vernonia scorpioides (Persoon) y subtomentosa de Candolle, ibid. TYPE: Brazil, 1834, M. Lund 479 (lectotype (here designated), c-pc (IDC B-800- eat ie eae Os ee ee (Persoon) 6 longifolia de ee ibid. Type: Brazil. 830, M. Salzman s.n. (holotype, c-pc (IDC B-800-2. 771: III. it): ee ear e longeracemosa de Candolle, ibid. Type: Brazil, 1832, Roebere 33 (1203) (lectotype (here designated), c-pc (IDC B-800-2. 771: III. 3 sae longeracemosa Martius ex de Candolle, Prodr. 5: 42. 1836. Vernonia lanuginosa Gardner, Jour. Bot. London 5: 219. 1846. Type: Brazil. Minas Gerais Prov., in fruticetis prope Formigas in Sertao, July 1840, Gardner 4764 (holotype, BM!; isotype, NyY!). Vernonia saepium Ekman, Ark. Bot. 17: 63. 1922. Type: Haiti. Départ du Sud, Morne de la Hotte ad Ma Blanche, prope Dayette, 7 Aug. 1917, Ek- man 463 (holotype, s!; isotypes, F!, cH!, ny!, s!, us!). Cacalia tournefortioides (HBK.) Kuntze, Rev. Gen. Pl. 2: 971. 1891. Cacalia scorpioides (Lamarck) Kuntze, ibid. Shrub, upright or sprawling, 1-3 meters high; stems densely pubescent, strigose to villous, ribbed. Leaves numerous, but not crowded along main stems; petioles 4-7(-10) mm. long, densely gold-pubescent; blades (1.8- _ 3-9.5 cm. long, (1.2—)2.2-6 cm. wide, length/width ratio ca. 1.5, widest at or below the middle, ovate to elliptic, some very broad, strigose to pilose-hispid and shiny above, hispid to hirsute below, tomentum den- sest in young leaves, the apex obtuse to acute or acuminate, the base cune- ate, tapering, the margins entire to rarely denticulate. Inflorescences of 410 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 scorpioid cymes, heads very close together; heads (14- to) 16- to 22- flowered, sessile or with woolly-pubescent peduncles to 1 mm. long; in- volucres campanulate, 3.5—4.5(—6) mm. high, 4-7 mm. wide, the phyllaries firmly appressed, with long hairs at margins, gold to green, the inner phyllaries often purple Bes and pubescent, linear-lanceolate with curled tips, 3.8-4.5 mm. long, 0.6-1.2 mm. wide, the curled tips 0.2-0.5 mm. long, the outer meee lanceolate with a much-constricted base, 1-2 mm. long, 0.5—-0.8 mm. wide; pappus white, the inner bristles 4.5-6 mm. long, the outer pappus of fimbriate scales, (0.8—)1—1.6 mm. long; corollas 6—7(—8) mm. long, purple to white, outer tips of petals with conspicuous white hairs often projecting into the clefts between the corolla lobes; anthers 1.5-2 mm. long. Achenes 1.5 mm. long, with sparse, stiff hairs between ribs. Flowering and fruiting from March to August. DISTRIBUTION, Open places, along disturbed roadsides, and near the ocean and towns, in Trinidad and Tobago and Haiti (Map 5). REPRESENTATIVE SPECIMENS. Haiti. DEPART pU Sup: Massif de la Hotte. Torbec, Forman, Ekman 7587a (¥). Tobago: near Little Bacolet Bay, Howard 10459 (GH, NY). Trinidad: North Post Rd., Britton, Hazen, & Mendelson 776 (GH), This species has a very characteristic aspect both in the field and in the herbarium, with its dense scorpioid cymes and broad leaves that are often folded. The corollas are also conspicuous because they are hairy, a feature unknown in other West Indian vernonias. The range of this species ex- tends throughout much of South America. — Subsection VIT. PANICULATAE 41. Vernonia blodgettii Small, Flora S.E. U.S. 1160. 1903. Synony- mies, typifications, and representative specimens given in Keeley and Jones (1977b). Herbaceous perennial, 1-1.5 meters high; stems erect, glabrous, often branched at base. Leaves mostly basal, sessile; blades 1.8-6.9 cm. long, 0.1-1 cm. wide, linear or nearly so, glabrous above, lightly gland dot- ted below, apex obtuse to acute, base attenuate, the margins slightly revolute, entire. Inflorescences loose, irregular, with few heads: heads about 21-flowered; involucres loosely and irregularly imbricated, cam- panulate, 5—8.5 mm. high, 5.5-10.5 mm. wide, the phyllaries firmly ap- oressed, ae Itoid to lanceolate, the inner phyllaries 3.9-6.7 mm. long, the outer phyllaries 1.7-3.5 mm. long, purple, glabrous to slightly pubescent, with tips acute to subacute, 0.1-0.5 mm. long; pappus light yellow, the inner bristles 5.5—-7.8 mm. long, the outer pappus of fimbriate scales, 0.5— 0.8 mm. long. Achenes pubescent, ribbed, 2.3-2.7 mm. long. Chromosome number # = 17. Flowering and fruiting from January to July. DISTRIBUTION. Limestone soils and pinelands, in Florida and the Baha- 1978] KEELEY, VERNONIA 411 SECTION TEPHRODES 42. Vernonia cinerea (Linnaeus) Lessing, Linnaea 4: 291. 1829. Syn- onymies, typifications, and representative specimens given in Keeley and Jones (1977b). Annual, 3-6 dm. high; stems tomentulose with T-shaped hairs, some- times becoming glabrate with age. Leaves scattered along stem; petioles margined, ca. 1.5-2.5 cm. long, pubescent; blades 1.5-2.5 cm. wide, 2-5 cm. long, lanceolate, pubescent above, pubescent with T-shaped trichomes and punctate beneath, apex acute, base attenuate, margins remotely toothed. Inflorescences of loose, open, and often spreading cymes; heads 12- to 16-flowered; involucres campanulate, 6-7 mm. high, 5-6 mm. wide, phyllaries loosely and irregularly imbricate, the inner phyllaries linear- oblong, 5—5.5 mm. long, 0.6-0.8 mm. wide, with acuminate to subulate purplish tips, the outer phyllaries 0.5-1 mm. long; pappus white, decidu- ous, inner bristles ca. 4 mm. long, outer bristles ca. 0.2 mm. long; corollas purplish lavender, 6-7 mm. long. Achenes rounded, nearly ribless, ca. 1.5 mm. long. Flowering and fruiting throughout the year. DisTRIBUTION. Antigua, Bahama Islands, Cuba, Curagao, Dominica, Dominican Republic, Grand Cayman, Grenada, Guadeloupe, Haiti, Jamaica, Martinique, Montserrat, Netherlands Antilles, Puerto Rico, St. Lucia, St. Kitts, St. Vincent, Tortola, Trinidad and Tobago, Virgin Islands. This species is extremely weedy and is found in almost all disturbed habitats. EXCLUDED SPECIES Vernonia anthelmintica (L.) Willd. Sp. Pl. 3: 1634. 1804. This Old World species is a well-known Vernonia that was imported to Jamaica for its medicinal properties. It was collected at Green Harbour at the turn of the century, but has not been re-collected since. For this reason it is not included in this treatment. Vernonia tetrantha (Urban) Ekman, Ark. Bot. 13: 8, 9. 1914. This taxon was described under Piptocarpha and has not been reevaluated since Ekman’s study. It may well be a Vernonia, but a decision awaits further study because the detailed comparison of genera is beyond the focus of the present study. Ekmania lepidota (Griseb.) Gleason (Bull. Torrey Bot. Club 46: 250. 1919) was previously treated under Vernonia. Gleason noted the unusual ring of fused outer pappus scales in this species and removed it from Vernonia to Ekmania. It appears to be distinct. Two species of subsection GraciLEes have been reported to occur in the West Indies. These are Vernonia gracilis HBK. subsp. tomentosa Ekman and V. tricephala Gardner. There is only one record of Vernonia gracilis subsp. tomentosa from Bequia, a small island in the Grenadines, in 1889 412 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 59 (Dalton & Smith 288, K). It is possible that this plant was collected there, but it has not been re-collected since. Vernonia gracilis subsp. gracilis is found in Colombia near the Caribbean, and in Brazil; it may be that the West Indian location was in error or that the plant was adven- tive on Bequia. Vernonia tricephaila, likewise, is found principally in South America. It was collected around the turn of the century at Cedros, Trini- dad (Broadway 2174, c-pc), but has not been reported since. ACKNOWLEDGMENTS Many people have given generously of their time and expertise through- out this study. I would like particularly to thank Dr. Samuel B. Jones, Jr., without whose guidance and support this project could not have been completed. I also thank Mr. G. R. Proctor, of the Institute of Jamaica, for great assistance in locating Vernonia in the field, and Nancy Coile, of the University of Georgia, for arranging loans and providing many help- ful suggestions. I would also like to thank my husband, Dr. Jon Keeley, whose enthusiasm and help were a constant source of support and en- couragement. The continued support of the Botany Department is gratefully acknowl- edged, as is the loan of herbarium specimens from the many herbaria cited in this revision. The typescript was prepared by Ruth Killinger, of California State University, Long Beach; her assistance is gratefully acknowledged. Financial support was provided by the National Science Foundation, grant number BMS74—-23659, and by the Department of Botany, Uni- versity of Georgia. LITERATURE CITED BAKER, J. G. 1873. ae a F. P. von Martius, FI. brasil. I. Ver- noniaceae. 6(2): 1-180. pls. EKMAN, E. 1914. ae Indian ae Ark. Bot. 13: FEATHERLY, H. I. 1954. Taxonomic terminology of the vie plants. 166 + . Iowa State College Press, Ames. Fournier, P. 1932. Voyages et découvertes eae des missionnaires na- turalistes francais. 258 pp. LeChevalier & Fils, Paris GLEASON, H. A. 1906. A revision of the North pares Vernonieae. Bull. Y. Bot. Gard. 4: 144-243 . 1922. Vernonieae. N. Am. Fl. 33: 52-95. GRISEBACH, A. H. R. 1861. Plantae Wrightianae e Cuba Orientali. Mem. Am. Acad. 8: 153-192. 1864. Flora of the ar West Indian Islands. Parts IV, V. 789 + xvi pp. J. Cramer, Wei 1866. Catalogus Aarne cubensium. 301 pp. Engelmann, Leipzig. 1879. Symbolae ad floram Argentinam. Abh. Ges. Wiss. Gottingen 24: 49-279, JONES, 5. B., JR. 1977. Vernonieae —a systematic review. Pp. 501-519 in KEELEY, VERNONIA 413 Biology and chemistry of the Com- 1978] V. H. Heywoop & J. Harporne, eds., positae. Academic Press, London. KeEELEy, S. C., & S. B. Jones, Jr. 1977a. Taxonomic implications of external pollen morphology to Vernonia (Compositae) in the West Indies. Am. Jour. Bot. 64: 576-584. . 1977b. Vernonia (Compositae) in the Bahamas — reexamined. Rhodora 79: 147-1 LessInc, C. F. 1829. Synanthereis herbarii regii Berolinensis dissertation prima. Linnaea 4: 240-356. De synanthereis. IV. /bid. 6: 624-721 LINDLEY, J. 1848. Illustrated dictionary of botanical terms. Reprinted 1964, School of Earth Sciences, Stanford University, Stanford, California. RIcHARD, A. 1950. Fanerogamia. Vol. 11 i R. DE LA SAGRA, ae pe politica y natural de la isla de Cuba. 339 pp. Arthus Bead. SCHULTZ-BiponTinus, C. H. 1847. Compositae. Linnaea 20: pores 1863. Adfiotationes in See Wrightianas Cubenses, a a Grise- 1: 231-237. bach determinatas. Jour. B ondon URBAN, I. 1899. Species novae ee ee Symb. Antill. 1: 456- 481. Reprinted 1964, Asher, Amster 1903. Nova genera et species II. bid. 3: 390-420. . 1911. Flora portoricensis. /bid. 4: 618-622. . 1912. Nova genera et species. V. /bid. 7: 421-432. Present address: DEPARTMENT OF BOTANY NATURAL History MUSEUM 900 ExPoOsITION BOULEVARD Los ANGELES, CALIFORNIA 90007 DEPARTMENT OF BOTANY NIVERSITY OF GEORGIA ee GEORGIA 30602 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 INDEX eee oo 60 Acac ee eee 60 — torulosa, 60 Acalypha wien. A Acanthaceae, 60, 61, Acanthocarpus, ae 137, 145, 146, — preissii, 136, 137, 142, 143 Acrotrema, 44 Actinidia, 284 Actinidiaccae, oe 287, 289, 292 67 141, 143, Allagoptera arenaria, 123 Allium cepa, 65 +! Aloé barbadensis, 65 — vera, Alpinia sect. Myriocrater, 150 Altamiranoa, 21 Alternanthera ramosissima, 60 2,6 Anacardioxylon ieee osu Anatomy and cen of nn Sec- tion Moutan, Wood, 274-29 Anatomy of Hibbertia (Dilleniaceae) in Relation to Ecology and Evolution, 32-49 Myrothamnus halen ees (Myrothamnaceae), The other Example of a “Split- tere as dition, 192-196 Annona squamosa, 60 Annonaceae, 60, 65, 286, 287, 289, 290, 292 Anthurium, 65 Apeiba, 299 = ome 299, 300, 302, 303, 305, 307— Aphlantes 130, 132 Araliaceae, 65, 66 Araucaria columnaris, 65 Ardisia escallonioides, ie Arecaceae, 63, 65-67, Acetasirum oan 65 Arenga undulatifolia, 116 Argemone mexicana, 60 Arundo donax, 60 Asclepiadaceae, 61, 66 ae cen densiflorus, 65 — setac pee Ae 65 Asplundia, 83 Asteraceae, 60-66, 192 Austin, Danitet F., and Daviy M. Mc- UNKIN. An Ethnoflora of Chokolos- Island, Collier County, Florida, 50- — alasia, Rhizophora in — Some Clari- fication of Taxonomy and Distribution, 156-169 Australian pine, 65 Avicennia, 344 — germinans, 60 Avicenniaceae, 60 Avocado, 63 Bactris, 106 ee 123 Balsam apple, 63 Banana, 66 Banded wild pine, 64 ELISABETH 5 p A. Scott, Fos- sil Dicotyledonous Woo Yellow- — aan tional Park, IT, 1- Bataceae, 60 Batis cae ae 60 BARGHOORN, ELso ee 1978] Bauhinia i cat 65 Baxteria, 129, Beach ae 62 Beach-orach, 60 Beach sage, 62 Beefwood, 65 Befaria, 315, 332, 336-338 — mexicana, 299, 300, 302, 303, — 309 Berberidaceae, 283, 284, 287-28 — subfam. Paeonioideae, 284 — subfam. Podophylloideae, 284 Bidens pilosa, 60 ied a 67 Bird p eens Ae 106 Black mangrove, 60 Blechum brownei, 60 Bloodberry, 64 Bloodleaf, 62 Blue porterweed, 6 Born, Bruce A., = W. Brim, RIcH- RD J. STEVENS. Ge- = . w o 2 S 5 | neae (Ericaceae), and Its ea in the Rhododendroideae, 311- Bombacaceae, 65 Bombax malabaricum, 65 65 Botryostege, ee 332, 335 — eas ma aac 335 aide ee 65 Bougainvillea, 65 — glabra Bowne sn he Brassaia are i 65 eryngioides to Perezia (Compositae, Mutisieae), Transfer of the, 352- Brim, Scott W., RicHarp J. Hepp neae (Ericaceae), and Its Position in the dodendroideae, 311-341 Bromeliaceae, 64, 67 Brown gum, 66 Brownetera, 261 econ um ie 240, 241 — pinn es eo ee from Co- INDEX lombia, A Speceacelan, 298 Bumelia celastrina 60 mice - . B. Primack, N. C. 7 fo sa cee fae sea eee retaceae) oe Status and Floral paneeaiag 34 Bursera simaruba, 51, 55, oa Burseraceae, 2, 60 Bush morning glory, 62 Buttonwood, 61 Byrnesia, 220 Cabbage, 65 Cabbage palm, 67 Cabbage palm fern, 63 Cacalia acuminata, 366 — a aan 393 — thomae, 373 — to ene 409 Callstemon fiance olads, 65 Callun Bae 284, 286, 287, 289, 290 Calyptrogyne BrAchy stachys, 123 Calyptronoma occidentalis, 118 Camptotheca, 14 Cananga odorata, 65 Candelabra cactus, 66 Cane palm Gancliaceac: 5a. 289 Canistel, 67 Cape ee ea 67 Capparace erie ae 60 Cardiospermum halicacabum, 61 Carex, 215 Carica papaya, 61 416 JOURNAL OF THE ARNOLD ARBORETUM Caricaceae, 61, 72 Catludovica, 74 Carp aaa 299 ameliae, 299-303, 305, 307-309 Caryophy ace ae, 217 Caryota m Cassia Segue: 61 Castanopsis acuminatissima, 150 Castor bean, Casuarina equisetifolia, 65 — glauca, 65 Casuarinaceae, 65 Catharanthus roseus, 65 Cat’s claw, — occidentalis, 19 Cenchrus echinatus, 61 — incertus, 6 Century plant, 65 Cereus peruvianus, 65 62 Corer lon a 108 Chaff-flow a. ne. 123 — elatior, 122 Chamaerops humilis, 111 ec ine blodeetti, 61 — hirta, 6 —hy ae 61 —hy jenena 61 — porterana, 61 Chamaesers 129-134, 137, 141, 143, 145, , 148 ee ee ae —macranthera, 134, 13 — serra, 134, 137, 146 Cheeseweed, 63 Chelyocarpus dianeurus, 114 poeneporinn 60, 64 2 Chokoloskee a Collier ee Flor- ida, An Ethnoflora of, 50- Christmas berry, 62 Christmas palm, 67 Christmas rose, 6 Chrysalidocarpus lutescens, 65 Chrysophyllum oliviforme, 61 Cistaceae, 218 Citrus aurantium, 65 Cladothamneae (Ericaceae), Generic Lim- [voL. 59 its in the Tribe, and Its ae in the Rhododendroideae, 311-34 Cladothamnus, 311, 312, Ae 332, 335, 33 — pyrolaeflorus, 336 —pyroliflorus, 311-316, 318, 319, 322, 323, 325-327, ae 330, 332, 333 oo 287-289, 291 Clemestaia,2 Clinostigma savaiiense, 114, 116 Citrus aurantifolia, 65 oo Pe — sin ce ie philippinum, 61 Cliftonia, 4, 5 Coccoloba diversifolia, 61 — uvifera, Coconut, 65 Cocos nucifera, 65 Codiaeum variegatum, 65 Coffee weed, 61 Coleus, 6 — blum Colombia a a asad oe (Scro- om Combretaceae: Lumni . zera rosea — Its Status and Floral Morphology, 342- 357 Combretaceae, 61, es 67, 342 Commelina ciftus, 6 dian eats 360-41 Compositae, Mutisicae® — of the ne Trixis eryngioides to Perezia, 352-35 Composie, 286, 387 — tri Sais ce subtribe Nassauviinae, ae — tribe Vernonieae, 360 Coniferae, 71, 72 Gonoearus erectus, 54, 61 Convolvulacea Conyza, 370 arborescens, 379 1 Cordia sebes , 65 Cordy australis, 65 Corn 1978 | an 11, 12 Cornu Cores im 219 Cot Cs eee 336 Crassula, 198, 206, 231-236 — sect. Pyramidella, 236 — sect. Stellatae, 236 — sect. Tillaeoideae, 201, 232-236, Southeastern or 48 Crassulaceae, 60, 62, 197-248 ub . Cotyledonoideae, 241, 242 —subfam. Crassuloideae, 230, 231, 241, 242 Echeverioideae, 200, 218-220 — subfam. Kalanchoideae, 237, 241, 242 — subfam. Sedoideae, 200, 206, 219 — subfam. Semper — subfam. CrIscI, JORGE and C1 IRO MartTicorENA. Transfer of the Brazil- ian Trixis eryngi s to Perezia oe Mutisieae), 352-359 Crossosoma, 293 Cross ssoinataccae 274, 284, 287, 289, 293 Crotalaria i inca 66 Cryosophila albida, 118, 122 Cryptostegia madagascariensis, 66 Cyclanthaceae, Two New Species and a New Subgenus of, 74-101 See 74-101 Cyclant Cy ae um, 61 Cyperaceae, 61, 62, 66, 215 Cyperus alternifolius, 66 — ligularis, 61 Cyrilla, 4,5 — racemiflora, 4 Cyrillaceae, 3-5, 11, 70 Spectral ao nicum, 3-5, 20, 26 — europaeum, 5 Dactyloctenium aegyptium, 61 INDEX Daemonorops grandis, 114, 116 , 66 Dates of een of Sargent’s Silva of North America, 68-73 6 Desmodium tortuosum, 61 Desmoncus Dismerpha, ae 214, 217, 219, 226-230 Dicx1sow, Wirttram C., PHILLIP - Rury, Xylem Anatomy of Hibbertia (Die ceae) in Relation to Eco and Evolution, 32-49 Dicliptera assurgens, Dicotyledonous ae from Yellowstone National Park, Fossil, IJ, 1-31 Dicranopygium, 74, 77, 79-81, 83, 87 — subg. Dicrano 8 1 — subg. Tomlinsonianthus, . 75, 79-81 eke ne ama 74, 79-8 — Cam ee ie e 81, 82, 84, 86 oe —robustum, 81 areas 42, 44 Dillenia, 44 Dilleniaceae: Xylem Anatomy of Hibber- tia in Relation to Ecology and Evolu- tion —49 Dilleiaceae, 32, 44, 48, 274, 284, 286, 7, 289, 290, 292- —su ae Dllenioien, 292 Dioscorea bulbife Distribution, Rhizophora in Australasia — Some Clarification of Taxonomy and, 156-169 Ditch-grass, Dizygotheca ae 66 Dropseed, 64 Dudleya, 218 UKE, C. B. Tomutnson, J. Bunt, anid R. B. a ae ro (Combret —Its Status and Floral Morphology, ee 418 Ebenaceae, 71 Echeveria, 200, 219, 220 — sect. Graptopetalum, 220 — sect. Thompsonella, 220 Ecology and Evolution, Xylem Anatomy of Hibbertia (Dilleniaceae) in Relation to, 32-49 Edenoxylon parviareolatum, < Egyptian grass, Ekmania, 411 — lepidota, 411 Elacis oleifera, 119, 123 Elliottia, 311-313, 329, 330, 332-338 —bracteata, 312-314, 316-320, 322, 323, 325-327, 329, 330, 332-335, 337 — — forma bracteata, 335 oe 335 — paniculata, 312-320, 32 oe 330, 332- 336 — pyroliflora, 334-337 —racemosa, 312-323, 325-327, 329-332, 334-33 antennas 41, 42 Epiphyllum, 66 Eragrostis ee is, 61 neric coe in the Tribe Cladothamneae, and Its Position in the Riededendeoidcad cae Ericaceae, 217, 312, 313, 320, 331, 332, 36-338 — subfam. Monotropoideae, 337 — subfam. Pyroloideae, 33 — subfam Rhododendroideae, 311-313, 336-338 — — tribe Befarieae, 336, ; —~— — tribe Cladothamnese, nee 341 — — tribe Diplarcheae 7 & —subfam. Vaccinioideae tribe Androme- deae, 315, 336 Eriobotrya japonica, 66 Erythrina herbacea, 61 Ethnoflora of Chokoloskee Island, Collier County, Florida, An, 50-67 Eucalyptus, 4 — grandis, 66 Eugenia axillaris, 61 —- foetida, — uniflora, 66 Eupatorium domingense, 396 —menthaefolium, 4 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Eupatorium obtusifolium, 372, 374 as 61 Euphorbia lactea, 66 — milil, — tirucalli, 66 ea ete nee 61, 63, 65-67 Euterpe cuatrecasana, 111 Evolution, Xylem Anatomy of Hibbertia (Dilleniaceae) in Relation to Ecology and, 32- Fabaceae, 60-67 Fagaceae, 5, 71 False siralin, 66 FrerNANDEZ, P., S. Lozano, and E. Mar- TINEZ-HERNANDEZ. Pollen of Tropical Trees. I. Tiliaceae, 299-309 Ficus aurea, 61 Fimbristylis caroliniana, 62 ad fa) eA ap oe, =e fo re) © Be) 4 Morphology, Lumnitzera rosea (Combretaceae) — Its Status and, 342- 357 men An os of Tend: ere County, Florida privet, 62 Forestiera oe 62 Form of the Perforation Plates in the Wide Vessels of Metaxylem in Palms, Chokoloskee 5-128 Fossil Diceiy Iedonous Woods from Yel- lowstone es Park, II, 1-31 Frangipani, Frostweed, Galactia macreei, 62 Geiger tree, Genera of satis in a. Southeast- ern United States, The, 197-248 Generic Limits in the The Cladotham- and Its Position in 341 tl Generic Limits in ve ee (Liliaceae sensu lato), -_ Gentianaceae, 19 Genus See (Phyllocladaceae), 9-273 Gesneriaceae, 192 1978 | INDEX 419 Giant reed, 60 Hibbertia cuneiformis, 33, 35, 37, 39, 40, Goatweed, 60 44 Goldenrod, 64 — dentata, 45 Gossypium hirsutum, 51 aminmnnda. 33, 35, 37, 40, 44 Gouania eee 62 ~- aa ee 33, 35, 38, 39 Grapefruit, — furfuracea, 33, 37 Soe te 200, 220 — glaberrima, 33, 37, . Gray nicker, 60 — huegelii, 33, 35, 37 Groenlandia, 187 — hypericoides, 33, 35, ee a 45,48 Grossulariaceae, 218 — lineata, 33, 37 Guava, 63 — lucens, 33, "35, 36, 40, 43, 44 Guinea grass, 63 — montana, 34, 35, 37 Guinea-hen weed, 63 — ngoyensis, 34— 36, 40, 44 Gumbo limbo, 60 — obtusifolia, 34-36, 39, 44 Guttiferae, 287, 289, 292 — potentilliflora, 34, 35, 37 — procumbens, 34, 35, 37, 41 Hamamelidaceae, 11, 12, 71 — saligna, 34-36, 39, 40, 44, 45 Hamelia p — scandens, 34-36, 40-43, 45 Hasseanthus, 218 — sericea, 34, 35, 37, 39, 40 Haynes, Rosert R. The Potamogetona- — serrata, 34, 36, 44, 45 ceae in fe Southeastern United States, — Stricta, 34, 35, 37, 43 170- — trachyphylla, 34-36, 40 Heart in 61 — uncinata, 34, 35, 37, 39, 40 Hespa, Ricuarp J., P. F. Stevens, — virgata, 34, 35, 37, 40, 43 Bruce A. Boum, and Scotr W. Brim. Hibiscus tiliaceus, 57 Generic Limits in the Tribe Cladotham- | Hydrophila, 232 neae (Ericaceae), and Its Position in Hylocereus undatus, 62 the Rhododendroideae, 311-341 Hylotelephium, 207, 209, 219 Hedge cactus, 6 Hyophorbe indica, 123 ee donnell-smithii, 299, 301-304, | Hyphaene guineensis, 106 06-309 Hyptis pectinata, 62 ie tropium angiospermum, 62 8 Helleboraceae, 2 Iguanura, 123 Ilex, 13, 14 icine eae, 70 Indian mallow, 64 Indigo berry, 63 Hensmania, 130 poy ian, 132, Ipomoea alba, 62 ernandiacea cee i Hibbertia ier in Relation to oe — indica, 62 prec and Evolution, Xylem Anat- h 39-49 —macrantha, 62 — triloba, 6 Hibbertia, 32-49 Peecne dius peel ane 37 Itaya pn ceiari: ey 122 — sect. Cyclandra, 36, 42 Ixora, 66 — — subsect. Subsessiles, 42 Sena ndn 66 — — subsect. Trimorphandra, 42 — sect. Hemipleurandra, 37, 42 Jamaica caper tree, 60 — sect. Hemistemma, 38, 42 Jambolan plum, 67 — sect. Pleurandra, 37, 4 Japanese rose, 61 — sect. Polystiche, 36 Juglandaceae, — sect. Spicatae, 3 Juglandinium caucasicum, 10 — altigena, 33, 36 — tasseewil, as , 33, 38 Juglandoxylon, 10 — banksii, 33, 38, Juglans, 9, — baudouinii, 33, 36, 39, 40, 44 — sect. Cardiocaryon, 9 — corlacea, 33, 35, 38, 40, 43, 44 — sect. Trachycaryon, 9 420 Juglans cinerea, 9 — fryxellu, 10 Juncaceae, 129 — tribe Calectasieae, 130 — tribe Xanthorrhoeae, 129, 130 — tribe Xeroteae, 129, 130 uncaginaceae, 171 Juniperus conterta, 66 a ak 200, 206, 217, 237-243 ; Beyopny llum, ak 243 ae Kalanchoé, 238, 240-242 — sect. Kitchingia, 238, 241 — grandiflora, 62 M., and Maynarp F. Jr. Wood Anatomy and eee of Paeonia Section Moutan, 274- ree | aus C. A Revision of the West Indian Vernonias (Compositac), 60-413 The Genus Phyllocladus 249-273 HSsvuan. (Phyllocladaceae), Key lime, 65 Kidney bean, . 0, 24 Kviorz, Larry H. Fo orm of the Perfora- tion Plates in the Wid Nese of Metaxylem in Palms, 105-12 Lactuca sativa, 66 — ae 62 Ledothamnus, 336 Ledum, 337 te 41, eae de byraaeforum, 336 nophyllum, 200, 21 Snare beats on — arborescens, 3 — buxifolia, ve JOURNAL OF THE ARNOLD ARBORETUM [voL, 59 Lepidaploa — 373 hyllos tachya, . Ligustrum amurense, 62 — ovalifolium, 62 Liliaceae sensu ae Generic Limits in the Xcroteae, 129-155 Liliaceae, 64, 65, — su whem: New hmae loldede: 130 — — tribe Dasypogoneae, 130 — tribe Xeroteae, 129-155 Lime, 65 Liri odendron, 292 Lithocarpus, 6, 8 Live-forever, 60, 207 Lomandra 129-134, 137, 148-150, 152 — sect. ae 137, 141, 146 — sect. Eulomandra, ee — — ser. Sparsiflorae — sect. Lomandra, a ue 138, 140, 141 — sect. Macrostachya, 138 — sect. Schoenolomandra, 137 — sect. Typhopsis, 137, 138 —banksii, 129, 134, 140, 5? 143, 145, 146, 141, 147, 150- — bracteata, 141, 143 — brevis, 141 —contertifolia, 140, 141 — cylindrica, 137, 139, 141-143 — densiflora, 141 — endlicheri, 130, 132, 139, 141, 146 — fibrata, 139, 141 — filiformis, 139, 141 — glauca, 140 — — subsp. collina, 137 — — subsp. glauca, 137 — hastilis, 135, 138 — leucocephala, 134, 138, 143, 146 1978 | Lomandra leucocephala leuco- cephala, 130, 132, 134 — — subsp. robusta, 138 — longifolia, 140, 141, 143, 150 — macrostachya, 1 — micrantha, 130, 13, 137, 139, 141, 146 — montana, re — mucronata, a — multiflora, na 138, 140-143, 147, 150, L253 — obliqua, 134, 137, 140 — ordii, 141 — papuana, 149 subsp. 8 129, 132-137, 141, 146, 148, — patens, 141 seater? 141, 145 — preissii, 139, 141 Se 139, 141 s ee — sororia, 134, 137, 139, 141, 146 1 Lozano, S., E. Martinez-HerNAnpbeEz, and FernAnpez. Pollen of Tropical ees. I. Tiliaceae, 299-309 Ludovia, 81, 91, 93, 97 — Bierhorstii, 75, 83, 85, 87-100 — integrifolia, 81, 85, 93, 95 Siar ea 81, 85, 93, 95, 97 Luechea — on 301, te ae Lumnitzera rosea Status and Floral Soa ee 342 ae ee 342, 344-350 — littorea, 342-350 —_—— i) vavemiora, 348, 349 — lutea, — racemosa, 342-350 Lycium carolinianum, 62 OOH Ly ee eae 16 Magnoliaceae, 70, 284, 287-290, 292 Magnoliaceox Mahogany, 67 INDEX Maloidoxylon, 16 Malvaceae, 64 Mangifera indica, 66 Mango, 66 Mantes saccifera, 123 zapota, 66 CLopomiro, and JorcE Crisci. Transfer of the Brazil- 1S ery ee to oe (Com- FERNANDEZ, eae of Tropical Trees. 5 Tiliaceae, 299- Mastic Me en St 62 fauritia vinifera, 108, Mauritiella pacifica, ae {cJuNkin, Davin M., and Dawnier F. AUSTIN Ethnoflora of Chokolos- kee ilaad Collier County, Florida, 50- 6 Melanthera parviflora, 62 Melia azedarach, 62 Meliaceae, 62, 67 Melicoccus baueas: 66 Mentzelia floridana, 62 Metaxylem in Palms, Form of the Per- foration Plates in the Wide Vessels of, 105-128 Mexican popp Milk and wine a 65 Milk pea, 62 Momor diea eae 63 Monstera deliciosa, 66 Morning glory, 62 Morphology, Lumnitzera rosea (Combre- taceae) —Its Status and Floral, 342- 357 Moserey, Maywnarp i Jr., and JosEPH M. KEEFE. Wood Anatomy and Phylo- geny of Paeonia Section Moutan, 274- Musa paradisiaca var. sapientum, 66 Musaceae, Mutingia anaes, 66 Myrica, 11-13, 21 — absarokensis, 10-13, 26 —cerifera, 11, 13 — rubra, 11, 13 sca Slay Gani 12 Moricaceie. 10-12 422 Myricoxylon hungaricum, 12 Myristicaceae, 287, 2 Myrothamnaceae Nodal Anatomy of Myrothamnus flabellifolius: An- other Example of a “Split-lateral’”’ Con- dition, 192-196 Wins dae. 192, 194 ies aes flabel eae (Myrotham- naceae), The Nodal Anatomy of: An- other Example of a rae lateral” Gon: dition, 192-196 Myrothamnus, — flabellifolius, 192-196 S | 3 5 n a i) 2s Murtavexe. 61, 63, 65-67 Natal grass, 64 Necklace Needle-leaf w tA ee 64 Nerium oleander Night-blooming pane 66 Night- ae cereus, 62 Nightshade, Nodal ra “ wn n io} iw wa [o) “ a saximontanum, 26 Oleaceae, 62 Orchid tree, 65 e, 63 Oxalis corniculata. 63 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 Pachynema, 39, 43, 45 so na OL at Pachy phy tum, 200, Paeonia Section a. w ie Anatomy and Phylogeny of, Paeonia, 274-29 — sect. Moutan, 274, 279, 281, 294 — — subsect. Delavayanae, 275 ae: Vaginatae, 275 —— sect. Onaepia, 274, 288 — sect. Paeonia, 274, 288 — californica, 279 — delavayi, 275, 28 lut ices 278, ie , 288 — lutea, 278, ae 278, 280, 282 —— potaninii, 275 — potaninii, 278, 278, 281 —-— forma alba, 282 — suffruticosa, 275, 278-281, 283 pete a 279, 283, 284, 286, 288- 290, 293, 283, 288, Palandra 7 de 119 Pa a - spatulata, 14 Palm Palms, on of the Perforation Plates in the Wide Vessels of Metaxylem in, 105- 128 eavaun maximum, 63 , 61 Parca floridana, 63 serena! quinquetolia, 63 aes Peac ee een Pedilanthus tithy oe 66 Pelargonium hortorum, 66 Pepper tree, 67 Perdicium magellanicum, 357 Perezia (Compositae, Mutisie eae), Tran fer of the Brazilizn Trixis ee < ° "w OL RS | ‘ mn — catharinensis, 357, 35 — eryngioides, oe a — kingii, 357, : — magellanica, 7 — multiflora, 387, 358 1978] INDEX Perezia multiflora subsp. multiflora, 357 — — subsp. sonchifolia, 35 — squarrosa, 357, 358 — — subsp aaa oe ubsp. squarrosa, Perforation Plates in i aide Vessels of Metaxylem in Palms, Form of the, 105- 128 Periwinkle, 65 Phaseolus aureus, 329 — vulgaris, 66, 328 Puitcox, D. A Spectacular Buchnera Geroanilaracens) from Colombia, 298 Philoxerus vermicularis, 63 Phlebodium aureum, 63 Phoenix Fie rag 63 — roebellini — rupicola, 6 ee ee 123 ope austra —com , 63 th ie The Genus Phyllocladus, ~273 Bia ee 49 Phyllocladus (Phylocasce), The Ge- nus, 249— Phyllocladus, yee 273 oe 249, 263, 266 aspleni folie: 249, 255, 258, 259, 261- 263 — alpinus, 249, 251, 253, 256, 258, 261-267 —-— var. aspleniifolius, 249-252, 255-264, 70 — Pile, 262, 263 — glaucus, 249, 251-262, 265-267, 270 See 249, 251, 253-261, 267- 269 —— var. protracta, 267 — major, 26 nce 267 —rh mola, 249, 261-263, 266 — a ~ tichomanoides 251, 253-260, 262, 265, 267, 26 —vV Phy logeny of Paeonia Section Moutan, Nood Anatomy and, 274-297 Physokentia rosea, 110, 114 ocarpa, 119, 123 Phytolaccaceae, 63, 64 Pieris japonica, 315 Pigafetta filaris, 123 Pinanga, 110, 123 Pinus, — auStralis, 335 Piptocarpha, 411 Piscidia piscipula, 63 Pisonia aculeata, 63 Pistacia, Pithecellobium unguis-cati, 63 Planera, 19 Plataninium, 1 Platycerium bifurcatum, 66 66 Povdocarpaceae, = Podocarpus — eae 262, 263 117 Poinciana, 6 Poinsettia, 66 — cyathophora, 63 — pulcherrima, 66 Poison ivy, Pollen of Tropical Trees. 299-309 I. Tiliaceae, Polyandrococos caudescens, 116 Polygonaceae, 61, 65, Polypodiaceae, 63, 66 Polyscias ae 66 — guilfoylei, Pomegranate, 2 Pondweed Family, 170 — subg. Potamogeton, 175 —— sect. Adnati, 175 — — sect. Axillares, 175, 176 Potamogetonaceae in the Potamogetonaceae, 170-191 Southeastern 424 Potamogetoneae, 170 Potato vine, 61 Pothos, 67 Ponter campechiana, 67 Prickly pear, 6 Ww . N.C. Duke, P. B. Tom LINSON, and S. Bunt. eee rosea éCombsctacede) — Its Status and Floral Morphology, 342-35 Privet, 62 Procrassula, 214, Progymnospern Progymnospe oo 249 Proteaceae, 4 Proustia domingensis, 396 Pruninium gummosum, 15, 16 Prunus, 15, 16, 2 — gummosa, 15-17, 20, 26 _ — serotina, 16 Pterocarya, 9, 10 — rhoifolia, 9, 10 Pterocaryoxylon, 9, 10, 21 —chinense, 10 onshouense, 10 oe lags 8-10, 26 —su bpannonicum, 10 Publication of Sargent’s Silva of North erica, Dates of, 68-73 eae Cristian, The Nodal Anatomy of Myrothamnus flabellifolius (Myro- thamnaceae) : ther Example of a “Split-lateral” aa 192-196 Pull-and-hold-bac Punica ote 67 Punicacea 7 Purdiaea, Purple on 66 e, Pyrola fruticosa, 336 Quamoclit, 62 Queen’s palm, 65 CHecem, 6, 7, 21 amethystianum, — setae 8 , 6, 8, 26 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 a se know oo 6, —lam —8, — [esa uercuxii, 8 Quercus, 8, 71 — ricardensis, 8 —rubida, 6 Randia aculeata, 63 Ranunculaceae, 274, 283, 284, 287, 2 2 293 Rattlebox, 61 Red mangrove, 63 Red a aoe tree, 65 Reed, ean simplex, 115, 12 Revision of the West indian Vernonias (Compositae), A, 360-41. Rhabdadenia biflora, 63 Rhamnaceae, 62, 72 Rhappidophors: aurea, 6 Rhapidophyllum hystrix, 108, 111, 123 Rhapis excelsa, 123 Rhizophora in Australasia — Some Clari- fication of Taxonomy and Distribution, 156-169 Rhizophora, 156-1 apiculata, re ae 165-168 — harrisonii, 15 — lamarckii, oe 162, 163, 165-168 — mangle, 54, 63, 156, 158-161, 163 —— var. samoensis, 156 — mucronata, 156, oS a 163, 167, 168 — — var. sla 158, — racemosa, 156 — samoensis, Te 164, 167, 168 —‘selala’, 159 — x seal 156-159, 161, 162, 164, 167, 20 156-159, 161-168 nes ease As 192 Rhodiola, 2 seen eer eae, Generic Limits in the Tribe Cladothamneae (Ericaceae), and Its Position ae the, 311-341 Rhododendron, 332 — subg. Anthoamndsan 217 37 ae 2 — Rhus, 2 “sub Toxicodendron 253 — crystallifera, 1-3, 1978] Rhus striata, 3 Rhynchelytrum repens, 64 Rhynchosia parvifolia, 64 Ribes, 218 Ricinus communis, 67 Rivina h , 64 — papuana, 147, 149, 150 Rosa, 67 Rosaceae, 15, 16, 66, 67, 70, 72, 287, 289, 292, 293 Chrysobalanoideae, 15, 16 Maloideae, 287-289, 293 Prunoideae, 15, 16, 287-289, — subfam. — subfam. — subfam. 293 —subfam. Rosoideae, 287, 289, 293 bfam. Spiraeoideae, 287, 289, 293 Rosaceoxylon, 16 Rosary pea, 60 Rury, Puityre M., G. LepyArp STEBBINS, and WititiAm C. Dicxkison. Xylem Anatomy of Hibbertia (Dilleniaceae) in —— to Ecology and Evolution, ee equisetiformis, 64 Rutaceae, 61, 65 Sabal, 67 Sansevieria metallica, 51, 64 Sapindaceae, 61, 64, 66, 70 Sapindus saponaria, 55, 64 Sapodilla, 66 Sapotaceae, 60-62, 66, 67, 71 Sargent’s Silva of North America, Dates of oe ieee of, 68-73 Satinlea Sori 287, 289, 292 Saw s, 61 eee 70, 2 218 — subfam. ee 200 INDEX 425 Saxifragaceae subfam. Saxifragoideae, 216 ee actinoporosum, 3 Schin — trent, 67 o S. BaRGHOORN, Fossil Di- cotyledonous Woods et Wolo watone National Park, II, 1-31 Scrophulariaceae: A Spectacular Buchnera from Colombia, 298 Scrophulariaceae, 60, 64 Sea blite, 64 Sea grape, 65 Seaside mahoe, 64 Sedella, 219 « ot 220 Sedge Sedum, ou 201, 206-220, 227-230, 235 — subg. — subg. — subg. — sect. Al — sect. Bergerosedum, 219 — sect. Dendrosedum, 219, 220 — sect. Epeteium, ee 214, 235 — sect. Eusedum — sect. Leteee e 220 — sect. Pachysedum, 219 — sect. Procrassula, 214, 235 — sect. Rhodiola, 208, 218 — sect. — sect. — sect. Sedum, 211, 213, 214, sect. Telephium, 208, 209, re 218 Selala, 158, 159 Sempervivum, 201 Serenoa repens, 53, 123 Sesbania exaltata, 64 Setcreasea purpurea, 67 Silva of North America, Dates of Publi- cation of Sargent’s, 68-7 Sindroa longisquama, 119 Slipper-flower, 66 50 rry, 61 Soap oe tree, 64 426 Solanaceae, 61, 62, 64, 66 num americanum, 64 Dates of Publication of Sargent’s Silva of North America, 68-73 Sour oran Seer ete awe States, The Genera of Crassulaceae in the, 197-248 Southeastern United States, The Potamo- getonaceae in the, 1 Spanish dagger, 67 racer lime, 66 Spa moss, 67 Spanish needles, 60 s, 64 Spathodea campanulat Spectacular Buchnera “Seroptriacee from Colombia, A, 2 Sphaeradenia, 83 “Split-lateral” Condition, Another Ex- ample of a: e Nodal Anatomy of eon ae et (Myro- thamnaceae), 196 Spondias ils 2 — mombin, SPONGBERG, — HEN A, Genera of he Crassulaceae in the Southeastern United States, 197-248 Sporobolus domingensis, 64 Spurge, 61 Squash, 66 Stachytarpheta jamaicensis, 64 Stachelina solidaginoides, 409 Staghorn fern, Staphyleaceae, 17, 18 STEBBINS, G, Lepyvarp, WILLIAM C. DIckI- son, and Puitrtie M. Rury. Xylem Anatomy of Hibbertia (Dilleniaceac) in Relation to Ecology and Evolution, Sterculiaceae, 65 JEN in the 129- Generic Limits eciltieess sensu lato), ? os Sen, : F., Bruce ae Boum, ScoTt = w % a neae (Ericaceae), and Its Position in the aetna aaa 311-341 Stonecrop, 2 Stonecrop amy, 197 Strangler ee bes 66 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 59 Streptocarpus hybrida, 329 Stylosanthes biflora, 64 it Swietenia oe 67 Synechanthus, Syzygium cuminii, 67 Tamarind, 67 Tepaundus indicus, 67 Tapirir nena and Distribution, Rhizophora in Australasia — Some Clarification of, 156-169 Tecoma stans, 64 Tecomaria capensis, 67 Telephium, 218 Seay catappa, 67 Tetror eae 219, 230 Thales 2 —ssptoiolin 262, 263 Thatch palm Theaceae, ie 2 — tribe perenne 292 — tribe Ternstroemieae, 292 Thespesia populnea, 64 Thompsonella, 200, 220 Thrinax radiata, 53, 67 Thuja orientalis, 67 Tick trefoil ete 9-; Rae 1, 61 Pollen of Tropical Trees. Dias, 299-. Tillae 227, 20, 232, 234, 235 Tillacastrum, 2 Tillandsia eee 64 fasciculata, 64 Rhiz Clarification of Taxon- 69 rosea (Combretaceae) — Its Status and Floral Morphology, 342-357 Torricellia, 11 Toxicodendron, 3 — radicans, 64 1978] Transfer of the Brazilian Trixis eryn- gioides to Perezia (Compositae, Muti- sieae), 352-35 aoe ee of Tropical. I. Tiliaceae, ie nae octandrum, 64 Tridax 4, Scan oh giracemosa, 4 — panieuiata, “312, se a5 on 335 akusimensis, 312, i 121 Trixis eryngioides to Perezia (Composi- tae, Mutisieae), Transfer of the Brazil- ian, 352-35 see 352, 356-358 alata, 35 —ery neioides: 352 —inula, 356 a spinal almond, 67 Tropical Trees, Pollen of. J. Tiliaceae, 299-309 Turpinia, 17, hep eet sis lamarense, 1, ae 20, 26 ae New Species and - New Subgenus of Cyclanthaceae, 74-1 Ulmaceae, 18-21 Ulminium, 1 Ulmus pacifica, 16 Umbrella plant, 66 Umbrella tree, 65 Urticaceae, 63 Vaccinium, 337 Veitchia merrillii, 67 Verbenaceae, 61, 62, 64 Verbesina virginica, 65 Vereia Werner. 360-4 — sect. Uepidaplo, 361, 362, 365 — — Arborescentes, 362-365, 369, Tee ee — — subsect. Graciles, 411 — — subsect. Pallescentes, 362, 408 — — subsect. Paniculatae, 1 — — subsect. Polyanthes, 362, "365, 399 INDEX 427 Vernonia sect. ine subsect. Sa- graeanae, 362, 365 2 — subsect. ee 362, 409 11 — sect. ee 407 — minata, pe 368, 378 — acufiae, 385 pecs, 364, 2 bsp. ce 368, 372- eT longistylis, 368, 372, 374, 375 78 — arborescens, 364, ae 379, 381, 382 — — var. divaricata, 390 —— yy divaricata, 378 ——fp sane 375, 381 — 6 ovatifolia, 373, 381, 382 a se Weel, 363, 392, 393 aronifolia, 368, 381, 402, 403 — barkeri, 365, 395, 396 — Berteriana, 375 eee ttii, 361, 410 orinquensis, 364, 376-378 — var. hirsuta, 37 —— var. resinosa, 377 — — var. stahlii, 377 — buchii, 391 — buxifolia, 365, 396, 397 — calida, 3 4 —calophylla, 384 — cinerea, 361, 411 —commutata, 363, 380, 394, 395 — complicata, 363, 380, 392, 393 —corallophila, 393 - deciliens. 363, 8 — divaricata, co a 378- 382, 384 — domingensis, 3 — ekmanii, ae 96 403-405 ee en 391, 392 — gleasonii, 373, — gnaphalifolia, 3 —gnaphaliifolia, 363, 378-380, 384, 385, 390-392 — var. platyphylla, : Sei subsp. ned ve 428 Vernonia gracilis subsp. tomentosa, 411 — grisebachii, 401 — harrisii, 364, 366-368 — havanensis, 365, 380, 399-402 — hieracioides, 365, 380, 400, 401 — 8 cubensis, 399 — icosantha, 381 — oo 407 .an ngustifolia, 407 var. obtusifolium, 407 — intonsa, 378 — Jenssenn, 364, 380, 386 — lanuginosa, 409 — leonis, 38: — leptoclada, 379, 380, 384-389, 406 —linguaefotia ; — nee 409 — longifolia, 373 390 seen ie 368, 381, 401, 402 — menthifolia var. grisebachi, 401 — microphylla, 398, 39 — moaensis, 385 — montana, 397 — neglecta, 385 —nematophylla, 393 nis, 401 — pallescens, 362, 368, 408, 409 — parvuliceps, 378, 379 — permollis, 378 —phyllostachya, 3 Se 363, a 380, 385, 386, 388, — awe phyla, < ee 2 oe 371, 372, 403 reducta, 3 ~ potrerillons, 407 — praestan 71 — punctata, 373 — purpurata, 365, 380, 403 398 ra, 395 rma angustifolia, 375, 393, 395 ipida. 364, 366-37 JOURNAL OF THE ARNOLD ARBORETUM [voL. 59 Vernonia rigida var. sagraeana, 407 —— var. valenzuelana, 407 — ee en 407 — saeplum — sagraeana, 365, —— var. angusticeps, 407 — scorpioides, oe, 383, 409, 410 380, 405-408 — segregata, 363, 370, 371, 380 — semitalis, 384 — sericea, 364, 374-377, 379, 395 — — subsp. racemosa, 375 — — — var. aneusitolias: 375 — shaferi, 3 — sintenisii, 373 — sprenge yaa, 365, 396, 404, — stenophylla, 363, 375, 376, 393. -395 9 — tournefortioides, 409 — tricephala, 411, 412 — trinitatis, 361, 363, 382, 383 — tuerckheimii, 365, 396-399 — urbaniana, 364, 381, 386, 387 vahliana —v alenznelana, 407 —v se 381 — venusta, 375 — seria 364, 366, 368, 369 4 — ana: 365, 381, 405, 406 — wrightii, 365, 381, 384, 385, 406 — yunquensis, 364, 381, 386, 389, 390 Vernonias (Compositae), A Revision of the West Indian, 360-413 Vigna luteola, 65 Villadia, 200, 219 ae creeper, 63 Vitaceae, 63 Waltheria, 65 — indica, 65 Welfia georgii Wendlaudiclia polyclads, 123 West Indian blackth West Indian a een A Revision of the, 360-413 WHEELER, ELISABETH F., RicHarp A. Scott, and Erso S, BarcHoorn. Fos- 1978 | sil Dicotyledonous Woods from Yel- lowstone National Park, II, 1-31 Wild See 63 Wild poinsettia, 63 WILDER, GEORGE J a New . Two New Species and Subgenus of Cyclanthaceae, 74- ae aee 287, 289 Woman’s tongue, 60 Wood Anatomy d Phylogeny of Paeonia Section Moutan, 274-297 Xanthorrhiza, 291 Xanthorrhoea, 129-— ae Xanthorrhoeaceae, 129 , 132 Liliaceae sensu ea Generic Limits in the, 129-155 32 — sect. Euxerotes ser. Fasciculatae, 138, 40 —— ser. Glomeratae, 138 — — ser. SG aanionie, 138, 140 — sect. Schoenoxeros, 137 INDEX 429 Xerotes sect. Xerotes, 138 ) Xylem Anatomy of Hibbertia (Dillenia- ceae) in oo to Ecology and Evo- lution, Yellow elder, 6 Yellowstone ae Park, Fossil Di- pace Woods from, II, 1-31 Ylang-ylang, 65 Yucca aloifelia, 57, 67 mia, 50 ae llia, 188 Tawnichelincens 170, 171 xylum fagara, 65 Zygophyllaceae, 192 JOURNAL oF tHe ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 59 1978 Dates of Issue No. 1 (pp. 1-104) issued 24 January, 1978. No. 2 (pp. 105-196) issued 3 May, 1978. No. 3 (pp. 197-310) issued 26 July, 1978. No. 4 (pp. 311-430) issued 15 November, 1978. Contents of Volume 59 Fossil Dicotyledonous Woods from Yellowstone National Park, ELISABETH F. WHEELER, RICHARD A. SCOTT, and ELtso S. BARGHOORN Xylem Anatomy of Hibbertia (Dilleniaceae) in Relation to Ecology and Evolution. WILLIAM C, DICKISON, = M. Rury, and G. LEDYARD STEBB An Ethnoflora of ee Island. Col lice County, Florida. DANIEL F. AUSTIN and Davip M. MCJUNKIN . Dates of Publication of Sargent’s Silva of North America. PAUL D. SORENSEN . Two New Species and a New Subgenus of Cyclanthaceae. GEORGE J. WILDER Statement of Ownership . Form of the Perforation Plates j in the Wide Vessels of Metaxylem in Palms. Larry H. KLoTz . Generic Limits in the Meroteac- (Liliaceae sensu lato). P. F. STEVENS Rhizophora in Australasia. — Some Clarification of Taxonomy and Distribution. . TOMLINSON , : The Potamogetonaceae i in the Southeastern United States. ROBERT R. HAYNES . The Nodal Anatomy of Myrothamnus flabellifolius (Myrothamnaceae): Another Example of a “Split-lateral’ Condition. CHRISTIAN PUFF The Genera of Crassulaceae in the. Southeastern United States. STEPHEN A. SPONGBERG The Genus Phyllocladus (Phyllocladaceae). HsuAN KENG Wood Anatomy and Phylogeny of Paeonia Section Mout JOSEPH M. KEEFE ne MAYNARD F. MOSELEY, JR. . 1978 103 170 192 7 249 A Spectacular Buchnera (Scrophu- lariaceae) from Colombia . PHILCOX . Pollen of Tropical Trees. I. Tiliaceae. E. MARTINEZ-HERNANDEZ, P. FERNANDEZ, and S. LOzAN NO. Generic Limits in the Tribe err (Ericaceae), and Its Position in the Rhododendroideae. BRUCE A. BOHM, ScoTT W. Brim, RICHARD J. HEBDA, and P. F. STEVENS . Lumnitzera rosea (Combretaceae ) — Its Status and Floral Morphology. P. B. TOMLINSON, J. S. BUNT, R. B. PRIMACK, and N. C. DUKE Transfer of the Brazilian Trixis eryngioides to Perezia (Compositae, Mutisieae ). JORGE VicToR Criscr and CLODOMIRO MARTICORENA . A Revision of the West Indian Vernonias (Compositac). STERLING C, KEELEY Index 342 360 414 Preliminary Announcement Thirteenth International Botanical Congress Sydney, Australia. 21-28th August, 1981 The Programme will consist of 12 sections — molecular, metabolic, cellu- lar and structural, developmental, environmental, community, genetic, systematic and evolutionary, fungal, aquatic, historical, and applied botany. There will be plenary sessions, symposia, and sessions for sub- mitted contributions (papers and posters). Chairman of the Programme Committee:—Dr. L. T. Evans. Field Trips will include visits to arid and semi-arid regions, eucalypt forest, rain forest, heath, coastal vegetation (e.g. Great Barrier Reef, mangroves) etc., and specialist trips. Chairman of the Field Trips Com- mittee: —Prof. L. D. Pryor. First Circular, containing details, will be mailed in August, 1979. Send your name and full address, preferably on a postcard, to ensure your in- clusion on the mailing list. Enquiries should be sent to the Executive Secretary, Dr. W. J. Cram. Congress address — 13th I.B.C., University of Sydney, N.S.W. 2006, Australia. Sponsored by the Australian Academy of Science. Journal of the Arnold Arboretum October, 1978 CONTENTS OF VOLUME 59, NUMBER 4 Generic Limits in the Tribe Cladothamneae (Ericaceae), and Its Position in the Rhododendroideae. BRUCE A. BOHM, ScoTT W. BRIM, RICHARD J. EBD A; and P.-P.-STRVENS< “peo a Lumnitzera rosea (Combretaceae ) — Its Status and Floral Morphology. P. B. TOMLINSON, J. S. BUNT, itch cPRIMACK.cand N.C. DURES aimee are ce nae Transfer of the Brazilian Trixis eryngioides to Perezia (Compositae, Mutisieae). JORGE VICTOR CRISCI and CrEODOMIROcMARDICORENA.. occ co spear) oka ee nner a ie pale ena A Revision of the West Indian Vernonias (Compositae). STERLING: G@SKBELEY:-< 52-5 oes -.r eT es pene (oF) Index . : ; g : 3 : : : ; : F teats [al SSS Volume 59, Number 3, including pages 197-310, was issued July 26, 1978.