THOMAS B. CROAT, 1994 33

Taxonomic Status of Neotropical Araceae

Thomas B. Croat Missouri Botanical Garden P.O. Box 299 St. Louis, MO 63166-0299 USA

INTRODUCTION of Araceae in South America is spotty, with some areas being relatively well known and The Araceae, with 106 genera and rough­ others, especially Andean areas of the con­ ly 3,200 species is nearly world wide in tinent, being very poorly known. Some ar­ distribution. There are two major centers eas for which floristic surveys have been of species diversity, in tropical Asia and conducted in the past thirty years or so are tropical America, with nearly an equal num­ relatively well known. This is especially true ber of indigenous genera, 43 for Asia and of Venezuela where extensive work has 36 for America (Croat, 1979). Of these to­ been done, especially by G. S. Bunting tals 33 are endemic to the American tropics (Bunting, 1975, 1979, 1986, 1988, 1989) but while 32 are endemic to Asia. Africa, a less also by Croat & Lambert (1986). The Ven­ important center of species diversity, has ezuelan flora contains 266 species with an only 19 indigenous genera with only 12 of additional 25 subspecies or varieties. them endemic. While the Paleotropics has The Guiana region is relatively well more genera than the Neotropics (60 versus known at least in part because it is relatively 36) the latter area contains roughly two­ species-poor rather than because of the ex­ thirds the species of the world's Araceae. tent of the collecting effort. Suriname was, Our level of knowledge of the systematics until recent years, the only part of the of the neotropical Araceae varies greatly Guianas that received much attention in from area to area, owing largely to recent regard to Araceae, this largely owing to the revisionary work or to the interest and area work of Jonker-Verhoef &Jonker (1953) in concentrated on by particular workers, e.g. Suriname. More recently, the whole region G. S. Bunting in Mexico (Bunting, 1965) is receiving more attention owing to work and Venezuela, Croat in Panama and Cen­ on the "Flora of the Guianas" project and tral America (Croat, 1978, 1983, 1986a, to the Araceae treatment being carried out 1986b, 1988a, 1991) and Croat & Grayum by Croat and collaborators. In addition, in Costa Rica. Central America is much more Bunting (in press) has completed the Ar­ well known than South America, largely aceae treatment for the Flora of the Vene­ owing to the more prolonged effort byaroid zuelan Guyana, the Venezuelan counter­ taxonomists in the region, but also owing part of this flora. to the fact that some parts of the area are Another example of a relatively well much less rich in species per unit area than known area is the state of Bahia in Brazil many parts of South America, especially the where Simon Mayo and other members of species-rich northwestern part of that con­ the Kew Garden staff, especially R. Harley, tinent (Croat, 1992). Most of the earlier work have made a number of expeditions. Mayo in Central America was undertaken by P. C. has also prepared a checklist for Brazil and Standley in a series of floristic works (Stan­ any reference to the number of species in dley, 1927, 1928, 1930, 1933, 1937, 1944). Brazil for any genus discussed in this paper Species diversity generally increases in relies heavily on his unpublished work. Central America as one approaches South Mayo has also worked closely with many America (Croat, 1986a, 1986b). Knowledge Brazilian botanists to encourage their par- 34 AROIDEANA, Vol. 17 ticipation in work with Araceae of Brazil fact that the species inhabiting the Amazon (see Aroideana for a discussion of these lowlands are in general wide-ranging, often efforts) . common species. Despite the presence of Parts of southern South America are by a few endemic species near the mouth of now also well known, where floristic treat­ the Amazon, e.g. Philodendron ecordatum ments for Argentina (Crisci, 1971) and Par­ Schott, the vast Amazonian region lying be­ aguay (Croat & Mount, 1988) have been tween the Atlantic coast and the foothills completed. In addition, a floristic treatment of the Andes has moderately few, mostly has been completed for the state of Santa Wide-ranging species. Species diversity in­ Catarina (Reitz, 1957). creases dramatically as one approaches the The published Flora of Peru (Macbride, foothills of the Andes in the west. Species 1936), though falling short of giving an ac­ occurring on the lowermost slopes of the curate picture ofthe species count for Peru, Andes tend to range widely in a north-south does come close to indicating the number direction, often from Colombia to Bolivia of species actually described for Peru, since, and thus tend not to be endemic. However, except for Anthurium sect. Pachyneurium some of the species of this region are cur­ (Croat, 1991), few groups have had many rently believed to be endemic. The degree species described from Peru since that pub­ of endemism increases as elevation increas­ lication. A more accurate accounting for the es on the slopes of the Andes and as the number of species of Araceae in Peru is terrain becomes more dissected with river published in the "Catalogue of the Flow­ valleys (Croat, 1994). ering Plants and Gymnosperms of Peru" To the east of the Amazon basin, es­ (Croat, 1993). Though not a thorough re­ pecially in the Guiana Highlands and in vision ofthe species occurring in Peru, this eastern Brazil, from the state of Bahia south list takes into account all species of plants almost to Uruguay, the rate of endemism is described for Peru as well as all species much higher. Nearly all the species occur­ represented only by identified herbarium ring in this region are endemic to eastern specimens which were able to be verified Brazil, and few range into the Amazon ba­ by experts for each family. The checklist sin. For example, five of the eight Anthur­ contains 210 species of Araceae for Peru ium species from Santa Catarina are en­ but it does not include any unpublished demic to eastern Brazil, including A. acu­ names and many species remain to be de­ tatum N. E. Br., A. gaudichaudianum scribed. Kunth, A. harrisii (Grah.) Endl., A. lacer­ While there is no completed Araceae dae Reitz, and A. pilonense Reitz. treatment for the flora for Ecuador, a re­ Though many species have been de­ cently published checklist for the Amazo­ scribed from eastern Brazil, by early aroid nian lowlands (Renner et al., 1990), which specialists, including Schott, K. Koch, and lists 92 species of Araceae (a few of them Engler, based on the early collecting efforts undescribed), gives some indication of the by botanists such as Glaziou, Regel, Riedl, species diversity of that part of Ecuador. and others, these areas remain poorly Unfortunately the Amazonian lowlands known, principally because of the taxo­ represent one of the most species-poor por­ nomic complexity in such groups as An­ tions of the country if its area is taken into thurium section Urospadix, which domi­ account, owing to the widespread nature nates the area. Perhaps the only group of of the species in that zone. aroids well known in the region are mem­ The above-mentioned treatments, though regional in nature, help to give an indica­ bers of the recently revised Philodendron tion of species diversity on the continent subgenus Meconostigma (Mayo, 1991). of South America, and any reference to spe­ The truly temperate parts of the conti­ cies counts for any genus is based on them. nent are devoid of aroids, and the subtrop­ The lowland Amazon basin is also rela­ ical portions of the continent, while con­ tively well known, principally owing to the taining a number of small, frequently en- THOMAS B. CROAT, 1994 35 demic genera in the tribe Spathicarpeae, Cauca Department, the department of An­ are also relatively species-poor. tioquia, and the department of Cundina­ Species diversity is high throughout the marca, especially around Bogota, were well extent of the South American Andes but collected in the late Nineteenth Century by especially along the northwestern slope ex­ collectors such as Lehmann, in time to have tending from Choco Province in Ecuador, their material included in the revisions of as well as on both the eastern and western both Schott (1860) and Engler (1905), many slopes in the Andes in the region of the areas have not been collected until recent equator and on the eastern slopes of the times. Andes in Peru. Species diversity is also rel­ Pichincha Province and a few other areas atively high in the Cordillera de Merida of of Pacific coastal Ecuador were well col­ western Venezuela, but remarkably less so lected by L. Sodiro (1901a, 1901b, 1903a, in the northern part of the Eastern Cordil­ 1903b, 1905a, 1905b, 1905c, 1906, 1907, lera of Colombia and on the entire western 1908) and 257 species (including 281 taxa) flank of the Eastern Cordillera in Colombia were described. Despite this, the region (Croat, 1992). Species diversity is also rel­ remains poorly known, largely because of atively low in the Central Cordillera of Co­ the inability to locate and study the widely lombia. In all cases high species diversity scattered and poorly documented collec­ is correlated with high precipitation and tions of Sodiro. Some Sodiro specimens are with the absence of prolonged dry seasons. deposited in European herbaria (Croat, Species richness is greatest between sea 1989), but most collections are depOsited level and middle elevations up to about in the poorly curated herbari um (Q PLS) of 1,500 meters. While some species may range the monastery in Quito where Sodiro orig­ up to about 3,750 m (Anthuriurn caucan­ inally worked. The collections may not be urn Engl.; Colombia: Antioquia: Villa Ar­ borrowed and the conditions in the her­ teaga, Hodge 7045), diversity drops off dra­ barium make their study there very difficult mati cally above 2,000 m elevation. Season­ (Croat, 1989). Nevertheless, some recent ally dry areas, such as the central plateau attempts at revisionary work have taken of eastern Brazil and the lower Amazon ba­ place in Ecuador in Pichincha Province. sin, are relatively species-poor, as are the Floristic surveys of the Araceae have been generally treeless llanos of Venezuela. made of the Reserva ENDESA on the west­ Endemism is also especially high in the ern slopes ofVolcan Pichincha. This work, Andes of western South America, including begun by Jimena Rodriguez de Salvador the eastern range of the Andes which ex­ while a student at the Universidad Catolica tends into Venezuela. Endemism is also in Quito, was subsequently augmented by high in the Guiana Highlands and in parts investigations of my own, and is nearing of North America, especially in Mexico and completion and is expected to be pub­ in lower Central America, in Costa Rica, and lished in 1995. The paper will be entitled Panama. For example, Mexico, with 41 taxa Anthllrillln of Reserva ENDESA (Pichincha of Anthuriurn has 26 endemic taxa. Gua­ Province , Ecuador). temala has only three endemic species. Both Other Andean regions are being studied Honduras and Nicaragua have a single en­ on a local basis. With the assistance of Jeff demic species each. Costa Rica has 68 taxa Lake of Grinnell College I have completed with 22 species endemic, and Panama has a revision of the Araceae for the Flora of La 150 with 82 species, 55 percent of the total, Planada, a natural history reserve in Narino. considered to be endemic. This revision is also expected to be pub­ Because of the high rate of endemism lished soon. and the very high speciation in many parts With the assistance of another student, of the Andes, our taxonomic knowledge of Dorothy Bay, I am also working on a revi­ all but a few areas of the Andes is very poor. sion of the Araceae of Bajo Calima. This Although selected areas of the Colombian flora is an exceptionally rich one with many Andes, such as the region of Popayan in new species, and despite about six separate 36 AROIDEANA, Vol. 17 collecting trips made to this region over the publication of the book, currently the past ten years it remains still poorly scheduled at Kew during fiscal year 1994. known. Bay will work on the flora as a part of a Ph.D. dissertation at St. Louis Univer­ ALLOSCHEMONE sity. The project is expected to take three This genus with two species in tribe Mon­ additional years to complete. stereae is a hemiepiphytic appressed This has been a general overview of the climber endemic to the western Amazon distribution of the neotropical Araceae. The basin in Brazil and Bolivia. Its placement remainder of this paper will deal individ­ has been somewhat controversial. Madi­ ually with different genera to discuss the son (1976) considered it to be a member of general level of knowledge of each genus, the genus Scindapsus, otherwise known the predicted number of species in each only from the Old World tropics but later genus, and what is currently known of their reestablished it as a distinct genus (Madi­ phytogeography. A review of the ecology son, 1977). It was also treated as a distinct and life forms of the genera is summarized genus by Grayum (1990) and Bogner & in Croat (1988a). Nicolson (1991). Alloschemone occiden­ Discussions of the suprageneric place­ talis (Poeppig) Engl. & K. Krause is still ment of the genera to follow are based on poorly collected and is known only from a the latest system of classification, an as yet few localities (Croat, 1985b). It has been unpublished work by S. Mayo, ). Bogner, collected on the Rio Solimoes at Tefe, on and P. Boyce (in prep.). This work reflects the Rio Madeira at Humayta, along the Rio many of the ideas expressed by M. Grayum Marie, a tributary of the Rio Negro, on the (1984, 1990) and by Bogner & Nicolson Rio )apura near Maraa the Department of (1991). It also takes into account unpub­ Pando, province of Frederico Roman along lished molecular work by French et al. (in the Rio Negro. press) on chloroplast DNA and an exten­ The genus is still poorly known in many sive cladistic analysiS of the family by Mayo respects and living material should be ac­ et al. (in prep.). Thus, though perhaps not quired for additional study. the final word on the suprageneric classi­ fication of the family, it constitutes a con­ ANAPHYLLOPSIS sensus of modern thinking on the classifi­ cation of the Araceae. For a discussion of This lasioid genus (tribe Lasieae; sub­ all but the most recent revision, including tribe Dracontiinae) of three terrestrial, rhi­ that of Hotta (1970) see Croat, 1990. zomatous species is endemic to northern In order to avoid confusion with names Brazil (Para) and the Guianas. One of the used in earlier works brief mention should three species, A. americana (Engl.) A. Hay be made of the major conclusions reached from French Guiana and Surinam, had long in the work being prepared by Mayo, Bog­ been considered a Cyrtosperma, but Hay ner & Boyce. These are as follows: the fam­ (1988) found that it differs from Asian ily has been divided into two major groups, members of Cyrtosperma by having fenes­ the Protoaroids consisting of subfamilies trate leaf blades, as well as both prophylls Gymnostachyoideae and Orontioideae and and cataphylls on the rhizome. In addition the True Araceae, conSisting of subfamilies to placing this species into the new genus Anaphyllopsis, he described two new spe­ Calloideae, Pothoideae, Monsteroideae, cies, A. pinnata A. Hay, from the upper Rio Lasioideae, and Aroideae. The most signif­ Negro in Amazonas State of Brazil, and A. icant change with earlier systems as out­ cururuana A. Hay, from the upper Rio Ta­ lined in Croat (1990) is that all genera with pajos along the Rio Cururu in Para State of unisexual flowers (previously incorporated Brazil. variously in Colocasioideae, Philodendro­ None of the species are particularly well ideae, and Aroideae) are now included in known and the two most recently described Aroideae. For further details, one must await are known only from a single collection THOMAS B. CROAT, 1994 37

each. Certainly more collections should be in synonymy, many regions were poorly made to determine the true nature of mor­ collected prior to Schott's or Engler's re­ phological variability and geographical visions, and thus have few species de­ range. scribed. Interestingly, these same areas, no­ Two other species previously thought to tably those along the Pacific slope of west­ belong to Cyrtosperma, C. wurdackii Bun­ ern Colombia, probably are the most spe­ ting and C. spruceanum (Schott) Engl., cies-rich. For this reason, substantial were placed by Hay (1988) in Urospatha numbers of new species exist. In some ar­ and Dracontium, respectively. Thus, Cyr­ eas which have now been well collected, tosperma is eliminated as a neotropical ge­ such as the region of the lower Calima River nus and is restricted to the Asian tropics. basin, a region commonly referred to as "Bajo Calima," the majority of species ap­ pear to be new to science. The region is ANTHURIUM largely a transition zone between Tropical There is now agreement as to the place­ wetforestand Tropicalforestand has yield­ ment of Anthurium in tribe Anthurieae in ed, so far, 133 species including 63 species subfamily Pothoideae. of Anthurium and 40 species of Philoden­ The genus Anthurium is the largest ge­ dron, with only a few species yet to be de­ nus in the family with an estimated 1,000 termined and most appearing to be new to species. Admittedly, the margin of error in science. The Pacific slope also appears to counting the number of species in this ge­ have a high rate of endemism, especially nus is great. Indeed, the margin of error in between different life zones and at different estimating the number of species existing elevations. For example, in the department for Anthurium and Philodendron, the two of Narino, three study sites, all located rel­ largest genera in the family, is greater than atively near one another and all located the total number of all other species in all within a Premontane forest life zone (ac­ other neotropical genera combined. While cording to the Colombian life zone map), the number of species of Anthurium in contained relatively few species which were Central America (221 spp.) and some parts shared between sites (Croat, 1992). of South America, e.g. Venezuela (68 spe­ Probably regions of Tropical forest, which cies including seven varieties or subspe­ are restricted to northwestern South Amer­ cies), the Guianas, the Amazon basin, and ica, will prove to be the most species-rich Paraguay, is well known, other areas, es­ of all. Some parts of this life zone, es­ pecially eastern Brazil and the Andean re­ peciallythe area south of Quibd6, have been gion of western South America are doubt­ relatively well collected, but the material ful. Mayo has produced a checklist for Bra­ has not yet been analyzed for a determi­ zil containing a total of about 100 species, nation of its species diversity. eight of which are new to science. Still, The situation in Ecuador is even more owing to the taxonomic uncertainty of many complex regarding exact numbers of spe­ species of section Urospadix, a complex cies, especially in Carchi and Esmeraldas group of species which are difficult to sep­ Provinces in northwestern Ecuador. Much arate, the total is likely to be higher than of Ecuador is probably as rich per unit area that projected by Mayo. as the Pacific slope of Colombia. Both slopes It is in the Andean region of the conti­ of the Andes in Central Ecuador in Imba­ nent where specific counts of species are bura, Pichincha, Los Rios and Cotopaxi, Bo­ most difficult to obtain. While a few areas, livar, Tungurahua, Chimborazo and Canar, notably, the departments of Antioquia and especially at middle elevations on the Pa­ Cundinamarca as well as the region of Po­ cific slope and at all but the higher eleva­ payfm in the department of Cauca, have had tions on the Amazonian slope, are rich in many species described and probably even species. The same is true of the Amazonian have a number of species (especially at the slopes of Loja and in the principally low­ higher elevations) which must be placed land provinces of the so called "Oriente," 38 AROIDEANA, Vol. 17

Sucumbios, Napo, Pastaza, Morona-Santia­ cies to occur. The same is essentially true go, and Zamora-Chinchipe. Renner et al. for the intermountain valleys in Central (1990) report 32 species for this region but Peru. Only in the foothills of the Andes on the actual number is certainly much higher. the eastern slopes do any significant num­ Much of the middle elevations of the north­ ber of species occur but species diversity eastern slopes of the Andes as well as the here does not come close to matching that southeastern slopes remain unexplored. in Ecuador. The species occurring there, Recent surveys of the Cordillera del Con­ including A. breviscapum Schott, also tend dor, along the Peruvian border, indicates a to be much more widespread but there are high rate of endemism. The same is true of also significant centers of species richness the Serrania de Cutucu further to the north. and endemism such as the areas around The Amazonian lowlands, though con­ Tingo Maria in Huanuco Department and taining some new species, have mostly rath­ at Tarapoto in San Martin Department. er wide-ranging species in common with The species distribution in the Amazon other areas of the Amazon basin. An ex­ lowlands of Peru mirrors that in the low­ ample of the area, now reasonably well col­ lands of Ecuador, tending to be wide-rang­ lected is that of Jatun Sacha which has a ing species, many of which also enter Bra­ flora of at least 50 species, including 25 zil. Examples include Anthurium atropur­ species of Anthurium. Of this total only pureum R. Schultes & Maguire, A. oxycar­ three species are believed to be endemic pum Poeppig, A. lore tense Croat, and A. to Ecuador. rubrinervium (Link) G. Don. In contrast to the remainder of Ecuador, The Flora of Peru (Macbride, 1936) is of the Pacific lowlands of western Ecuador relatively little value. The treatment con­ south of Esmeraldas are either mostly de­ tains 55 species of Anthurium, with eleven forested or seasonally dry and contain rel­ names being synonyms of other species atively few aroid species. This is particu­ treated and an additional two names of larly true of the southwestern part of the plants which do not occur in Peru. The re­ country in El Oro and Loja. maining names are probably correct, but The region of northwestern Peru, like that may represent as few as one-third of the of southeastern Ecuador is relatively dry species in the country. The up to date with a depauperate aroid flora. Species di­ checklist for Peru (Croat, 1993) has 76 taxa versity is substantially greater in the moun­ but many remain to be described. tains of northern Peru in the state of Ama­ Bolivia is relatively poor in Anthurium zonas, but the region has yet to be well species but has a significant number of en­ collected. Most collections existing are the demic species. No aroids occur in the dry result of University of California anthro­ eastern portion of the country, but a num­ pologist Brent Berlin and his workers who ber of species of Anthurium occur in the collected in the region of the Rio Cenepa relatively humid intermountain valleys be­ and the Rio Santiago. This area is home to ginning in Nor Yungas east of La Paz and the rare and poorly collected genus Filar­ extending to the lowlands of Beni and to um. The drier highland regions of Peru are a lesser extent to the east in Cochabamba. relatively poor in species. One particularly Paraguay has a depauperate aroid flora variable species ranging throughout the with only 16 species, including two species highlands from southern Peru to Central of Anthurium. Argentina has 17 species, Ecuador is A. dombeyanum Brongn. ex including only one Anthurium, A. para­ Schott. The middle elevations are particu­ guayense Engl. That species ranges north larly rich in members of Anthurium sect. to Bolivia and southern Brazil. Uruguay and Pachyneurium (see discussion of this sec­ Chile have no Anthurium (indeed Chile tion below). has only the pantropical genus Pistia) and The entire Pacific coastal region of Peru, Uruguay has only four species in three gen­ much like the southwestern part of Ecuador era). In contrast to the northern hemi­ is arid, indeed too arid for many aroid spe- sphere where several genera, including Ar- THOMAS B. CROAT, 1994 39 isaema, Calla, and Lysichiton, range well well represented in Panama with 50 species north of 45 degrees latitude, few Araceae but has an even greater concentration in range very far into temperate South Amer­ northwestern South America, especially in ica. The southernmost is Pistia, found at 35 the wettest forest types at lower elevations. degrees latitude in the basin of the Rio de It is particularly well represented in Trop­ la Plata near Buenos Aires. ical forest and Tropical Premontane forest Sections of genus Anthurium are not in Choco Department of Colombia. For ex­ equally distributed. The sectional classifi­ ample, 27 species are already known from cation discussed here, somewhat modified one small area at Bajo Calima near Bue­ from that of either Schott (1860) or Engler naventura in the Department of Valle, in (1906) is outlined by Croat & Sheffer (1983). transitional zone between Tropical wet Some sections, such as section Pachy­ and Pluvial forest. neurium, are widespread, ranging Relatively few species in the section oc­ throughout most of the range of the genus, cur in the Amazon basin. Most of the spe­ with the exception of the wettest part of cies in the group are new to science. Be­ the range, such as in northwestern Colom­ cause of their generally similar appearance, bia. Most members of the section, especial­ the species are easily overlooked or even ly series Pachyneurium, are most common mixed up by collectors. Probably no sec­ in areas of seasonally dry forest. The center tion has such a high percentage of species of diversity for sect. Pachyneurium is in new to science, perhaps owing to their in­ Ecuador, where 36 species occur, and in conspicuous nature (by aroid standards). Peru with 28 species. These countries are For example, in Panama, 39 taxa (of a total also the centers of endemism with 23 and of 80 new taxa described) are members of 13 species respectively. Central America is sect. Porphyrochitonium. The greatest con­ a second center of diversity for sect. Pa­ centration of section Porphyrochitonium is chyneurium, especially Costa Rica and on the Pacific coast of Colombia, most ap­ Panama, with 23 and 22 species (seven and pear to be endemic and undescribed. Schott five species endemic) respectively (Croat, (1860) described only one member of the 1991). A portion of this section, including section, A. scherzerianum, and Engler all the cordate species, is restricted to Costa (1906) described only 17 additional spe­ Rica and Panama. Another group, series cies while Sodiro (1901a, 1901b, 1902, Muitinervia, is nearly restricted to Ecuador 1903a, 1903b, 1904, 1905a, 1905b, 1906, with only a few outlying species occurring 1907, 1908) described only 12 species. in Colombia (A. carchiense Croat, A. na­ Though most of the species of section Por­ paeum Engl., and A. narinoense Croat) or phyrochitonium described by Engler were Peru (A. soukuPii Croat and A. ottonis K. included in his section Urospadix, a few Krause). were included in other sections, including Section Calomystrium is also wide­ Tetraspermill1n (A. margaricarpllm Sodi­ spread, but with the vast majority of its spe­ ro) and in Xialophyllium (A. angosturense cies occurring between Costa Rica, Vene­ Engl., A. filiforme Engl., and A. tenllinerl'e zuela, and Peru, especially at middle ele­ Sodiro). A few species placed in sect. Por­ vations on both slopes of the Andes. A sig­ phyrochitonium in the Central American nificant percentage of the species, perhaps revision of Anthurium are now believed to as much as 50 percent is new to science. be aberrant members of sect. Calomys­ The percentages of new species are partic­ trium. This group of species occur in or ularly high in Panama, with 15 of 18 species along streams and has mostly slender leaves in the section new to science (Croat, which are not noticeably punctate on the 1986b). The percentages in South America, lower surfaces. The group consists of A. especially in Colombia where the diversity amnicola Dressler, A. antioquiense Engl., is greatest, could be as high or higher. A. antropbyoides Killip, A. rupicola Croat, Section Porphyrochitonium has a range A. sytsmae Croat, and A. werJlii ined. Croat similar to that of sect. Calomystrium and is sp. nov. Because of their short internodes 40 AROIDEANA, Vol. 17

and slender leaves they were confused with Central American species of Anthurium in sect. Porphyrochitonium but breeding general are not closely related to those of studies have shown that they do not inter­ South America. For example, of the 221 breed with other members of that section Central American species, only two, A. but instead breed readily with sect. Calo­ scandens (Aublet) Engl. and A. gracile mystrium and are believed to be specially (Rudge) Schott, are truly wide-ranging, i.e. adapted members of that group_ Perhaps from Mexico to Brazil. While quite a large they independently evolved slender leaves number enter northern Colombia, or barely which survive much easier in rapidly mov­ enter Panama from South America, only ten ing water. additional species from Middle America Section Digitinervium is a small group enter into non-Colombian portions of South with leathery, pli-veined, glandular-punc­ America (Croat, 1986a). tate leaf blades and scalariform venation. It Another example pointing out the iso­ ranges from Costa Rica to Venezuela and lated nature of Central America is the Peru, particularly at middle to high eleva­ breeding behavior of Anthurium sect. Pa­ tions. The section is centered in the moun­ chyneurium. While Central and South tains of Ecuador, and most species were American species respectively will readily described by Sodiro (loc. cit.). Only a few interbreed within their own group, few species exist in this section and relatively Central American species will cross with few of them are believed to be new to sci­ South American species (Croat, 1991). ence. In contrast to those groups mentioned Section Cardiolonchium is also heavily above, which have the majority of their spe­ concentrated in the Andean region at low cies in the Andean regions, section Uros­ to middle elevations, especially on the padix is almost exclusively known from western slopes of the Andes. A few species eastern South America and is most heavily occur in the lower part of Central America concentrated on slopes of the eroded pla­ and one species, A. costatum K. Koch & teau of central and southern Brazil. Here it Bouche occurs in the coastal range of clearly represents the most dominant grou p northern Venezuela. One species, A. rub­ of Araceae. Mayo reports approximately 60 rinervium (Link) Don (probably to in­ species of sect. Urospadix, and there are clude both A. polyrrhizum K. Koch & Au­ numerous new species in the region. Some gustin and A. aliena tum Schott) is wide­ members of the section range as far north spread in the Amazon baSin, ranging from as the Guiana Highlands, and at least one lowland Ecuador and Peru to the Guianas. member believed to be in the group, A. Section Cardiolonchium, with frequently lilacinum Bunting, even occurs in the Cor­ velvety, often discolored leaf blades, was dillera de la Costa in northern Venezuela. represented among the earliest introduc­ Another unique section, Chamaere­ tions into European glasshouses. Many spe­ pium, represented only by A. radicans K. cies are narrowly restricted and now rare. Koch & Haage, is restricted to eastern Bra­ Most have probably already been described zil. Though now represented in cultivated but a few new species remain. collections in many places, it is apparently Despite having velvety leaves, a few Mex­ rare in the wild. ican species appear to be unrelated to Sec­ Few groups of Anthurium are restricted tion Cardiolonchium. They include A. clar­ to Central America, but one as yet unnamed inervium Matuda, A. lezamae Matuda, and group possibly deserving sectional status, A. leuconeurum Lem. These species will represented by plants with cordate, glan­ hybridize with other Mexican species in­ dular-punctate blades, is restricted to Me­ cluding species in unrelated groups such soamerica and centered in Mexico. Exam­ as A. pedatoradiatum Schott but not with ples include A. lucidum Standley ex Cardiolonchium species in other areas. This Yuncker, A. chiapasense Standley, and A. again points out the isolated nature of the verapazense Engl. This group, as well as Mexican species of Anthurium. Indeed, many other endemic Mexican species, ap- THOMAS B. CROAT, 1994 41

pears to share little in common with other Peru, with greatest concentrations of spe­ species from other parts of Central America. cies in the Andes of Colombia and Ecuador. The A n thll rill In flora of Costa Rica and Pan· This group probably includes many new ama, though also highly endemic, at least species. It almost certainly includes two shows signs of relationships with species disparate groups which will have to be rec­ from South America. Most of these species ognized, one represented by A. cuspida­ occur in groups which are much better rep­ tum Master and its allies and the other A. resented in the Andes of South America. panduriforme Schott and A. corrugatum The section Polyphyllium, consisting of Sodiro. two scandent species, A. flexile Schott and Section Belolonchium sensu Schott is al­ A. clidemioides Standley, is also restricted most exclusively Andean, occurring pri­ to Central America. The group is distin­ marily in cloud forests at higher elevations guished by having several internodes be­ and characterized by the widespread A. pul­ tween successive flowering internodes and chrum Engl. which usually is the dominant in having roots produced along the inter­ species at elevations above 1,000 m in many nodes (Croat & Baker, 1978). parts of Ecuador and Colombia. The section Another largely Central American group also ranges into western Venezuela (e.g. A. is section Schizoplacium. This small group, berryi Bunting, A. betanianum Croat, A. with deeply palmately lobed leaves, is ginesii Croat, A. nubicola Bunting, and A. known primarily from Mexico, the West In­ signatum K. Koch & Mathieu). Engler's cir­ dies (represented there by A. palmatum [L.] cumscription of section Belolonchium was Don) and northern Venezuela (represent­ much broader and included many species ed by A. longissimum Pittier). which occur at lower elevations, even at sea The section Semaeophyllium (deeply tri­ level. lobed blades) is concentrated in southern Section Tetraspermium is widespread, Central America and in the Andean portion owing to two species, namely A. scandens of northern South America. Despite previ­ (Aublet) Eng!. and A. trinerve Miq., but a ous work by Madison (1978), the group has number of mostly Andean species remain a number of new species, especially in South undescribed. Most species range from very America. A new revision of the section is dry areas, such as Tropical dry forest (A. currently being carried out. pohlii Eng!.) to Tropical moist forest, but a Section Xialophyllium is an unnatural few species, such as A. scandens ssp. pus­ group (Croat & Sheffer, 1983) with at least illum Sheffer and A. lacinosum Sodiro, two elements. One group, with thin, bul­ range into Premontane wet forest and Pre­ late leaves or at least prominently raised montane rain forest. tertiary veins, is represented by such spe­ Another small but widespread section is cies as A. microspadix Schott, A. colum­ Dactylophyllium, a group with palmately bianum Engl., and A. myosuroides (Kunth) compound blades. It ranges throughout Endl. Another group, with more coria­ most of the range of the genus (except in ceous, smooth leaves is represented by such Paraguay and Argentina) and is most abun­ species as A. mindense Sodiro, A. sarmen­ dant in regions of Tropical mOistforest. The tosa Engl., and A. popayanense Engl. Sec­ section is most diverse in eastern South tion Xialophyllium is primarily Andean but America and in the Amazon basin. It is rep­ ranges from Mexico to Bolivia with the resented by such species as A. eminens greatest concentration at middle to high Schott, A. 'pentaphyllum (Aublet) Don and elevations in the Andes of Colombia, Ec­ A. clavigerum Poeppig. A few species from uador, and Peru. This group probably con­ eastern South America may be new to sci­ tains a number of new species. ence. Other primarily Andean groups are sec­ A well known but insignificant group is tions Polyneurium and Belolonchium. Sec­ Leptanthurium. It is represented by A. tion Polyneurium ranges from Nicaragua gracile (Rudge) Schott, one of the most to the mountains of western Venezuela and widespread species in the genus. This is 42 AROIDEANA, Vol. 17 the only Anthurium species with a chro­ Aroideae, occurring usually as understory mosome base number of 10 and is believed plants, from near sea level to 2,300 meters. to be unrelated to any described and the The genus Asterostigma consists of only section is not closely related to any other six species, all but a few of them rather section. A few additional undescribed An­ poorly known. Those in the Andes, such as dean species probably also belong here, A. integrifolium Madison and A. pavonii but their chromosomes have not yet been Schott, occur in cool mountain valleys, of­ studied. ten on steep slopes. Only the latter occurs Two monotypic sections are very geo­ in Peru. Asterostigma riedelianum (Schott) graphically isolated. Section Chamaere­ Kuntze is endemic to the state of Bahia in pium with a single species, A. radicans eastern Brazil, while A. tweedianum Schott Koch & Haage occurs in eastern Brazil while is endemic to Santa Catarina in southeast­ A. gymnopus Griseb. in section Gymno­ ern Brazil. Asterostigma luschnathianum podium is endemic to Cuba. Neither group Schott is known from Minas Gerais and Rio appears to have any close relatives. de Janiero states, and A. lividum (Lodd.) A large number of Anthurium species re­ Engl. is the most widespread species, rang­ main unclassified as to section. Species in­ ing from Minas Gerais to Rio Grande do volved mostly consist of larger cordate­ SuI. leaved species which clearly do not belong Collecting Asterostigma poses no partic­ in any ofthe other described cordate groups, ular problem, although they are not com­ e. g. Cardiolonchium, Calomystrium, Po­ mon, and few species have been well col· lyneurium, or Belolonchium. Probably sev­ lected. Recollections are necessary, es­ eral small sections will have to be created pecially of the Brazilian species to recon­ to accommodate them. Engler placed most firm Engler's treatment. of such species known to him in section Another problem with Asterostigma con­ Belolonchium. Another somewhat smaller cerns the name A. cubense (A. Rich.) K. group which does not fit into any recog­ Krause ex Bogner, purportedly from Cuba nized section includes more or less oblong (Bogner, 1969). The species is related to to lanceolate-leaved plants lacking glan­ A. riedelianum 0. Bogner, pers. comm.). dular punctations, e.g. A. michelii Guillau­ This name is based on a mixed collection min. Both groups contain new species. with Xanthosoma cubense (Schott) Engl. Future work on Anthurium must con­ Although Asterostigma has never been col­ centrate in two areas: the Andean countries, lected in Cuba and is probably restricted especially Colombia, Ecuador, and Peru as to southern South America, the true nature well as in eastern Brazil. The problems in of A. cllbense has, in my opinion, never these two areas are very different. In eastern been determined. Brazil the majority ofthe species have been described but are still poorly defined and BOGNERA little understood. They are very similar to one another, and the type localities are very This monotypic genus (tribe Dieffenba­ much degraded. Some species may be lost chieae), represented by B. recondita (Mad­ altogether. ison) Mayo & Nicolson, is known only from In the Andes of western South America the western part of the state of Amazonas species diversity is much higher, and there on the Rio Javari. Madison (1980) initially are many undescribed or poorly known described it as a species of Ulearum. Since species. A great deal of time and space will it is in cultivation at the Munich Botanical be needed just to sort out the specimens Garden and elsewhere, it is now quite well to determine how many species there are. known. It is very similar to Dieffenba­ chieae, according to J. Bogner (pers. ASTEROSTIGMA comm.). The tribe Dieffenbachieae is This is a genus of terrestrial, tuberous placed in the Dielfenbachia Alliance in herbs in tribe Spathicarpeae of subfamily subfamily Aroideae. Little remains to be THOMAS B. CROAT, 1994 43 done with Bognera except to determine sii Bunting, C. aristeguietae Bunting, and the exact extent of its range. C. striatipes K. Koch & Bouche. All have been transferred to Xanthosoma (Madison, 1981) or by Croat & Lambert (1986). This CALADIUM is an indication of the close proximity of A modest-sized genus of 17 species (tribe these two genera taxonomically. While most Caladieae in the Caladium Alliance of sub­ species of Xanthosoma and Caladium are family Aroideae) ranging from sea level to distinctive by a suite of characters, many at least 1,000 m, this principally South appear quite intermediate. Madison (1981) American group is often confused with determined that the only definitive char­ Xanthosoma. While chiefly tuberous, and acter for separation of the two genera is in preferring semi deciduous, seasonally dry the nature of pollen presentation, with Ca­ forest, the plants are sometimes rhizoma­ ladium having pollen borne in monads and tous and caulescent as in Caladium lin­ Xanthosoma having tetrads. While work by denii (Andre) Madison. This species in­ Madison was important in defining the two habits wet forest areas in Colombia and genera, it is believed that his treatment of Panama. Species in semideciduous forest Caladium, wherein 39 taxa were synony­ often lose their leaves and go dormant dur­ mized under C. bieolor, was unrealistic. ing the dry season. Bunting (1979), for example, has recog­ The genus ranges from southwestern nized as distinct two species synonymized Costa Rica to the West Indies and Trinidad, by Madison, e.g. C. pieturatum K. Koch and Venezuela, northern Brazil (Amazonas and C. smaragdinllm K. Koch & Bouche. Para, except for the widespread C. bieolor Caladium appears to be less richly de­ [Ait.] Vent.), and Peru. Mayo reports only veloped in the Andes of western South four species for Brazil, including C. bieolor, America than in northeastern South Amer­ C. humboldtii Schott, C. maerotites Schott, ica, but numerous collections in the Ca­ and C. pieturatum K. Koch. Only C. bieolor ladium bieolor complex remain unidenti­ and C. pieturatum are reported for Peru fied. (Croat, 1993). A major center of diversity appears to be Venezuela and the Guianas. Only two spe­ CHLOROSPATHA cies, C. bieolor and C. lindenii (Andre) Madison, occur naturally in Central Amer­ This is a genus of fifteen species of un­ ica and the latter is restricted to Panama. derstory herbs in middle elevation wet for­ Caladium bicolor has been introduced est at elevations of 275 to 1,800 (2,500) widely throughout Central America. Ven­ meters. The genus is placed in tribe Cala­ ezuela has eight species, several of them dieae (Aroideae). recently described (Bunting, 1975) and still Madison (1981) provided a revision of poorly known. At least five species, includ­ Chlorospatha including ten species. Sub­ ing C. aturense Bunting, C. bieolor, C. sequently an additional four taxa were de­ humboldtii, C. maerotites, and C. pietura­ scribed by M. H. Grayum (1986, 1991). tum, occur in southeastern Venezuela on Three species occur in Panama including the margin of the Amazon baSin, and an C. eroatii Grayum, C. eroatii var. ennea­ additional species, C. se!JomburgkiiSchott, phyllum Grayum, and C. hammelianum occurs in the adjacent Guianas. The Cor­ Croat & Grayum. Four species occur in Co­ dillera de Merida also has several species, lombiaincluding C. gentryiGrayum, C. kol­ including C. eoeruleseens Bunting, C. bii Engl., C. lehmannii (Engl.) Madison, smaragdinum K. Koch & Bouche, and C. and C. mirabilis (Masters) Madison. Ec­ steyermarkii Bunting. uador is the center of species diversity with Several species treated as Caladium by six species, C. atropurpureum Madison, C. Bunting (1979) have proven to belong to besseae Madison, C. eastula (Madison) Xanthosoma. These include C. akkerman- Madison, C. eutlleuense Madison, C. dod- 44 AROIDEANA, Vol. 17 sonii (Bunting) Madison, C. ilensis Madi­ Dieffenbachia going by that name in Ven­ son, and C. longipoda (K. Krause) Madison. ezuela, the Guianas, and in Central Amer­ ica. Living material from Martinique must be studied to see if it is the same as that DIEFFENBACHIA from the Dominican Republic. The genus comprises the monotypic tribe Dieffenbachia is, in general, well col­ Dieffenbachieae (Aroideae) and has an es· lected since plants provide no particular timated 85 species, consisting of mostly obstacle, being terrestrial and readily ac­ caulescent, understory, terrestrial herbs, cessible. However, some are very large, and occasionally occurring along clearings or all are fleshy. Many are rich in oxalic acid rarely in swampy areas. It ranges from Mex· so the sap may burn the skin, and they may ico (Veracruz) to the West Indies, Trinidad, be very foul during the drying process. Some the Guianas, Paraguay, and northeast Ar· collectors avoid them for these reasons. The gentina, and is certainly the most overall chief problem with their taxonomy, how­ poorly known aroid genus in the neotrop· ever, is species circumscription. Species are ics. In Central America, where it is dra­ either highly variable or there are many matically more well known than in South species. In Panama and Costa Rica where America, there are 26 species; 20 believed they have been studied most closely, each new to science. Dieffenbachia is believed population is in some way different from to be just as diverse in South America as in the next, yet with distinct patterns of vari­ Central America, and it may have an equally ation that lead one to believe they might large percentage of species new to science. all be closely related. Helen Young, during However, more species have already been her studies at La Selva (Organization for described from South America than from Tropical Studies) in Costa Rica, found that Central America. the beetles pollinating Dieffenbachia There are seven species in Venezuela, six sometimes cross-pollinated different spe­ species in the Guianas, and at least six spe­ cies of Dieffenbachia and plants pollinated cies in Brazil. Macbride (1936) treated 11 in such manner did produce viable fruit species of Dieffenbachia for Peru, but a (Young, 1986, 1988a, 1988b). It is not yet modern checklist (Croat, 1993) treats only known if the offspring are capable of re­ 9 species. In all, 21 species (recognized by producing, but the possibility exists that Engler in his 1915 revision) have been de· much of the variation seen in the field is scribed from South America, but little is the result of hybridization. The genus war­ known about the majority of them. rants a thorough biosystematic study. Plants The tribe Dieffenbachieae is a member are readily brought into cultivation, and of the Dieffenbachia Alliance in subfamily many species are now growing at the Mis­ Aroideae. souri Botanical Garden and elsewhere. One species name used throughout much Dieffenbachia species are rarely found in of the Neotropics and certainly represent­ great numbers at any locality. It is rare to ing a variety of species (perhaps a different find more than one or two species at any species in Central and South America) is particular neotropical site. La Selva, an area Dieffenbachiaseguine (Jacq.) Schott (Stan· of Tropical wet jorest, is unusually rich in dley, 1944); (Standley & Steyermark, 1958); having seven species. Barro Colorado Is­ (Engler, 1915); (Matuda, 1954); (Jonker· land in central Panama, is typical for the Verhoef & Jonker, 1953); and (Bunting, genus in having two species, both extreme­ 1979). Since Jacquin's original material ly common. Four species have been col· came from the West Indies, probably from lected at Cerro Neblina in southern Vene­ Martinique, it is necessary to study there to zuela, and four species are also known from determine the true nature of D. seguine. Tropical wet jorest at the Jatun Sacha Re­ Only a single species of Dieffenbachia was serve at 450 m along the Rio Napo in Ec­ found in the Dominican Republic, and it uador. At study sites in Colombia, four spe­ appears to be a different species than the cies have been collected at Bajo Calima in THOMAS B. CROAT, 1994 45

Premontane pluvialJorestat near sea level, (Seemann) Engl. from Costa Rica and Nic­ but no species were collected in Premon­ aragua; D. pittieri Engl. from Costa Rica; taneJorestat 1,800 m elevation at La Plan­ and D. soconuscum Matuda ranging from ada in Narino. Typically, Dieffenbacbia is Mexico to Panama. much more abundant at lower elevations Venezuela has four species, D. asperum than at higher elevations, especially above K. Koch, D. cbanguango Bunting, D. mar­ 1,000 m elevation. garetae Bogner and D. polypbyllum L. Dra­ contium margaretae Bogner has been col­ DRACONTIOIDES lected only from Apure, Venezuela and Par­ This is a monotypic genus in tribe Las­ aguay. This species differs from all other by having linear leaves and it ieae, subtribe Dracontiinae, with a spongy Dracontium rhizome. It is represented by D. desciscens may belong to a different genus. Three spe­ cies are known from the Guianas, D. as­ (Schott) Engl., occurring in swampy areas perum K. Koch, D. Joecundum Hook, f., of eastern Brazil in Baia and Espiritu Santo. and D. polypbyllum L. Brazil has at least Relatively few collections exist but the ge­ four species, including D. asperum, D. lon­ nus poses no particular problem taxonom­ gipes Engl., D. margaretae Bogner, and D. ically. It is in cultivation in several botan­ ulei K. Krause. The latter, as well as D. lor­ ical gardens, including the Missouri Botan­ etense K. Krause were reported from Peru ical Garden, and it will be studied as an (Macbride, 1936). Dracontium loretense outgroup in a cladistic analysis of Dracon­ ranges to Ecuador and is the most common tium by Guanghua Zhu. species in the upper Amazon region. The genus is apparently rare in Colombia but is DRACONTIUM known from only a few sites in Choco De­ This small, tuberous lasioid genus (tribe partment at 240-260 m in Premontane wet Lasieae, subtribe Dracontiinae) of about 18 Jorest. Elsewhere on the Pacific slope Dra­ species ranges from southern Mexico to contium is known from Ecuador at EI Cen­ Panama, Colombia, Ecuador, and Peru on tinela. It is possibly the same species as the Pacific coast as well as to Venezuela, those collected in Colombia. the Guianas, Brazil, Bolivia, and Paraguay. The genus Ecbidnium Schott with two It also occurs in Puerto Rico. It has long species, recognized by Engler (1911) has been among the most troublesome neo­ been synonymized with Dracontium (Bog­ tropical aroid genera from the taxonomic ner, 1985). standpoint owing to both morphological The genus is currently being revised by and ecological characteristics. Firstly, pop­ Mr. Guanghua Zhu, a graduate student at ulations are frequently rare. Only 29 col­ the University of Missouri, St. Louis, work­ lections have been personally collected by ing with me at the Missouri Botanical Gar­ me in my 25 years of collecting Araceae. den. About half of the species are currently Secondly, type collections are often incom­ in cultivation at the Missouri Botanical Gar­ plete, since plants often are sterile or have den. inflorescences with no leaves. Thirdly, plants are often huge, often to 2 m or more FILARUM in height and usually consist of a single leaf. Collections rarely contain the whole This rare, monotypic genus is in tribe leaf, or they are folded and difficult to study. Zomicarpeae in the Caladium Alliance of Thus, specimens are not only rare but also, subfamily Aroideae. It is in much need of in many cases, inadequate. Dracontium is further investigation, including anatomi­ represented in the West Indies by probably cal, cytological, and molecular studies. The a single species, purportedly D. polypbyl­ genus is represented by F. mansericbense lum L.; Central America has four species of Nicolson, a tuberous understory herb in Dracontium including D. costaricense primary forest, known from Loreto Depart­ Engl., from Costa Rica and Panama; D. gigas ment as well as the type locality at Pongo 46 AROIDEANA, Vol. 17 de Manseriche in the department of Ama­ diversity. Seven taxa, including five species, zonas, along the Rio Marai'ion. More col­ occur in the region. These include H. flex­ lections are needed to determine the ex­ uosa (Kunth) Bunting, H. melinonii (Eng!.) tent of its range and morphological vari­ Jonker-Verhoef & Jonker, H. spruceana ability. Schott (two varieties), H. steyermarkii Bunting, and H. tenuispadixBunting. Even GEARUM single localities, such as the Rio Mawrin­ uma, near the base of Cerro Neblina in This is a monotypic tuberous genus (tribe southern Venezuela, may have as many as Spathicarpeae in Dieffenbachia Alliance of four species. As many as six species, many Aroideae) growing in seasonally inundated of them remaining unidentified, occur in areas. It is endemic to southern Brazil in the department of Caqueta in eastern Co­ Goias and was very poorly known until rec­ lombia at Araracura on the Rio Caqueta. ollected by A. de Araujo Dias in 1978. More Elsewhere in the Neotropics, Heteropsisap­ collections are needed to determine the pears to be relatively rare, though individ­ true extent of its range and morphological variability. uals of a species may be locally abundant. Central America has but a single species, currently deemed to be H. oblongata Kunth, GORGONIDIUM and only a single species, H. ecuadorensis The genus (tribe Spathicarpeae as above) Sodiro, occurs along the Pacific slope in consists of three species in the southern Ecuador. Four species are reported for Peru, Andes of Bolivia, Peru, and northeastern including H. linearisA. C. Smith, H. oblon­ Argentina. None of the species have been gijolia Kunth, H. peruviana K. Krause, and particularly well collected, but a modern H. spruceana Schott var. robusta Bunting. revision exists (Bogner & Nicolson, 1988). Most of the same species are also known Gorgonidium vermicidum (Speg.) Bogner for Ecuador. & Nicolson occurs in Argentina. G. mira­ Five species occur in Amazonian Brazil, bile Schott occurs in Bolivia (Depts. La Paz the majority of which probably enter the and Cochabamba), and G. vargasiiBogner lowlands of either Ecuador or Peru (or occurs in Peru (Departments Cuzco and both). These include H. linearisA.C. Smith, Apurimac at 2,800-3,000 m but collected H.longispathaceaEng!., H. macrophyllaA. recently at a lower elevation in Lima Prov­ C. Smith, H. oblongijolia Kunth, and H. ince, Lima Department). The plants have spruceana Schott. At least two species oc­ rhizomatous tubers and occur in seasonally cur in eastern Brazil: H. rigidijolia Eng!. dry areas at up to 3,000 meters. More col­ ranges from Sao Paulo State to the state of lections need to be made to determine the Santa Catarina of Brazil (Reitz, 1957), and true nature of morphological variation in H. salicijolia Kunth ranges from Sao Paulo all of the species. State to perhaps as far north as Bahia. Nei­ ther occurs in the Amazon basin. HETEROPSIS Heteropsis is not a particularly trouble­ some genus taxonomically. Species seem Heteropsis in tribe Heteropsideae of sub­ to be distinct and not markedly variable, family Monsteroideae is a genus of loosely but the genus is difficult to collect since climbing hemiepiphytes ranging from Nic­ plants often flower high up in the canopy. aragua to the Guianas, Brazil, and Bolivia. In addition, they are exceedingly difficult It is absent from the West Indies, including to bring into cultivation because of their Trinidad, despite the fact that it is partic­ slender, somewhat woody stems with long ularly abundant in Venezuela. Species are internodes. Some success in recent years heavily concentrated in eastern South has been made in growing them from seeds, America, especially in forests north of the and one chromosome count has been made Amazon River. Southeastern Venezuela and for H eteropsis oblongijolia Kunth from Cos­ the adjacent Guianas represent a center of ta Rica. In several areas of Latin America THOMAS B. CROAT, 1994 47 some species are being exploited com­ also widespread in the northern Amazon mercially for weaving into baskets or used basin. At least three additional species oc­ in making furniture. This may create a prob­ cur in the northern Amazon basin, includ­ lem with local extinction in these some­ ing H. erythropus (Schott) Engl., from times rare plants and would justify prompt southeastern Colombia and northwestern revisionary attention. Brazil, H. crinipes Engl., from southwest­ ern Colombia to northern Peru, H. peltata from the slopes of the Andes in Peru, and HOMALOMENA H. solimoensisG. Barroso in the upper Am­ The neotropical species of this chie.tl.y azon basin in western Brazil and eastern paleotropical genus are as yet poorly known. Peru. The genus consists of between 10 and 17 The thorniest problem in the genus is taxa comprising terrestrial, rhizomatous the taxonomy of the peltate-leaved plants herbs, with frequently aromatic sap (anise­ which have gone by the name H. peltata. scented), and frequently armed petioles Plants in this group are wide-ranging and and/or pubescent parts. In these features the distribution of the H. peltata complex they are nearly unique among neotropical is particularly curious. In Central America aroids. Still, they may be easily confused peltate plants are rare, having been col­ with Philodendron. A definitive means of lected only in the Isthmus of Panama at separation is the presence of staminodia only a few localities, including the Pipeline among the female .tl.owers. Road north of Gamboa, Matachui in Colon Homalomena is placed in tribe Homa· Province and at Puerto Obaldia in the Com­ lomenae in the subfamily Aroideae. The arca of San BIas. In South America, similar neotropical species of the genus were re­ plants have been collected at a few sites in vised during the mid 1980s by Mark MotHer the department of Choco, Colombia, in the as a part of a Ph.D. thesis at the University region of the Rio Palenque Biological Re­ of South Florida in Tampa, under the di­ serve south of Santo Domingo de los Co­ rection of Dr. Fred Essig. Though MotHer lorados, and at La Centinela (both in the died in 1986 the thesis is being reworked province of Los Rios, Ecuador). The only extensively by Dr. Richard Wunderlin, who other place where similar plants have been seems inclined to reduce the number of collected in South America is on the other recognized species to only seven. Certainly side of the Andes in Huanuco Department this does not re.tl.ect an accurately true num­ of Peru. This is a curious distribution, and ber of species in the genus. An example of I believe, not likely to have resulted from this reduction is the placement of H. pel­ the vagaries of collecting. It is possible that tata Masters under H. wendlandii Schott. at least some of these populations actually I have studied both of these species in the represent different taxa. It is necessary to field and there is no doubt as to their dis­ cultivate representatives of these species tinctness. The study is not yet complete, so together to determine if there are differ­ the final word is still out on Homalomena. ences not apparent in the dried herbarium Many species are still poorly known and material. collecting, especially in the Andean region of Colombia and Ecuador, is much needed. ,ASARUM A few species, such as H. wendland;;, are well known and well collected. Central This monotypic, submerged aquatic ge­ America has four species, including H. pel­ nus in tribe Caladieae in the Caladium Al­ tata, H. roezlii (Masters) Regel, H. wallis;; liance of Aroideae represents perhaps the Regel, and H. wendland;;. Three addition­ most unusual aroid. Plants are persistently al species occur in northern Colombia, in­ submergent except for the in.tl.orescence at cluding H. moJlleriCroat & Grayum sp. nov., .tl.owering time .]asarum steyermarkii Bun­ H. speariae Bogner & MotHer, and H. pic­ ting (Bunting, 1975) occurs only in a few turata (Linden & Andre) Regel, the latter rapidly flowing streams off the Gran Sabana 48 AROIDEANA, Vol. 17 in the Guiana Highlands of Venezuela in four species occur in the Guianas, five are the state of Bolivar and in Guayana in the reported for Brazil by Mayo and seven taxa Mazaruni River basin. Though rare, it is now are reported for Peru (Croat, 1993). Mon­ reasonably well known owing to the work stera is primarily centered in the northern of J. Bogner, who collected it and has it in Andes and lower Central America (Costa cultivation at the Munich Botanical Garden Rica and Panama) (Madison, 1977). The (Bogner, 1977, 1984, 1985). most widespread "species" in the genus is M. adansonii (Schott) Madison which, un­ MANGONIA der three varieties, ranges throughout the neotropics from Nicaragua to the West In­ This tuberous genus of two species, a dies, the Guianas, and eastern Brazil to Santa member of the tribe Spathicarpeae (Die! Catarina, and in the Amazon basin to Peru fenbachiaAlliance of Aroideae) is endemic and Bolivia in South America. Madison to southeastern temperate South America. (1977) reported that species diversity in Mangonia tweediana Schott occurs in South America was greatest in an arc rang­ southern Brazil and adjacent Uruguay. ing from Peru north to the Guianas. Mon­ Mangonia uruguayana (Hicken) Bogner stera was divided by Madison into four sec­ is endemic to Uruguay. Plants occur in sea­ tions, Monstera, Marcgraviopsis, Tornelia, sonally dry areas. Though still poorly and Echinospadix. In Madison's revision known, it has been studied in recent years section Monstera is the largest section (13 by Bogner (1973). species) with sect. Marcgraviopsis having six species and Ecinospadix and Tornelia MONSTERA with a single species each (M. tuberculata The genus Monstera, one of four neo­ Lundell and M. de/iciosa Liebm. respec­ tropical genera in the tribe Monstereae tively). Based on the modern day distri­ (Aroideae), was last revised by Mike Mad­ bution of the sections (all in Mexico versus ison (1977) in his Ph.D. thesis at Harvard two in South America) , Madison (1977) be­ University. The revision treated 22 species, lieves that Monstera is a remnant of the including three varieties, but perhaps a Paleogene tropical flora of North America more realistic estimate of the number of and that it spread into South America dur­ species in the genus, as it is known today, ing the Neogene times. Nevertheless, it is is closer to 60 (Croat, 1988a). For example, clear that considerable evolution of the ge­ Madison recognized 20 taxa for Central nus has taken place in South America. Nu­ America, but studies there by Croat and merous species in South America, especial­ Grayum, who are working on a revision of ly in the Amazon lowlands, do not fit spe­ the Araceae of Costa Rica, have found that cies he circumscribed. there are at least 39 species of Monstera for Monstera is not highly diverse at any of Central America. There are 22 species in the wetter forest sites on the Pacific slope. Costa Rica alone, equalling the total num­ For example, only two species have been ber of species included in Madison's 1977 found at La Planada in Narino and at Bajo revision. Although Monstera is particularly Calima in Valle Department of Colombia, rich in Central America, the South Ameri­ two well studied areas involved in a com­ can continent is expected to have a pro­ parative study sponsored by the National portionately larger increase owing to the Geographic Society. fact that it is not only a vastly larger area, Monstera represents a difficult group to but also because it remains poorly collect­ study because of the difficulty of collecting ed. Indeed, since many of the Central them. They are usually firmly appressed­ American species are endemic, the number climbing hemiepiphytes or sometimes of species for the genus will probably ex­ loosely scandent canopy vines, growing ceed sixty. higher in the canopy (e.g. M. pittieri Engl. Venezuela currently is known to have and M. tuberculata Lundell) than most ar­ seven taxa (Croat & Lambert, 1986); about oids, and on average they are very large, THOMAS B. CROAT, 1994 49 contributing to the lack of many well-pre­ genus is even more poorly known than An­ pared, and well-documented collections. thurium. This is especially true of the larg­ Another feature making the genus difficult er species, which are poorly collected and to understand is the well developed hetero­ represented by poorly made, poorly de­ blastic leaf development (Madison, 1977) scribed collections. Much of what was said which they exhibit. Juvenile, preadult, and about the areas of high species diversity for adult plants often exhibit marked differ­ Anthurium is true of Philodendron as well, ences. Seedlings exhibit scototropic growth but Philodendron is much more abundant until they become established on trees at lower elevations than is Anthurium. Ow­ when they change their growth patterns and ing to the much more degraded nature of begin growing toward light (Strong & Ray, lowland forest throughout much of the 1975). Neotropics, plants are often more difficult to find. This coupled with the fact that MONTRICHARDIA Philodendron species are often much larg­ er in size and frequently high in trees means This genus (tribe Montrichardieae) of two that they are often neglected by the average species is widespread, occurring at low el­ collector. evations and ranging from Guatemala and There are 107 taxa including 96 species Belize to Panama, Colombia, the Guianas, of Philodendron in Central America with Brazil, and Peru. Plants are shrub-like with the majority (87 species) occurring in Cos­ erect, sometimes armed stems and some­ ta Rica and Panama. There are 121 taxa in­ what hastate simple blades, and occur prin­ cluding 108 species in Venezuela (versus cipally in standing water along the margins 77 taxa, including 71 species of Anthuri­ of streams or lakes and in swampy areas and um), ten species in Trinidad, 43 species in estuaries but may also occur in the under­ the Guianas, 100 taxa including 97 species story near watercourses. Montrichardia ar­ in Brazil, 62 in Peru, four in Paraguay, and borescens (L.) Schott, which makes up the five in Argentina. Ten species range as far northern range of the genus, occurs in the south as Santa Catarina State in Brazil (Reitz, south of Venezuela in the states of Bolivar, 1957), three of them being members of Amazonas, and Apure, but it is mostly re­ subgenus Pteromischum, i.e. P. obliquifol­ placed with M. linifera (Arruda) Schott in ium Engl., P. ochrostemon Schott, and P. the Amazon basin. An unusual population, sonderianum Schott. All species in Santa believed to be M. arborescens, located in Catarina are endemic to eastern Brazil ex­ the lower Rio Negro north of Manaus has cept P. bipinnatifidum Schott ex Endlich­ blades which may sometimes be complete­ er. That latter ranges to Argentina, Para­ ly three-parted. This population deserves guay, and Bolivia. more investigation since this feature is not Any counts for the Andean countries known elsewhere in the genus. Otherwise would have a margin of error of more than the genus warrants very little work even 30 percent, but the richness of the Philo­ though no modern revision exists. dendron floras more or less mirrors that of Anthurium since all of these countries have PHILODENDRON a range of lowland, middle elevation, and The genus is a member of tribe Philo­ higher elevation forest. Perhaps only in Bo­ dendreae of the Philodendron Alliance of livia would the Philodendron species out­ subfamily Aroideae. It is the second largest number the Anthurium species, this owing in the family and is as widespread as An­ to the fact that there is a broad stretch of thurium, ranging from Central Mexico to lowland forest in Beni contiguous with the the West Indies, the Guianas, and Argen­ Amazon basin and the forests of higher el­ tina. Like Anthurium it is difficult to know evations are often very cold, often quite exactly how many species there are, but it arid, and have relatively few species of Ar­ can be reliably estimated that there are as aceae compared to forests of similar ele­ many as 700 species. In some respects the vations at higher latitudes. 50 AROIDEANA, Vol. 17

Nevertheless, Philodendron definitely the vast majority of species in the genus, becomes more dominant in areas with large and there are numerous taxonomic prob­ expanses of lowland forest (especially sea­ lems in the group. Not only are many spe­ sonally dry forest), such as in the Amazon cies poorly known or new to science but basin or in Venezuela. For example, while the sectional classification of Engler must Anthurium species outnumber Philoden­ be carefully reinvestigated. Many sections, dron species in the Andean countries, especially but not exclusively Philoden­ Philodendron is much the larger genus in dron and Calostigma, appear to be artificial Venezuela, with 118 versus 77 species of with unrelated elements. It is probable that Anthurium. In contrast, Costa Rica and the reliance principally upon ovular struc­ Panama which reflects the distribution of ture alone, which appears to characterize Araceae in the Andes of South America have Engler's revision, is artificial. Still, no better proportionately more species of Anthuri­ system of grouping species has been de­ um. Costa Rica has 75 species of Anthuri­ vised, so until a more thorough investiga­ um and 51 species of Philodendron while tion is made it seems best to discuss the Panama has about 160 species of Anthuri­ genus in terms of the existing classification. um and only 77 species of Philodendron. Some sections, especially Macrolonchium, Sectional groupings of Philodendron are with P.fragrantissimum (Hook.) Kunth, P. much more poorly understood than those simsii Kunth, P. melinonii Brongn. ex Re­ of Anthurium, but the three subgenera are gel, and P. roraimae K. Krause (all admit­ distinct and their distributions are now rea­ tedly similar species), include what appear sonablywell known. The subgenus Mecon­ to be aberrant elements in them. For ex­ ostigma (Mayo, 1991) with 15 species is ample, this group also contains P. pinna­ largely distributed in southern Brazil but tiftdum (Jacq.) Schott with its few-ovulate ranges north into the Amazon basin. The locules and pinnately lobed blades, which subgenus Pteromischum, with an estimat­ would appear to be more well placed in ed 75 species occurs throughout most of section Polytomium, along with P. angus­ the range of the genus from Mexico (San tisectum Engl. and P. distantilobum K. Luis Potosi) to the West Indies (three spe­ Krause. Considering the immense variation cies), the Guianas, and Argentina, but has in some sections, especially in sections of a major center of distribution in lower Cen­ Philodendron and Calostigma, the section tral America and northwestern South Amer­ may prove to be unworthy of recognition. ica. Relatively few species occur in the south The majority of species in the subgenus of the continent, but considerable numbers Philodendron are in sections Philodendron (each 10 species) occur in the Guianas and or Calostigma (recently reclassified from in the Amazon basin (at least 12 species). sections Polyspermium and O/igosper­ Venezuela has a total of about 15 species. mium (Schott) Pfeiffer, respectively (Mayo, A few species, especially P. inaequilater­ 1990). Both sections occur nearly through­ um Liebm. are wide-ranging and extend out the range of the genus. Subsection Be­ from Mexico to Venezuela and Ecuador, but locardium of section Calostigma, with of­ none of the species in the Amazon basin ten elongate leaves with frequently tumid range into Central America or to the west­ petioles and a dark, purplish ring demar­ ern coast of South America, according to cating the petiole and blade, is particularly M. Grayum (pers. comm.). A common dis­ distinctive and is well distributed in lower tribution pattern is from the Atlantic slope Central America and on the lower slopes of lower Central America to the Pacific coast of the eastern slopes of the Andes. of Colombia to Ecuador. This same pattern Section Macrolonchium (mentioned is evident for species of subgenus Philo­ above) occurs in Central America, northern dendron but the largest share of the Central Venezuela, the Guianas, and the Amazon American species of this group is endemic basin. It has only about six species. to Central America. Section Baursia has only a few species The subgenus Philodendron comprises in Central America, in northern Venezuela THOMAS B. CROAT, 1994 51 or on the Pacific slope of Colombia and structure of their ovules, differ remarkably Ecuador. The majority of the species in the. in distribution. Section Macrogynium, rep­ section occur in eastern Brazil, the Amazon resented only by P. hederaceum (Jacq.) basin, or the lower eastern slopes of the Schott, ranges from Mexico (both slopes) Andes. Section Trltomophyllum, with its tri­ to Panama, northern Colombia, Venezuela, lobed blades and single ovule per locule and the Guianas, primarily in areas of sea­ is primarily Central American but ranges sonally dry forest. In contrast, section down the western coast of South America Camptogynium, represented by only P. to Ecuador. It is a small group with prob­ longistilum K. Krause is known only from ably fewer than ten highly variable taxa, the upper Amazon basin in Brazil. While several of which are new. The section, as most of the species in subgenus Philoden­ defined by Engler, certainly appears to be dron from the lowland Amazon basin are an unnatural one. reasonably well known, as are those from Section Schizophyllum ranges from Central America, Venezuela, the Guianas, northeastern Venezuela and the lowland and eastern Brazil, the species from the An­ Guianas to southern Brazil, and throughout dean countries of western South America the Amazon basin to Bolivia in the south­ are even more poorly known than those of west. Only P. pedatum (Hook.) Kunth is Anthurlum from the same region. Much of very widespread, and most of the approxi­ the material collected in the Andean coun­ mately six species are restricted to the low­ tries by earlier collectors, especially Leh­ er Amazon basin and to eastern South mann in Colombia, Fendler and Pittier in America. A new species exists in French Venezuela, and Poeppig in Peru, was stud­ Guiana. ied by either Schott or Engler, and many Section Polytomium, a small group with species were described as new. The same mostly large, pinnately lobed leaves and is true to a lesser extent with Philodendron few-ovulate locules, ranges from Mexico to collected by Luis Sodiro in Ecuador, but the Greater Antilles, northern Venezuela, whereas Sodiro was keenly interested in northern Colombia, and the Amazon basin. Anthurlum and described 281 taxa of An­ Only P. lacerum (Jacq.) Schott occurs in thurium (including 257 species) chiefly the West Indies (Cuba, Hispaniola, andJa­ from Central Ecuador, he described no maica), while three species occur in Cen­ Philodendron. Thus many of the species tral America: P. radiatum Schott, P. war­ from Ecuador remain undescribed. Perhaps scewicziiK. Koch, and P. dressleriBunting. more importantly, the areas which are prob­ Philodendron fendleri K. Krause occurs in ably richest for Araceae, namely the low­ northern Venezuela and Trinidad while P. land Pacific coastal regions, were not elegans K. Krause occurs in northern Co­ opened up to easy access until relatively lombia. Two species, P. distantilobum K. recent times. More specimens have been Krause and P. angustisectum Engl., are re­ made from these regions in the past 25 years stricted to the Amazon basin. than in all the preceding centuries. These A newly described section, Philopsam­ regions of very wet forest contain an ex­ mos Bunting, with probably fewer than 15 ceedingly rich, mostly endemic flora re­ species, is known only from the Amazon maining very poorly known. The majority basin and especially the region of the Gui­ of species appear to be new to science. An ana Highlands. The group is distinguished example is the site known as "Bajo Cali­ primarily by the elongate, many-veined leaf rna," a region of lowland Premontane wet blades. The most widespread species in the pluvial transition forest in Valle Depart­ group, P. pulchrum G. Barroso, ranges from ment near Buenaventura, Colombia, which southern Venezuela to Brazil (Roraima and has yielded 40 species of Philodendron. Amazonas) to Colombia (Vaupes and Pu­ Relatively few of these species have yet been tumayo) and Peru (Amazonas, Loreto and identified, and a large percentage of the San Martin). species is clearly new to science. The re­ Two small sections, both unique in the gion to the north, in Pluvial forest in the 52 AROIDEANA, Vol. 17 department of Chaco, is likely to be even Poeppig, and R. wendlandii Schott. In the richer. Colombia and Ecuador together current revision being carried out by Croat, probably have more than all the rest of South 78 taxa and 75 species of Rhodospatha are America. The Pacific slope of Colombia is recognized. Species diversity is highest in much dissected with rivers, and consider· Venezuela and in the western Andes of able endemism is apparent from investi· South America as well as in Central Amer­ gations there. For example, studies in Na­ ica. Central America has 14 species and 15 rino in three sites of Premontane wet/orest taxa, most of them in Costa Rica and Pan­ (Croat, 1992) have shown that even sites ama. Venezuela has 13 species, the Guianas relatively near one another, e.g. Rio Nembi has four species, and eastern Brazil has four and Rio Nambi, separated by only about 15 species, including R. oblongata, R. lati/ol­ km and at the same elevation, shared few ia, and two new species. Rhodospatha lao species in common. Investigation of wetter ti/olia and R. oblongata are widespread in life zones in Narino are expected to provide the upper Amazon basin and range to east­ even more startling differences. In short, ern Ecuador and Peru. Macbride reported much work remains to be done with Philo­ these two species in his treatment for the dendron in South America. Flora of Peru but included seven additional names which are now placed in 5tenosper­ PISTIA mation. Only one additional species, R. moritziana was cited in the recent check­ This monotypic floating aquatic genus in list for the Flora of Peru (Croat, 1993), but tribe Pistieae (Aroideae), represented by there are believed to be as many as 16 ad­ Pistia stratiotes 1., is both well collected ditional species in Peru, most of them new and moderately well studied, but no recent to science. Species diversity is especially study has been made, the last by Engler in high in Ecuador with approximately 27 spe­ Das Pflanzenreich (Engler, 1920). It ranges cies, most of them new to science. from southern Florida to the West Indies Rhodospatha is a difficult genus because and throughout Central and south America of the lack of definitive key characters. The to as far as 35 degrees south in the region inflorescences are less valuable than in most of Buenos Aires in Argentina (Crisci, 1971). genera and the spathe, which usually falls off just after anthesis, is often lacking. Spathe RHODOSPATHA and spadix color is not greatly variable and This genus of hemiepiphytic appressed fruits are almost always lacking, and in cases climbers or caulescent terrestrial herbs in where they are present, are not particularly tribe Monstereae ranges from Mexico to variable in any important respect. Plants are Panama, Venezuela, Trinidad, the Guianas, difficult to collect, generally being high up eastern Brazil, and the Amazon basin to Peru on tree trunks where they usually require and Bolivia. The last revision of Rhodos­ a clipper pole to bring them down. Char· patha by Engler in Das Pflanzenreich in­ acters separating species often involve col­ cluded eleven names. Three of the species oration of dried plants, which, though con­ proved to be members of other genera: sistent for a species, is difficult to quantify. Monstera (M. eostarieensis [Engl. & K. Krause] Croat & Grayum), 5tenosperma­ SCAPHISPATHA tion (5. tuerekheimii [Engl.] Croat), and Heteropsis (H. melinonii [Engl.] Jonker· This monotypic tuberous genus of herbs Verhoef &Jonker). Three additional names in tribe Caladieae (Caladium Alliance, have been synonymized: R. longipes Engl. Aroideae) ranges from eastern Bolivia to and R. blanda Schott under R. oblongata eastern Brazil. It was very poorly known Poeppig and R. pieta Nicolson under R. until recollected byJ. Bogner in 1976 (Bog­ moritziana Schott. This left only five spe­ ner, 1980) in the state of Ceara. The plants cies: R. densinervia Engl., R. belieonii/olia flower before the leaves are produced. The Schott, R. lati/olia Poeppig, R. oblongata peltate leaf blades look much like those of THOMAS B. CROAT, 1994 53 the common Caladium bicolor. The only This genus of five to seven species has had species, S. gracilisBrongn. ex Schott, poses no serious revision since Engler's time, no special taxonomic problem, but more though two species were treated in the Flo­ collections are needed to determine geo­ ra of Paraguay (Croat, 1988b) and in the graphic distribution. Flora de Argentina (Crisci, 1971). Spathicarpa occurs in eastern Brazil SCHlSMATOGLOrnS (from Bahia south to Rio Grande do Sui) as well as Uruguay, Paraguay, Bolivia, and The genus is primarily Old World, but northeast Argentina. Along with Mangonia two species occur in the Venezuelan and Pistia, it is one of the few aroid genera Guianas, the lowlands of southern Vene­ in Uruguay. One species, S. hastijoliaHook., zuela, and in adjacent Suriname and the occurs in Bolivia, Argentina, and Uruguay Amazon basin. Schismatoglottis bolivarana as well as in Paraguay and eastern Brazil, Bunting & Steyermark is endemic to Ven­ where it ranges from Minas Gerais to Rio ezuela, known from both Bolivar and Ama­ Grande do Sul. Spathicarpa lanceolata zonas, whereas S. spruceana (Schott) Bun­ Engl. occurs in Paraguay and southern Bra­ ting, which includes two varieties, occurs zil (Santa Catarina and Rio Grande do sui). in the same states but also in Brazil and The remaining species are all Brazilian and Peru. Relatively few collections exist for S. are more poorly known. Spathicarpa bur­ bolivarana. Schismatoglottis americana chelliana Engl. occurs in Goias and Ceara Jonk. & Jonk. occurs only in Suriname on and S. gardneri Schott in Ceara, Pernam­ Tafelberg. buco, Piaui, and Goias in the northeast of Schismatoglottis is a member of tribe Brazil, while S. tU'eediana Schott occurs in Schismatoglottideae of the Schismatoglot­ Sao Paulo State. tis Alliance in subfamily Aroideae. Engler (1920) treated S. gardneri as a variety of S. sagittijolia, whereas Mayo in­ SPATHANTHBUM cludes it as a distinct species. This genus This tuberous-stemmed genus of two is a highly variable one, especially in leaf species is still poorly known. One of the size and shape. More collections need to species, S. orbignyanum Schott, is known be made of the last three species men­ only from northern Argentina, Bolivia, and tioned, and other species need to be Peru, ranging up to 2,400 m elevation. This brought into cultivation to determine the species, once poorly collected, is now validity of the last taxonomic treatment. known from a number of well documented collections. Spathantheum heterandrum SPATHIPHYLLUM (Baker) N. E. Br., is now considered to be a synonym of the former. The second spe­ This genus of about 60 terrestrial or ru­ picolous species in the tribe Spathiphyl­ cies, S. intermedium Bogner is a distinct leae (Monsteroideae) ranges from central undescribed species from Peru. Spathan­ Mexico to Trinidad, the Guianas, Brazil, and theum is a member of tribe Spathicarpeae Peru primarily at lower elevations. Bunt­ in the Aroideae. ing's revision (1960) included 36 species, but many species have been subsequently SPATHlCARPA collected during the past three decades. This small group in tribe Spathicarpeae Baker & Burger (1976) revised the Costa (with the same subfamilial placement as Rican species and described two new taxa. Spathantheum) consists of small herbs with Twenty-four species and two varieties oc­ tubers, rhizomes, or rhizomatous tubers oc­ cur in Central America (including several curring as understory herbs in humid or new species); eleven species and one sub­ marshy forests, but also in dry forest (caa­ species occur in Venezuela; five in the tinga), where the vegetative parts may die Guianas; four in Brazil; and at least seven back during the dry season (Croat, 1988a). in Peru. There are relatively few species in 54 AROIDEANA, Vol. 17 the Amazon basin, but one widespread spe­ there is only one species, the widespread cies, S. cannifolium (Dryander) Schott is S. angustifolium Hemsley. Even in Tropi­ the most common and widespread species cal wet forest at the Jatun Sacha reserve on in the genus ranging from Trinidad and the the Rio Napo there is only one species lowland Guianas to just south of the Ama­ known. There are relatively few, mostly zon River in Brazil and west to Colombia widespread, species in the lowland Ama­ and Peru as far south as Pasco Department. zon basin (surely fewer than five species). Curiously, it has not been found in Bolivia. Next to Dieffenbachia, Stenospermation is While most Spathiphyllum species are un­ probably the most difficult genus taxonom­ derstory herbs, some species, such as S. ically and certainly one of the most poorly friedrichsthalii Schott, often occur in open known of all aroid genera in the neotropics. marshy areas or along the banks of streams The taxonomic difficulty is owing largely to or lakes. Others, including S. quindiuense the absence of pronounced differences be­ Engl., are rheophytes, occurring along or tween species. Whereas most genera of Ar­ in streams. Costa Rica and Panama probably aceae have valuable taxonomic characters constitute a center of diversity with 17 spe­ in leaf venation, Stenospermation leaf cies, but Mexico is also diverse with six blades offer little in the way of taxonomic species. In South America, the Guiana re­ value, mostly lacking differences in the gion (including the Venezuelan Guyana) midrib on either surface and usually lack­ is a center of diversity with eight species ing pronounced primary lateral veins al­ occurring there (Bunting, 1960). together. Yet conventional sorting of spec­ imens by texture, color, thickness, size, el­ evation, and lifezone clearly show that there STENOSPERMATION are many species, judged by standards of A primarily epiphytic genus in tribe Mon­ speciation in other genera. However, be­ stereae (Monsteroideae) with approxi­ cause pronounced differences are lacking mately 60 species, the majority of which are between the species, the construction of new to science. The genus ranges from dichotomous keys is extremely difficult. Guatemala to the Guianas, Brazil, and Peru, but it is known primarily from the Andes, SYNANDROSPADIX especially in Colombia and Ecuador. Only ten species occur in Central America, nine The genus is a member of tribe Spathi­ of these in Panama (Gomez, 1983). Costa carpeae in the Dieffenbachia Alliance of Rica has seven species, but few occur in subfamily Aroideae. It is a monotypic, tu­ Middle America. For example, there are only berous genus from northern Argentina, Bo­ two species each in Guatemala and Nica­ livia, and Peru. Synandrospadix vermitox­ ragua. Seven species occur in Venezuela, icus (Griseb.) Engl. is well known and pos­ four species occur in the Guianas, and Mayo es no taxonomic problems. It has been well reports only two species for Brazil. Araceae described and illustrated by Crisci (1971) listed in The Catalogue of the Flowering in the Flora Argentina. It is in cultivation Plants and Gymnosperms of Peru (Croat, in the Miami area as well as at the Munich 1993) report nine species for Peru. Botanical Garden, the Missouri Botanical Stenospermation occurs primarily at Garden, and elsewhere. middle elevations in areas of very wet for­ est. Relatively few species occur at lower SYNGONIUM elevations except in the wettest life zones. For example, in Colombia, the La Planada Syngonium, of tribe Caladieae in the Ca­ reserve in Narifio at 1,800 m has seven spe­ ladium Alliance of Aroideae, is a genus of cies, and Bajo Calima in Valle Department, hemiepiphytic, appressed climbers and located near sea level, has eight species, vines. The genus, recently revised (Croat, whereas in Tropical moist forest on Barro 1981), has an estimated 36 species. The Colorado Island in Panama (Croat, 1978) center of diversity for Syngonium is in Cos- THOMAS B. CROAT, 1994 55 ta Rica and Panama which together have a itate male flowers. Two species, T pere­ total of 17 species. Costa Rica has 14 spe­ grinum (Schott) Engl. and T weddellian­ cies, including four undescribed species, urn Brongn. ex. Schott, occur in Paraguay while Panama has 11 species. Mexico is a (Croat, 1988a). The former also occurs in secondary center of diversity with eight southern Brazil and in northern Argentina. species. Only three species from Central Taccarum weddellianum is the most wide­ America, S. angustatum Schott, S. macro­ spread species in the genus, ranging to Bo­ phyllum Engl., and S. podophyllum Schott, livia, Peru, and Brazil (Acre and Matto are very wide-ranging, with all others en­ Grosso). Several other species occur in Bo­ demic to either Mexico or to Costa Rica and livia, including T cardenasianum Bogner Panama. Syngonium podophyllum is par­ and T caudatum Rusby. Taccarum war­ ticularlywide-ranging, from Mexico to Bra­ mingii Engl. occurs in southern Brazil in zil. Sao Paulo State. The West Indies have but a single spe­ Only Taccarum weddellianum is very cies, S. auritum L. Relatively few species well known (Croat, 1985c). It is in culti­ of Syngonium occur in South America, vation at the Missouri Botanical Garden and though several collections made in recent elsewhere. years probably represent additional new species. Eleven species are endemic to ULEARUM South America with four of them occurring The genus, in tribe Zomicarpeae of the in Peru; S. gentryanum Croat, S. hastifol­ Caladium Alliance in Aroideae, is a genus ium Engl., S. podopbyllum Schott, and S. of understory rhizomatous herbs from "ter­ yurimaguense Engl. ra firme" forest in the upper Amazon basin Syngonium is divided into four sections, of western Brazil and Peru. Engler de­ all of which occur in Central America. The scribed Ulearum sagittatum from an Ule sections include section Syngonium, with collection (6323) from Pongo de Cain­ pedate leaves, the largest and most wide­ arche (now called Shanusi) in San Martin spread section occurring throughout the department of Peru. Josef Bogner recently range ofthe genus; section Cordatum, with acquired material of the genus from Brazil. simple cordate blades, ranging from Mex­ It had been collected by Jacques Jangoux ico to Brazil; section Oblongatum, with along the RioJurua Miry in the state of Acre. simple, oblong blades, from Costa Rica and This was another collecting locality of Ule Panama; and section Pinnatilobum, with and it is not distant from the Peruvian bor­ pinnately lobed blades, known only from der. Bogner has recognized a new variety, Mexico and Guatemala (Croat & Bogner, U sagittatll1n var. l'iridispadix, from this 1987). Syngonium exhibits pronounced material in Acre. Both varieties are in cul­ heterophily (Ray, 1981, 1986, 1987) and tivation at the Munich Botanical Garden. each growth phase from juvenile to pread­ A collection described by Madison (1980) ult to adult plants must be collected for as Ulearum recondita Madison has proven proper study. to represent a new genus, Bognera.

TACCARUM UROSPATHA Taccarum is a tuberous-stemmed genus The genus (of tribe Lasieae subfamily La­ of five species in tribe Spathicarpeae (Die! sioideae) consists of fewer than ten species jenbachia Alliance of Aroideae) restricted of rhizomatous herbs occurring in swampy to temperate South America in Southern areas or along streams or lakes at low ele­ Brazil, Bolivia, Paraguay, and Argentina. It vations, mostly in open habitats including grows in damp soil in the understory or wet places in savannas. In Central America along roadsides in humid to seasonally dry there is but a single species, U grandis areas. The genus is recognized by its tu­ Schott, occurring in swamps very near the berous stem, highly divided leaf and stip- Caribbean coast. Another species, U sag- 56 AROIDEANA, Vol. 17 ittifolia (Rudge) Schott, is widespread in tane wet forest at 1,800 m and two species the Amazon basin. Engler treated 12 spe­ occur at Bajo Calima in Valle Department cies, based mostly on leaf shape, but blade in Premontane pluvial forest at near sea shape is immensely variable in the well level. Three species occur at Jatun Sacha known species, and a reevaluation of the along the upper Rio Napo in a Tropical wet remaining species must be made. Jonker· forest life zone, four species occur in Trop­ Verhoef & Jonker (1953) have already re­ ical wet forest at La Selva in Heredia Prov­ duced three species, namely U. decipiens ince of Costa Rica (Croat & Grayum, ined.), Schott, U. dubia Schott, and U. hostmanii and two native species occur in Tropical Schott under u. sagittifolia (Rudge) Schott, moist forest on Barro Colorado Island in and in Central America U. tonduzii Engl. Panama (Croat, 1978). is a synonym of u. grandis Schott (previ­ Xanthosoma ranges through a wide va­ ously U. Jriedrichsthalii Schott). The type riety of life zones from Tropical moistforest of the latter name has proven to be a mem­ to Tropical forest and from sea level to at ber of the genus Sagitta ria in the Alisma­ least 2,200 meters. I t may occur in the forest taceae (Hay, 1992). In addition, consider­ understory, along the edges of roads and ing the high degree of variability in the streams, or in open swampy areas. In sea­ genus, it is possible that the Central Amer­ sonally dry areas the stems are tuberous, ican species is synonymous with the wide­ usually depressed-globose with roots spread U. sagittifolia. Perhaps the total around the upper periphery, or in wetter number of species will be even further re­ areas it may be caulescent or even arbores­ duced. cent with trunks to 4 m tall. Like many other According to Hay (1992) the recently de­ genera, the leaf blades range from simple scribed genus Urospathella Bunting (Bun­ to compound. Xanthosoma remains one of ting, 1988a), based on a plant originally the more poorly understood genera in the described as Cyrtosperma wurdackii Bun­ neotropics. This is not so much owing to ting, is a synonym of Urospatha. Hay (1988) real differences in distinguishing species transferred the species to Urospatha as U. as it is to the poor preservation of most wurdackii (Bunting) Hay. specimens and the fact that such a large percentage of the species described were derived from cultivated collections of un­ known origin. The plants are sometimes XANmOSOMA very large with very fleshy and juicy parts, This genus (tribe Caladieae of subfamily especially the petioles, though the leaves Aroideae) of approximately 45 species are usually membranaceous. This makes ranges from Mexico to the West Indies, drying the plants difficult. Blades, when Trinidad, the Guianas, and Argentina. As folded, often stick together and cannot be many as 12 species occur in Central Amer­ unfolded. It is imperative that collections ica, and four in the West Indies. Twenty be properly cut into sections or folded in species are known from Venezuela, per­ such a manner that they can be studied after haps owing as much to the more than twen­ they are dried. ty years of activity by G. S. Bunting in Ven­ Xanthosoma is being revised by Sue ezuela than to real differences in the dis­ Thompson at the Carnegie Museum of Nat­ tribution of Xanthosoma. Two native spe­ ural History (CM). While no specimens have cies occur in Trinidad; nine species in the been available for study, a review of the Guianas; ten in Brazil; two in Paraguay; and type photo collection at the Missouri Bo­ ten in Peru (Croat, 1993). Only a few spe­ tanical Garden indicates that there are per­ cies usually occur at a typical site, and low­ haps names that should be synonymized. land and upland sites tend to have about However, it is difficult to predict any spe­ the same number of species. For example, cific numbers. Certainly the genus needs to at study sites in Colombia, three species be extensively collected with detailed notes occur at La Planada in Narino in Premon- and photographs throughout its range, ex- THOMAS B. CROAT, 1994 57 cept for some of the widespread weedier E. Br. was described from Colombia based species. on a Linden collection without exact lo­ cality. The latter should be recollected. ZOMICARPA Zomicarpa is a member of tribe Zomi­ LITERATURE CITED carpeae in the Caladium Alliance of tribe Baker, R. A. & W. C. Burger. 1976. Key Aroideae. It is a genus of terrestrial, tuber­ and commentary on the species of Spa­ ous herbs comprised of three species from thiphyllum (Araceae) in Costa Rica, in­ eastern Brazil (Ceara to Pernambuco and cluding S. silvicola, sp. nov. Phylologia Bahia). Engler (920) reported all three 33(7):447-453. from a single locality, near Ilheos in Bahia Bogner, J. 1969. A propos du jenera An­ with Z. pythonium (BI.) Schott also occur­ dromycia A. Rich. (Aracees). Adan­ ring elsewhere in Bahia at Almada. Until sonia Ser. 2, 9(1):125-130. recently two of the three species, Z. steig­ ---. 1973. Otra especie de Mangonia eriana Schott and Z. riedeliana Schott, (Araceae) del Uruguay. Darwiniana were known only from Ilheos until]. Bog­ 18(1/2):70-79. ner collected Z. riedeliana in the state of ---. 1977. jasarum steyermarkiiBun­ Ceara. The characteristics separating the ting (Araceae). Aqua Planta 2/3:4-7. species do not appear to be strong ones, ---. 1980. The genus Scaphispatha and the populations should be recollected Brongn. ex Schott. Aroideana 3:4-12. to determine the validity of the species, ---. 1984. Das Pflanzenportrait: jasa­ since it is curious that three species of a rum steyermarkiiBunting. Aqua Plan­ rare genus should be so narrowly circum­ ta 4-84:15-16. scribed geographically. Zomicarpa steig­ ---. 1985. jasarum steyermarkiiBun­ eriana keys out as having trisect leaves, ting, an aquatiC aroid from Guyana whereas the other two have pedatisect highland. Aroideana 8:55-63. leaves with either five or seven lobes. Con­ Bogner, J., S. J. Mayo & P. Boyce. 1994. Sidering the variation in leaf morphology The Genera of the Araceae. Royal Bo­ of other genera, the characteristics sepa­ tanic Garden, Kew (in prep.) rating these three species may not hold up. Bogner, J. & D. H. Nicolson. 1988. Revi­ Bogner (pers. comm.) reports that in Ceara, sion of the South American genus Gor­ the plants are only evident during the rainy gonidium Schott (Araceae) Spathicar­ season and are dormant during the dry sea­ peae. Bot.jahrb. Syst. 109(4):529-554. son. ---. 1991. A revised classification of Araceae with dichotomous keys. Wi/ld­ ZOMICARPELLA enowia 21:35-50. This genus is a member of tribe Zomi­ Bunting, G. S. 1960. A revision of Spathi­ carpeae in the Caladium Alliance of Aro­ phyllum (Araceae). Mem. N. Y. Bot. ideae. It consists of two species with creep­ Gard. 10:1-53. ing rhizomatous stems and is endemic to ---. 1965. Commentary on Mexican the upper Amazon region in western Brazil Araceae. Gentes Herb. 9:289-382. (and no doubt adjacent Peru). The genus ---. 1975. Nuevas especies para la re­ was considered to be monotypic until re­ vision de las Araceas Venezolanas. Acta. cently, when]. Bogner collected a new spe­ Bot. Yen. 10:263-335. cies, Z. amazonica, in loamy soil in "terra --. 1979. Sinopsis de las Araceae de firme" forest along the Rio Javari in western Venezuela. Rev. Fac. Agron. (Mara­ Brazil. It is in cultivation at the Munich cay). 100-4):139-290. Botanical Garden. The same species was --. 1986. New taxa ofVenezuelanAr­ collected by Mur~a Pires at Benjamin Con­ aceae. Phytologia 60:293-344. stant in 1943, but remained unidentified ---. 1988. New taxa of Venezuelan Ar­ until recently. Zomicarpella maculata N. aceae II. Phyt%gia 64:459-486. 58 AROIDEANA, Vol. 17

---. 1988a. Urospathella, a new genus 1991. A revision of Anthurium of Venezuelan Araceae. Phytologia 65: sect. Pachyneurium (Araceae). Ann. 139-141. Missouri Bot. Gard.78:539-855. ---. 1989. Notes on Araceae. Ann. Mis· --. 1992. Species diversity of Araceae souri Bot. Gard. 76:917-919. in Colombia: a preliminary survey. Ann. ---. (In press). Araceae. In]. Steyer­ Missouri Bot. Gard. 79:17-28. mark, P. Berry & B. Holst (eds.), Flora --. 1993. Araceae. In L. Brako &]. o/Venezuelan Guyana. Timber Press, Zarucchi (eds.), Catalogue of the flow­ Portland, OR. ering plants and Gymnosperms of Peru, Crisci, J. V. 1971. Flora Argentina: Ara­ 71-82. Monogr. Syst. Missouri Bot. ceae. Revista del Museo de la Plata Gard.45. (Nueva Serie) 11:193-284. --. 1994. Floristic comparison of six Croat, T. B. 1978. Flora o/Barro Colorado Ecuadorian Florulas. In S. P. Churchill, Island (Panama Canal Zone). Stan­ H. Balslev, & J. Luteyn (eds.), Neo­ ford University Press, Stanford, CA. tropical Montane Forests: Biodiversity ---. 1979. The distribution of Araceae. and Conservation. New York Botani­ In K. Larsen & L. B. Holm-Nielsen cal Garden. (eds.), Tropical Botany, 291-308. Ac­ Croat, T. B. & R. A. Baker. 1978. Studies ademic Press, London. in Araceae II, Anthurium sect. Poly­ --. 1981 (982). A revision of Syn­ phyllium. Selbyana 2:230-238. gonium (Araceae). Ann. Missouri Bot. Croat, T. B. &J. Bogner. 1987. Aroid pro­ Gard.68:565-651. file 11: Syngonium steyermarkii Croat. ---. 1983. A Revision of Anthurium Aroideana 10:20-22. (Araceae) for Mexico and Central Croat, T. B. & N. Lambert. 1986. The Ar­ America. Part 1. Mexico and Middle aceae of Venezuela. Aroideana 90-4): America. Ann. Missouri Bot. Gard. 70: 3-213. 211-240. Croat, T. B. & D. Mount. 1988a. Araceae. ---. 1985a. The large mono cots of In R. Spichiger & ]. M. Mascherpa Panama. In W. G. D'Arcy & M. D. Cor­ (eds.), Flora 0/ Paraguay, 7-46. Con­ rea (eds.), NaturaIHistoryo/Panama. servatoire et]ardin Botanique, Geneva Missouri Botanical Garden, St. Louis, and Missouri Botanical Garden. Im­ MO. primeries Populaires, Geneve. ---. 1985b. A new collection of the ---. 1988b. Araceae.ln R. Spichiger & rare Alloschemone (Poepp.) Engl. & K. J. M. Mascherpa (eds.), Flora del Par­ Krause. Aroideana 8:80-82. aguay. Conservatoire et Jardin Bota­ ---. 1985c. Aroid Profile No. 10. Tac­ nique. Geneva & Missouri Bot. Gard. carum weddellianum. Aroideana 8: Pp.7-46. 94-95. Croat, T. B. & R. S. Sheffer. 1983. The ---. 1986a. A revision of Anthurium sectional groupings of Anthurium (Ar­ (Araceae) of Mexico and Central Amer­ aceae). Aroideana 6:85-123. ica. Part 2. Panama. Monogr. Syst. Bot. Engler, A. 1905. Pothoideae.lnA. Engler Missouri Bot. Gard. 14:1-205. (ed.), Das Pjlanzenreich IV. 23B (Heft ---. 1986b. The distribution of An­ 21), 1-330. thurium (Araceae) in Mexico, Middle --. 1911. Lasioideae. In A. Engler America and Panama. Selbyana 9:94- (ed.), Das Pjlanzenreich IV. 23B(Heft 99. 21), 1-130. --. 1988. Ecology and lifeforms of --. 1915. Anubiadae, Aglaonema­ Araceae. Aroideana 11(3-4):4-55. teae, Dieffenbachieae, Zantedes­ ---. 1989. Important collections of chieae, Typhonodoreae, Peltandreae. New World Araceae. Taxon 37:43-55. In A. Engler (ed.), Das Pjlanzenreich ---. 1990. A comparison of aroid clas­ IV. 23Dc(Heft 64), 1-78. sification systems. Aroideana 13:44-63. --. 1920. Aroideae-Pistioideae. In A. THOMAS B. CROAT, 1994 59

Engler (ed.), Das Pjlanzenreich IV. aceae). Contr. Gray Herb. Harvard 23F(Heft 73), 1-274. Univ. 207:3-100. French, J. C., M. Chung & Y. Hur. cpDNA ---. 1978. The species of the Anthur­ phylogeny of Ariflorae. In P. Rudall et iumwith palmately divided leaves. Sel­ al. (eds.), Monocotyledons: Classifica· byana 2:239-282. tion and Evolution. Royal Botanic Gar· ---. 1980. A new species of Ulearum dens, Kew. (In press). from Brazil. Aroideana 3:101-102. Gomez, A. 1983. A revision of Stenosper­ --. 1981. Notes of Caladium (Ara­ mation (Araceae) in Central America. ceae) and its allies. Selbyana 5:342- Masters Thesis, St. Louis University, St. 377. Louis, MO. Matuda, E. 1954. Las Araceas Mexicanas. Grayum, M. H. 1984. Palynology and phe­ Inst. Biol. Mexico 25:176. nology of the Araceae. Ph.D. Disser­ Mayo, S. J. 1990. History and infrageneric tation, University of Massachusetts, nomenclature of Philodendron (Ara­ Amherst. ceae). Kew Bull. 45(1):37-71. 1986. New taxa of Caladium, ---. 1991. A revision of Philodendron Chlorospatha, and Xanthosoma (Ara­ subgenus Meconostigma (Araceae). ceae: Colocasioideae) from southern Kew Bull. 46(4):601-681. Central America and northwestern Co­ Mayo, S.J.,]. Bogner, & P. C. Boyce. 1994. lombia. Ann. Missouri Bot. Gard. 73: A new cladistic classification of the Ar­ 462-474. aceae. Kew Bull. (In press). --. (In prep.) The Genera Ara­ ---. 1990. Evolution and phylogeny of 0/ ceae. with contributions from J. C. Araceae. Ann. Missouri Bot. Gard. 77: French and R. Hegnauer: illustrations 628-677. by Elenolr Catherine), Royal Botanic --. 1991. Chlorospatha kressii (Ara­ Garden, Kew. ceae) , a new compound-leaved spe­ Nicolson, D. 1983. Sodiro's publications cies from Choco Department, Colom­ on Araceae. Huntia 5(1):3-15. bia. Novon 1:12-14. Ray, T. S. 1981. Growth and heterophylly Hay, A. 1988. Cyrtosperma (Araceae) and in an herbaceous vine, Syngonium its Old World allies. Blumea 33:427- (Araceae). Ph.D. Dissertation, Harvard 469. UniverSity, Boston, MA. --. 1992. Tribal and subtribal delim­ ---. 1986. Growth correlations with the itation and circumscription of the gen­ segment in the Araceae. Amer. ]. Bot. era of Araceae tribe Lasieae. Ann. Mis­ 73:993-1001. souri Bot. Gard. 79:184-205. --. 1987. Cyclic heterophylly in Syn­ Hotta, M. 1970. A system of the family gonium (Araceae). Amer.]. Bot. 74:16- Araceae in Japan and adjacent areas. I. 26. Mem. Fac. Sci. Kyoto Imp. Univ. Ser. Reitz, D. R. 1957. Araceas catarinenses. Bio!. 4:72-96. Sellowia 8:20-70. Jonker-Verhoef, A. M. E. & F. P. Jonker. Renner, S. 5., H. Balslev & L. B. Holm-Niel­ 1953. Araceae. In A. A. Pulle &J. Lan­ sen. 1990. Flowering plants of Ama­ jouw (eds.), Flora o/Suriname, Vol. I, zonian Ecuador, a checklist. AAU Re­ Part 2, 1-80, 380-412. Royal Tropical ports 24:54-58. Institute, Amsterdam. Schott, H. c. 1860. Prodromus Systematis Macbride,J. F. 1936. Araceae in The Flora Aroidearum. Vindobonae. of Peru. Field Mus. Nat. Hist., Bot. Ser. Sodiro, L. 1900. Anturia Ecuadorensia 15:431-434. Nova. Rev. Chilena Hist. Nat. 4(6):77- Madison, M. 1976. Alloschemone and 82. Scindapsus (Araceae). Selbyana 1: 325- ---. 1901a. Anturios Ecuatorianos: di­ 327. agnoses previas. Anales Univ. Centro ---. 1977. A revision of Monstera (Ar- Ecuador 15(108):1-18. 60 AROIDEANA, Vol. 17

---. 1901b. Anturios Ecuatorianos: di· --. 1928. Flora of Panama Canal Zone. agnoses previas. Pp. 1-17. Quito. Contr. U. S. Nat. Herb. 27:1-416. ---. 1902-1903. Anturios Ecuatorianos. --. 1930. Asecondsupplementtothe Anales Univ. Centro Ecuador 15-17: Flora of Barro Colorado Island, Pana· 1903a. [Periodical edition covering pe· rna.}. Arnold Arb. 11:119-129. riod from Jan. 1902 through March --. 1933. The flora of Barro Colorado 1930], see Nicolson (1983). Island. Contr. Arnold Arb. 5: 1-178. --. 1903b. Anturios Ecuatorianos. --. 1937. Flora of Costa Rica. Field Monografia II. Contribuciones al con· Mus. Nat. Hist. 18:131-146. ocimiento de la flora Ecuatoriana. Pp. --. 1944. Araceae. Part II. Fascicle 3. i-xxxii 1-231, 1-7 (Adiciones). Quito. In R. E. Woodson Jr. & R. E. Schery, ---. 1905a. Supplemento I. Anturios Flora of Panama. Ann. Missouri Bot. Ecuatorianos. Anales Univ. Centro Ec· Gard.31:1-60. uador 19-20. Standley, P. C. & J. A. Steyermark. 1958. ---. 1905b. Supplemento I. Anturios Araceae, Flora of Guatemala. Field Mus. Ecuatorianos. Monografia II. Contri· Nat. Hist., Bot. Ser., pp. 304-363. buciones al conocimiento de la flora Strong, D. R. & T. S. Ray, Jr. 1975. Host Ecuatoriana. Pp. I-III, (1-3), 4-100, tree location behavior of a tropical vine (101-112), pIs. A, I-X. Quito. (Monstera gigantea) by skototropism. --. 1905c. Anturios Ecuatorianos. Rev. Science 190:804-806. Chilena Hist. Nat. 9(4·6):191-211, 242- Young, H. J. 1986. Beetle pollination bi· 263, 274-281. ology of Dieffenbachia longispatha ---. 1906. Supplemento II. Anturios (Araceae). Amer.}. Bot. 73:931-944. Ecuatorianos. Anales Univ. Centro Ec· --. 1988a. Differential importance of uador 22(156) :20-39. beetle species pollinating Dieffenba­ --. 1907. Supplemento II. Anturios chia longispatha (Araceae). Ecology69: Ecuatorianos. Pp. 1-23. Quito. 832-844. ---. 1908. Aroideae. In Sertula Florae --. 1988b. Neighborhood size in a Ecuadorensis. Anales Univ. Centro Ec· beetle pollinated tropical aroid: effects uador 22(162):257-278. of low density and asynchronous flow­ Standley, P. C. 1927. The flora of Barro ering. Oecologia 76:461-466. Colorado Island. Smithsonian Misc. Colect.78(8):1-32.