j i A | V\ \ ARNOLD ARBORETUM 1872-1972 CENTENNIAL JOURNAL OF THE ARNOLD ARBORETUM Volume 53 1972 HARVARD UNIVERSITY Dates of Issue No. 1 (pp. 1-140) issued 10 January, 1972. No. 2 (pp. 141-272) issued 19 April, 1972. No. 3 (pp. 273-408) issued 4 August, 1972. No. 4 (pp. 409-582) issued 13 November, 1972. Contents of Volume 53 The Genus Piper (Piperaceae) in New Guinea, Solomon Is- lands, and Australia, WEE-LEK CHEW . The Genera of J uglandaceae i in ‘the Southeastern . States. THOMAS S. ELI Revision of poet (Cruciferae). ELIZABETH A. SHAW . een A Revision of the Neotropical Genus Lisianthius (Gentianaceae). RICHARD E. WEAVER, JR. ; The Malpighiaceae in the Southeastern United States. KENNETH R. ROBERTSON Morphological Studies in Cordyline (Agavaceae) IL. Vegetative Morphology of Cordyline terminalis. J. B. FIsHER and P. B. TOMLINSON Notes on Asiatic-Polynesian-Australian Species of elon Il. B. A. KRUKOFF Statement of Gane ; George B. Hinton, Collector of Plants i in Southwestern Mexico. With portrait and map. JAMES HINTON and J. RZEDOWSKI . The Genera of Geraniaceae in the Southeastern United States. KENNETH R. ROBERTSON On the Origin of the Orchidaceae, ie LESLIE A. GARAY Cytomorphology of Arborescent Moraceae. A. S. HANS . The Phrymaceae in the Southeastern United States. JOHN W. THIERET . A Revision of the Neotropical ‘Genus Lisianthius barca )—( Seema ARD E. WEAVER, . A Revision of the Gets Genus Lisianthius (Gentianaceae ) — (concluded). RICHARD E, WEAVER, JR. . Seeds from Surface Soils ina Tropical Region of Veracruz, Mexico. SERGIO GUEVARA S. and ARTURO GOMEZ-POMPA . Floral Anatomy of Myrtaceae, II. Eugenia. RUDOLF SCHMID . Venation Patterns in ~ Leaves of ‘Ephedra. ADRIANCE §S. FOSTE . 1972 128 140 Dioecism in Citharexylum (Verbenaceae). P. B. TOMLINSON and PRISCILLA FAWCETT. . . . . . 386 Notes on West Indian Orchids, II. LESLIE A. GARAY and HERMAN R. Sw 390 Notes on Tibouchina and Charianthus he sae ia in the Lesser Antilles. RICHARD A. How 399 Three New Species of Gaaienivn (Rutaceae) from Darién Province, Panam DUNCAN M. PORTER . : ie ee ee ee ET ee Se eS The Genera of ae sg in ‘the Southeastern United States. STEPHEN A. SPONGBERG pike RN He egy 7 31 AOD Studies on Bigelowia (Asteraceae), IL. Xylary Comparisons, Woodiness, and Paedomorphosis. LoRAN C. ANDERSON cae eaten gir OO Notes on West Indian Orchids, TIL. LEsLiE A. Garay and HERMAN Ri OWHET 245 es A Bi The Genera of at ammeter in the Southeastern United Stat DUNCAN M. PorTER Ne a ee eee EY The Genus Macrocarpaea (Gentianaceac) in Costa Ric RICHARD E. WEAVER, JR... ig CER. ala ia Cytological Studies on Himalayan eliaceae. P. N. MEnRA, T. S. SAREEN, and P.K. KHosta. . Jae OSs Index . 569 ARNOLD ARBORETUM L672 i672 CENTENNIAL JOURNAL OF THE ARNOLD ston as Volume §3 Number I paencon . 7 HARVARD UNIVERSITY er 1 srt : _ : : hha : Journal of the Arnold-Arboretum _ "Published quarterly in January, April, July, and October by the Arnold Arboretum, Harvard University. Oat es é ce should be ‘seit to Ms. Ellen Bernstein, Divinity Avenue, Cambridge, Massachusetts 02138, JOURNAL OF THE ARNOLD ARBORETUM Voz. 53 JANUARY 1972 NUMBER 1 THE GENUS PIPER (PIPERACEAE) IN NEW GUINEA, SOLOMON ISLANDS, AND AUSTRALIA, 1 WEE-LEK CHEW THIS PAPER, the first of a number to be published seriatim, is the out- come of an attempt to determine the collections made by the Royal So- ciety Expedition to the Solomon Islands in 1965. It soon became clear in the investigation that reference had to be made to the New Guinean species of which more than one hundred have been described or recorded. Thus I was led to examine the very rich collections housed in the herbaria at Bogor (where types and authentic materials of earlier binomials are kept), Brisbane, Canberra, Lae, Singapore, and Sydney. In all, about 560 collections representing thirty-seven species have been studied. This re- sulted in twenty-two species being found conspecific with the fifteen dealt with in this paper. Also some of these have their ranges of distribu- tion extended, while two Australian species have been found to be similar to New Guinean ones. The presentation of the species in this series is very concise. Synonymy and literature references are kept at the minimum and descriptions are reduced to a diagnostic type. Detailed presentation is being reserved for the final comprehensive revision. 1. Piper abbreviatum Opiz in Presl, Rel. Haenk. 1: 157. 1828; Quisumbing, Philip. Jour. Sci. 43(1): 58. 1930. P. bivalvantherum C.DC. Nova Guinea Bot. 8(2): 417. 1910, syn. nov. P. internibaccum C.DC. l.c. 420. 1910, syn. nov. P. fragrans Trelease, Jour. Arnold Arb. 9: 147. 1928, syn. nov. Type. Philippine Islands. Luzon: Haenke s.n. (PR). Icon.: Quisumbing, l.c. fig. 24 & 25. 1930. Dioecious glabrous climbers. Lamina shortly petiolate, elliptic, ovate to broad ovate, ca. 10 cm. long, 5 cm. broad, length:breadth ratio ca. 2:1, glabrous on both sides; apex acute to acuminate; base cuneate to rounded, symmetrical; lateral veins 4 pairs, the lowermost pair very short, aris- ing from the base, the 2nd and 3rd usually long, arising from the midrib 2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 a little above the base, the distal 4th pair very short, often absent, arising from the midrib near the apex. Stipules to 1.5 cm. long, often shorter, usually as long as petioles. Jnflorescences shorter than leaves, peduncular stalks ca. 1.5 cm. long, usually as long as petioles, males thin, to 6 cm. long, females shorter and thicker, ca. 2 cm. long, 0.8 cm. broad. Male flowers 2-staminate; stamens ca. 1 mm. long; anthers reniform to subglo- bose, 2-valved, slightly shorter than filaments; bracts peltate, subsessile, orbicular, the very short pedicel stout, hirsute. Female flowers sessile; stig- mas 3- or 4-fid, sessile; bracts orbicular, peltate, sessile. Fruits sessile, entirely concrescent at maturity. DIsTRIBUTION. Philippines, Java, Celebes, Moluccas, New Guinea, Bis- marck ee and Solomon Islands. West New SOUTHERN. Noord-rivier. Bivakeiland: Versteeg 1058, May (type of ? F Ssinantberie BO); Versteeg 1021, May (type of P. interni- baccum, 80). VoGELKop. Aifat River Valley: van Royen & Sleumer 7052, Oct. (LAE); Djitmau: Versteegh BW. 7367, May (CANB, LAE). N.E. New Guinea. Morose Dist. Lae: Millar GH 99, March (prt) & NGF 9930 (BRI, LAE); Oomsis: Brass 29201, Apr. (CANB, LAE); White NGF. 10474, March (BRI, LAE); Wau-Salamaua Rd.; Womersley & Millar NGF 8485, Jan. (LAE). Papua. CENTRAL Div. Iawarere: Brass 701, Nov. (type of P. fragrans, BRI); near Sogeri: Schodde 2927, Sept. (BRI, LAE). MILNE Bay Dist. Dulubi village: Henty NGF 16960, June (srr, CANB, LAE, NSW); Modewa: Brass 28864, Dec. (LAE). NORTHERN Div. Pitoki village: Hoogland 3973, Sept. (BRI, CANB, LAE). Bismarck Archipelago. NEw Britain. Trans Vudal: Streimann NGF 44396, Aug. (LAE Solomon. Islands. BOUGAINVILLE. Buin: Kajewski 2047 & 2175, Aug. (RI, Nsw); Lake Loloru: Craven & Schodde 145, July (Bri, CANB, LAE); Lavarack & Ridsdale NGF 31475, Feb. (LAE); Schodde & Craven 3771, Aug. (BRI, CANB, LAE). Two types of leaf forms have been observed to occur in the individual plant. Leaves borne on the climbing stems are usually broad to very broad ovate with somewhat cordate bases. Those borne on free side branches tend to be narrow ovate to elliptic and often larger (e.g. Schodde & Craven 3771 and Kajewski 2047). It is on such a variation that De Candolle’s P. bivalvantherum and P. internibaccum as well as Trelease’s P. fragrans are based. The New Guinean and Solomon Islands’ populations differ slightly from those in the Philippines, In the former, the lower part of id lamina is usually furnished with three pairs of ESR veins, «the lowermost pair of which is often very short and faint. In the Philippine eel however, the lowermost pair is altogether absent. I am certain Rumphius’ Pharmacum magnum parvifolium from Ter- nate, which Merrill (1917) failed to interpret, is referable to this species. The characteristic abbreviated infructescence and the venation of the small ovate leaves as clearly illustrated by Rumphius, leaves its identity in no doubt whatsoever. 1972] CHEW, THE GENUS PIPER, 1 3 Piper abbreviatum is not to be confused with P. fragile with which it has many characteristics in common (see note under that species). 2. Piper eriaaiae Sp., Pl.. 29. 1753: CDC. in. DC. Prodr, 16(1): 285. peas adunca (L.) Kunth, Linnaea 13: 633. 1839. Artanthe adunca (L.) Migq. Comm. Phyt. 49. tab. 4. 1840; Syst. Piperac. 449. 1844, Icon.: Migq. lc. tab. 4. 1840; Koord. Exk. Fl. Java 4: 451. fig. 728. 1924. Monoecious shrubby trees. Lamina ovate, ca. 16 cm. long, 5 cm. broad, length:breadth ratio ca. 3:1, hispidulous; apex long acuminate; base lightly asymmetrically cordate: lateral veins 6—7 pairs arising alternately from the midrib, 1 or 2 short ones directly from the base. Petiole slightly longer than the sinus of the lamina base, shorter than the peduncular stalk. Inflorescences often bisexual, usually curved, as long as leaves, with flowers borne in dense spirals; peduncular stalk ca. 2 cm. long. Male flowers 2- or 3-staminate; bracts peltate, somewhat triangular. Female flowers sessile; stigmas 3-fid, sessile. Fruits sessile, obpyramidal, some- what truncate. DIstRIBUTION. Mexico, Central America, northern South America, and West Indies. Naturalized i in many places in Malesia West New Guinea. NorTHERN. Hollandia. Polimac: v.d. Sijde BW 4155, Jan. (LAE). WatcEo IsLanp. Majalibit Bay: van Royen 5278, Jan. (LAE N.E. New Guinea. Morose Dist. Lae Botanic Gardens: Native coll. ‘NGF 4719, Dec. (BO, BRI, LAE, NSW, SING); Huon Peninsula. Pindiu: Hoogland 8847, Apr. (BRI, CANB, LAE Being a native of the New World Tropics, this species bears no rela- tionship whatsoever to any Malesian and Australasian native. Its closest American ally is Piper hispidum which has also been introduced into Java. From the few collections available for my study, P. hispidum seems to be only a less hispid version of P. aduncum. I have made no comments on the type of Piper aduncum. Being a Linnaean species, based on pre-Linnaean ones, its typification is very in- volved, and can only be sorted out in the older herbaria of Western Eu- rope where much pre-Linnaean literature and many collections are still extant. For the present work, I adopt the interpretation of Miquel and De Candolle. 3. Piper amboinense (Miq.) C.DC. in DC. Prodr. 16(1): 347. 1869; Merrill Interpret. Herb. Amb. 182. 1917; Holth. & Lam, Blumea 5: 173. 1942, Chavica amboinensis Mig. Ann. Mus, Bot. Lugd. Bat. 1: 134. 1863, incl. var. latifolia Mig. I.c. 140. 1863. 4 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 53 Type: Moluccas. Amboina: Forsten, May (1). FicureE 1. Dioecious climber, entirely glabrous. Lamina very shortly petiolate; ovate to oblong, somewhat rugose, 30 cm. long, 14 cm. broad, length: breadth ratio ca. 2:1; apex bluntly acuminate; base unilaterally cordate; lateral veins 4—5 pairs, the broader side often with 2 or 3 veins more, the two distal pairs arising from the midrib, the rest directly from the base. Petioles ca. 1 cm. long. Inflorescences shorter than leaves; peduncular stalks longer than petioles. Male flowers 2-staminate; anthers reniform, FicuRE 1. Piper amboinense. 2 twig, from type. 1972] CHEW, THE GENUS PIPER, 1 5 bracts peltate, rounded. Female flowers sessile; stigmas 3-fid, sessile. Infructescences often slightly longer than the leaves. Fruits sessile, some- what cylindrical, free. DistR1IBuTIon. Celebes, Moluccas, and New Guinea. Celebes. MAKAssar PENIN. Pangkadjene: Teysmann 11717 (so). Moluccas. AmBoIna. Batoe Gadjah: Robinson Pl. Rumph. Amb. 58 (Nsw); Ema: Teysmann s.n. (Bo). West New Guinea. Sine loc.: Boorsma 2 (BO). This species is very close to Piper majusculum from which it differs principally in the shorter female inflorescences, the less cordate lamina base, the very short petioles, and the fewer, differently patterned lateral veins. The species has been erroneously recorded for Borneo by Holthuis and Lam (1942) who based their claim on one of Elmer’s collections dis- tributed as Piper amboinense. I have seen this collection and agree with Merrill (in Pl. Elm. Born. 40. 1929) that this is a misidentification. Elmer’s specimen belongs to P. majusculum. The New Guinean collection cited above seems to be not only the first record but also the only one for that island. I have included in the cita- tion materials from Celebes and Moluccas as these collections had ac- tually been authenticated by De Candolle and, as far as Teysmann’s ma- terials are concerned, actually by Miquel himself. 4. Piper caninum Bl. Verh. Bat. Genoot. 11: 214. 1826; C.DC. in DC. Prodr. 16(1): 341. 1869; Warb. Bot. Jahrb. 13: 283. 1891; K. Sch. & Laut. Fl. Schutzgeb. 257. 1900. Cubeba canina (Bl.) Miq. Syst. Piperac. 293. 1843. P. banksii Miq. Versl. Med. Akad. ser. 9. 2: 61. 1868 ; C.DC. Bot. Jahrb. 55: 207. 1918; Benth. Fl. Austral. 6: 205. 1873, syn. nov. P. lauterbachii C.DC. in K. Sch. & Laut. FI. Schutzgeb. 255. 1900, sym. nov. P. macrocarpum C.DC. Nova Guinea Bot. 8(2): 421. 1910, syn. nov. P. pubipes C.DC. I.c. 421. 1910, syn. nov. P. kietanum C.DC. Denkschr. K. Akad. Wiss. M.-N. KL. Wien 89: 527. 1914, Syn. nov, P. flavifructum Trelease, Jour. Arnold Arb. 9: 147. 1928, syn. nov. Type. Java: Blume (1). Icon.: Bl. Verh. Bat. Genoot. 11: fig. 26. 1826; Quisumbing, Philip. Jour. Sci. 43(1): 121. fig. 62. 1930. Dioecious climber. Twigs and petioles pilose. Lamina shortly petiolate, Ovate, ca. 14 cm. long, 7 cm. broad, length:breadth ratio ca 2:1, glabrous on the upper side, subglabrous to pubescent on the lower side; apex long acute to acuminate; base slightly asymmetrically rounded to shallowly cordate; lateral veins 3-4 pairs, the distal pair arising from the midrib at ca. 2~3 cm. from the base, the rest directly from the base. Petioles ca. 6 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 2 cm. long. Female inflorescences up to as long as leaves, not densely flowered; peduncular stalk pilose, slightly longer than petioles. Female flowers sessile when young; stigma 4-fid, reflexed, sessile; bracts peltate sessile. Infructescences often longer than leaves. Fruits ca. 4 * 3 mm pedicellate, ovoid to globose. ? DistRIBUTION. Widely distributed in Malesia through New Guinea to the Solomon Islands and Australia. West New Guinea. CENTRAL. Sabang: Branderhorst 330 (isotype of P. macro- carpum, BO). Sine loc.: Atasrip 64, Exped. Wichmann (isotype of P. pubipes, BO); Boorsma 4 (Bo); Branderhorst 425, June (Bo). -E. New Guinea. Mapanc Dist. Aiome: Frodin NGF 26954, June (LAE). Morose Dist. Finschhafen: Lauterbach 1337 (isotypes of P. lauterbachii, Bo, BRI, SING); Oomsis: Brass 29175, April (LAE). SEPIK Dist, Ambunti: Hoogland & Craven 10114, May (srt, Lag). Papua. CENTRAL Div. Dieni: Brass 3926 & 3966, May (BRI); Domara River: Brass 1644, June (isotype of P. flavifructum, prt), MILNE Bay Drv. Normanby Island. Waikaiuna: Brass 25390, April (LAE). Bismarck Archipelago. New Britain. Gasmata: Sayers NGF 24171 , March (LAE, NSW). Solomon Islands. BouGaInvitte. Buin: Kajewski 1782, May (BRI). GUADAL- CANAL. Mt. Tutuve: Kajewski 2646, May (srt), New Gerorcta, Vaimbu River: Cowmeadow BSIP 4762, Sept. (LAE). SAN CRISTOBAL. Maru Bay: Gafui BSIP 12829, Nov. (LAE). Australia, QUEENSLAND. Cape Tribulation: Mason s.n., July (Nsw); Rock- ingham Bay: Dallachy s.n. (Bo, Nsw). This widespread and consequently very well known species needs no introduction. Though quite variable in lamina form and infructescence size, it is easily recognized throughout its entire range of distribution. I have examined the types and authenticated materials of the synonyms cited above and have found that they come within the circumscription of this species as established by Blume and adopted by De Candolle and subsequent authors. 5. Piper decumanum [Rumph.] Linnaeus in Stickman, Herb. Amb. 19. 1754; Amoen. Akad. 4: 128. 1759; Merrill, Interpret. Herb. Amb. 181. 1917. Sirium decumanum Rumph. Herb. Amb. 5: 45. tab. 27. 1747. Chavica majuscula (Bl.) Miq. Syst. Piperac. 271. 1843, partim, Ch. imperialis Miq. Ann. Mus. Bot. Lugd. Bat. 1: 134, 140. 1863. Ch. rumphii Mia. l.c. 141. 1863. P. forstenii C.DC. in DC. Prodr. 16(1): 348, 1869; Nova Guinea Bot. 8(6): 1007. 1914; Gibbs, Phyt. Fl. N. Guinea 207, 1917. P. rumphii (Migq.) C.DC. in DC. Prodr. 16(1): 354. 1869, P. majusculum auct. non Bl., C.DC. ibid. 350. 1869, partim. Type. Moluccas: Rumph. Herb. Amb. 5: tab. “7 Baw by & 1972] CHEW, THE GENUS PIPER, 1 7 Icon.: Rumph. /.c.; Quisumbing, Philip. Jour. Sci. 43(1): 43. fig. 14. 1930. Dioecious climber, completely glabrous. Lamina very shortly petiolate, ovate, often rugose, ca. 40 cm. long, 18 cm. broad, length:breadth ratio ca. 2:1; apex acute to bluntly acuminate; base asymmetrically cordate to auriculate; lateral veins 4—6 pairs, the distal pair arising alternately from the midrib, the rest from the base, the larger side usually with 2-3 veins more. Petioles often shorter than the auriculate base, sheathing at the lower half. Male inflorescences about as long as the leaves. Female inflorescences much longer; peduncular stalks as long as petioles. Female flowers sessile; stigmas 2-lipped to 3-fid, borne on tapering styles. In- fructescences to 60 cm. long. Fruits ca. 3 X 1.5 mm., oblong, sessile, congested. DIstTRIBUTION. Philippine Islands, Celebes, Moluccas, and New Guinea. West New Guinea. NorTHERN. Sawaii: Gjellerup 599, Aug. (BO). VOGELKOP PENINSULA. Manokwari: Schram BW 12924, March (LAE). WAIGEO ISLAND. Radjah Ampat: van Royen 5215, Jan. (CANB, LAE). N.E. New Guinea. Mapanc Dist. near Aiome: Womersley NGF 24752, June (LAE). SEPIK Dist. Ossima: Streimann & Kairo NGF 39196, Jan. (LAE). Linnaeus based Piper decumanum entirely on Rumphius’ Sirium decu- manum from the Moluccas. Unfortunately, he later (1762) extended the concept to include an entirely different species described by Plumier from the American Tropics: unfortunately, because this led to some confusion on the use of the name P. decumanum. The error was noticed by Miquel (1843) and De Candolle (1869) who quite rightly separated Plumier’s species from the Moluccan one; bu both erred in other respects. They incorrectly retained the epithet for the American species and placed our Moluccan one under Piper majusculum, an entirely different species (see notes there). It was not until 1917 that the matter was corrected by Merrill. ; The basal leaves of this species differ greatly from the upper ones in being almost symmetrically ovate with deeply symmetrically cordate bases. Both Sirium decumanum of Rumphius and Chavica rumphii of Miquel are based on plants with such leaves. I have seen authenticated materials of Piper forstenii, under which name De Candolle recorded our species from New Guinea, and I am satisfied that P. forstenii was correctly referred to P. decumanum by previous authors. 6. Piper fragile Benth. Lond. Jour. Bot. 2: 234. 1843; C.DC. in DC. Prodr. 16(1): 358. 1869; Nova Guinea Bot. 8(2): 417. 1910. Chavica benthamiana Miq. Syst. Piperac. 233. 1843. Type. New Guinea: Hinds s.n. (K). Icon.: Quisumbing, Philip. Jour. Sci. 43(1): 100. fig. 50. 1930. 8 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Dioecious climber, entirely glabrous. Lamina shortly petiolate, ovate, ca. 8 cm. long, 6 cm. broad, length: breadth ratio ca. 4:3; apex acute; base rounded, often subpeltate, rather coriaceous; lateral veins 2 pairs, often with one more very short and faint pair at the very base, all aris- ing from the base. Petioles ca. 1/4 the length of lamina. Inflorescences shorter than leaves; the females very short; densely flowered in both; peduncular stalks as long as petioles; bracts peltate, sessile. Male flowers with 2 stamens; filaments very short, oblong; anthers ovoid, as long as filaments, 2-valved. Female flowers sessile; stigmas 4-fid, reflexed, sessile. Infructescences shorter than leaves, often congested and curved. Fruits ca. 3 mm. diam., almost globose, concrescent at the lower half. DistRIBUTION. Philippines, Moluccas, New Guinea, and Solomon Is- lands. West New Guinea. Sine loc.: Boorsma 9 (Bo). .E. New Guinea. MApanc Dist. Bagabag Island: Vandenberg & Mann NGF 42263, June (LAE); Serang: Millar NGF 37693, July (LAE). West SeprK Dist. Selio Island: Millar & Vandenberg NGF 40900, June (LAE). Papua. MILNE Bay Div. Wagalasa Island: Mann & Osborn NGF 43037, March (LAE). Solomon Islands. BouUGAINVILLE: Kajewski 1595, March (prt). GUADALCANAL. Rove West: Nakisi BSIP 7919, June (LAE); Corner RSS 193, Nov. (SING). This species can easily be confused with Piper abbreviatum with which it has a few characteristics in common. In both, the lamina are small and ovate with few veins, the female inflorescences rather short and “abbre- viated,” and the fruits concrescent. A detailed examination of these characters, however, reveals important differences between the two species. In P. fragile, all the lateral veins arise from the very base which is often subpeltate; whereas in the other species this is not so. In the female flowers and fruits, the sigmas are 3-fid and the fruits entirely concrescent in P. abbreviatum; but they are 4- or 5-fid and the fruits only partially con- crescent in P. fragile. Piper varibracteum C.DC. (1910) of the Philippines probably is a synonym of Piper fragile, but the type material of it must be examined before its relationship to P. fragile can be assessed with certainty. 7. Piper gibbilimbum C.DC. Nova Guinea Bot. 8(2): 416. 1910; Bot. Jahrb. 55: 211. 1918. P. squamuliferum C.DC. Nova Guinea Bot. 8(6): 1009. 1914, syn. nov. Type. West New Guinea. Noord-rivier, near Alkmaar: Versteeg 5115, July (Bo). FIicurE 2 Dioecious trees. Lamina very shortly petiolate, asymmetrically ovate ca. 15 cm. long, 9 cm. broad, length:breadth ratio ca. 3:2 to 5:3 gen- erally glabrous, sometimes lightly hirsute on the veins on the undet side rown puncticulate on both surfaces; apex acuminate; base unilaterally 1972] CHEW, THE GENUS PIPER, 1 Or YQ eA » ze \ x ci ay i He : | Ni ts Vv Was f e. nz Ficure 2. Piper gibbilimbum. A, 2 twig; B, stigma, from Brass 12252 to asymmetrically cordate, those of the basal leaves usually deeply sym- metrically cordate; lateral veins 3 or 4 pairs, all arising from the base, the broader side often with one more faint vein. Stipules thin, ca. 2 cm. long, usually as long as petioles. Inflorescences usually shorter than adult leaves: peduncular stalks usually as long as petioles, male and female equal in length. Male flowers 2-staminate; stamens ca. 0.4 mm. long; 10 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 anthers reniform, 2-valved, on extremely short filaments; bracts orbicular, subsessile, peltate, the very short pedicel lightly hirsute. Female flowers sessile; stigmas 3-fid, sessile; bracts long pedicellate, peltate, orbicular, the pedicel very lightly hirsute. Fruits sessile, free, crowded. DISTRIBUTION. New Guinea. West New Guinea. NorTHERN. Balim River: iat 11671, Dec. (80); Iden- burg River: Brass 12252, Jan. (BO, BRI). SOUTHERN. Noord-rivier. Alkmaar Versteeg 1515, July (type of P. gibbilimbum, sys Mt. Hellwig: Roemer 1268 (type of P. squamuliferum, BO); Wissel Lakes: Enarotali: Vink & Schram BW 8593, May (LAE NE. New Gninea. EASTERN HIGHLANDs. Chimbu Divide: Womersley et al. NGF 6097, Nov. (BRI, LAE); Kundiava: Millar & Nicholson NGF 13847, Sept. (BRI, CANB, LAE, NSW). MApANG Dtsv. Kar Kar Isl.: Mann & Vandenberg NGF 43192 (LAE). Morose Dist. Edie Creek: Millar & Holttum NGF 15824, Aug. (BRI, LAE, NSW); Kikiepa: Womersley & Thorne NGF 11860, May (RI, BO, CANB, LAE, NSW). SEPIK Dist. Bliri River: Darbyshire & Hoogland 8254, July (zo, CANB, LAE); Oksapmin: Henty et al. NGF 38972, Oct. (BRI, LAE, NSW). WESTERN HIGHLANDS. Nondugl: Womersley NGF 4349, April (BRI, LAE). Papua. MILNE Bay Dist. Maneau Range: Mt. Dayman: Brass 23235, July (CANB, LAE). SOUTHERN HIGHLANDS. Kuave: Pullen 2778, July (LAE); Lake Kutubu: Gray NGF 8109, May (BRI, LAE, NSw); Schodde 2355, Sept. (LAE). Piper squamuliferum belongs to this species. The presence of bracts at the base of the ovary, a characteristic claimed to be exclusive to P. squamuliferum, has also been observed in P. gibbilimbum, though not as constantly as in the former species. This tree species is one of the very successful colonizers of disturbed forests in New Guinea. Spread over an altitudinal range of from 800 to 2600 meters, it is found to be particularly common at about the 1500 meter contour. Piper gibbilimbum is close to P. plagiophyllum from which it can be distinguished by the different venation pattern and the longer peduncular stalks. 8. Piper macropiper Pennant, Outl. Globe 4: 242. 1800; Merrill, Jour. Arnold Arb. 29: 191. 1948 Sirium arborescens tertium Rumph. Herb. Amb. 5: 46. tab. 28, fig. 1. 1747. P. arborescens Rumph. Roxb. Hort. Beng. 80. 1814, Fl. Ind. 1: 161. 1820; Mig. Syst. Piperac. 320. 1844; C.DC. in DC. Prodr. 16(1): 358. 1869; P. miniatum Bl. Verh. Bat. Genoots. 11: 166. 1826; K. Sch. & Laut. FI. Schutzgeb. 261. 1900; C.DC. Nova Guinea Bot. 8(2) 415. 1910. Chavica miniata (Bl.) Mig. Syst. Piperac. 234. 1843 r. pendulum Warb. Bot. Jahrb. 13: 283. 1891, syn ? P. quinquenervium Warb. I.c. 284. 1891, syn.? P. brassii Trelease, Jour. Arnold Arb. 9: 147. o SYM, NOV. P. morianum Trelease, I.c. 148. 1928, syn. no 1972] CHEW, THE GENUS PIPER, 1 11 Type. Moluccas: Rumph. Herb. Amb. 5: tab. 28, fig. 1. 1747. Icon.: Bl. Verh. Bat. Genoots. 11: fig. 6. 1826, as P. miniatum; Miq. Illus. Piperac. tab. 28. 1847, as C. miniata; Olidinmnblig Philip. Jour. Sci. 43(1): 21. 1930, as P. arborescens. Dioecious climber; nodes often thickened. Lamina very shortly peti- olate; ovate, ca. 16 cm. long, 6 cm. broad, length:breadth ratio from 2:1 to 5: 2, glabrous on the upper side, glabrous to pubescent on the lower; apex acuminate: base round to shallowly cordate, very slightly asymmetri- cal, one side often produced into a minute auricle; lateral veins 2 or 3 pairs arising from the base, prominent. Stipules to 2 cm. long, early caducous. Inflorescences much longer than the leaves, often to 20 cm.; peduncular stalks ca. 3 cm. long. Male flowers 2- or 3-staminate; stamens very short; anthers reniform, 2-valved, often appearing truncate. Female flowers sessile, free, crowded; stigmas 3-fid, sessile. Fruits sessile, ovoid to oblong, crowded, not concrescent at maturity. Distrisution. Widely distributed from India, Boone through Malesia and New Guinea to the Solomon Islands in Melanesia West New Guinea. NorTHERN. Idenburg River: Brass 13710 & 13976, Jan. (BRI) N.E. New Guinea. Morose Dist. Finschhafen: Hellwig 624, Apr. (Bo); Lau- terbach 1489 (Bo, BRI); near Mumeng: Hartley 9765, Jan. (BRI, CANB, LAE, NSW). EASTERN HIGHLANDS. Kainantu: Henty NGF 29361, March (BRI, CANB, NSW). Papua. CENTRAL Div. Iawarere: Brass 683, Nov. (isotype of P. brassii, BRI); Mori River: Brass 1562, May (isotype of P. morianum Solomon Islands. BOUGAINVILLE. Buin: Craven & Schodde 201, Aug. (BRI, LAE); Kajewski 2133, Aug. (BO, BRI, NSW, SING). GUADALCANAL. Tutuv e Mts.: Kajewski 2579, Apr. (prt). SAN CRISTOBAL, Waimamura: Brass 2581, pn (BRI). SANTA YSABEL. Meringe: Brass 3347, Dec. (BO, BRI); Whitmore BSIP peor, Oct. (LAE). Variation in size, shape and vestiture of the lamina and twig is great in Piper macropiper. The length of the lamina, for instance, varies from 8 to 20 cm. In shape, the lamina can be very narrowly to very broadly ovate with a fairly noticeable cordate base. As for vestiture, the variation extends from completely glabrous to fairly densely pubescent. As all man- ner of gradations have been observed between the extremes, and as many of these variations have actually been observed to occur on an individual, Trelease’s species as well as the numerous varieties recognized by pre- vious authors are not maintained here. Nevertheless, this species re- mains one of the most distinctive in the genus. In the Herbarium at Bogor, Indonesia, there is a collection of Lauter- bach no. 1489 and one of Hellwig 624 named by De Candolle as Piper pendulum and P. quinquenervium respectively. These collections have been found by me to belong to Piper macropiper, and the names they bear are placed in the synonymy above, each with a query because their types have not been examined. 12 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 This species has suffered two name changes, the latest of which is most unfortunate because it was caused by the discovery of an overlooked pub- lication (vide Merrill 1948). 9. Piper majusculum BI. Verh. Bat. Genoots. 11: 210. 1826; C.DC. in DC. Prodr. 16(1): 350. 1869, partim; Quisumbing, Philip. Jour. Sci. 43(1) : 45. 1930. Chavica majuscula (Bl.) Mig. Syst. Piperac. 271. 1843, partim. P. insignilimbum C.DC. Candollea 2: 219. 1925, syn. nov. Type. Java. Mt. Salak: Blume (tL). Icon.: Bl. Verh. Bat. Genoots. 11: fig. 25. 1826; Quisumbing, Philip. Jour. Sci. 43(1): 46. fig. 16. 1930. Dioecious climber. Lamina shortly to very shortly petiolate, ovate to pentagonal, ca. 30 cm. long, 15 cm. broad, length-breadth ratio ca. 2:1, sparsely hirsute on the midrib below; apex shortly acuminate; base asym- metrically cordate, one or both sides auriculate; lateral veins ca. 6 pairs, all arising from the midrib, the auricles usually with 2-3 more short and rather faint veins arising from the base. Petioles ca. 2 cm. long, usually shorter than the sinus of the lamina base. /nflorescences very much longer than leaves, very densely flowered; peduncular stalks longer than petioles. Female flowers sessile; stigmas 3-fid, sessile; bracts peltate, sessile, gla- brous. /nfructescences ca. 35 cm. long, 2 cm. diam., pendulous. Fruits concrescent, DIsTRIBUTION. Widely distributed in Malesia, through New Guinea to the Solomon Islands. t New Guinea. Brak. Oregontrail: Vink BW 12052, July (cans, ref). Vocerxor PENINSULA. Manokwari: Versteegh BW 7572, April (LAE); Nabire Kanehira & Hatusima 11531 & 12643 (Bo). Sine loc.: Boorsma 3 & 13 (syn- types of P. insignilimbum, Bo). Papua. GuLF Div. Kikori. Biara: Gray & Floyd NGF 8050, July (sri, LAE). Solomon Islands. GUADALCANAL. Mt. Mambulu: Nakisi BSIP 7914, June (Lak). SANTA YSABEL. Maloku: Beer BSIP 7259, May (LAE). I have examined the holotype of Piper insignilimbum and having com- pared it with authentic materials of Blume’s species, I am quite satisfied that it is conspecific with Piper majusculum. The only difference I can find between the two is in the petiole which seems to be slightly shorter in the New Guinea and Solomon Island plants. It is not surprising that Miquel and De Candolle confused Rumphius’ Sirium decumanum (= Piper decumanum) with Piper majusculum. In the first place, the lamina of the latter are rather similar to the basal leaves of P. decumanum in that both are cordate and often auriculate at the base. Rumphius’ illustration clearly depicts a plant with such basal leaves. Secondly, both species are notable by their greatly elongated fe- male inflorescences which, being densely covered with flowers, become 1972] CHEW, THE GENUS PIPER, 1 13 pendulous at maturity. The two species, however, differ clearly in the following characters: (a) the stigmas in P. majusculum are 3-fid, small and sessile, in P. decumanum they are 2-lipped or 3-fid and borne on ta- pering styles, (b) the fruits are concrescent in P. majusculum but free in P, decumanum. 10. Piper mestonii Bailey, Rep. Exped. Bellenden-Ker 54. 1889; Lam, Vegetationsbilder 15(5/6): 5. 1924. P. rueckeri K. Sch. & Hollrung, Fl. Kais. Wilh. Land. 36. 1889; C.DC. Bot. Jahrb. 55: 204. 1918, syn. nov. P. bilobulatum C.DC. Nova Guinea Bot. 8(2): 418. 1910, type only, syn. nov. P. rhodocarpum Trelease, Jour. Arnold Arb. 9: 149. 1928, syn. nov. Type. Australia. Queensland. Harvey’s Creek. Russell River: Meston & Bailey s.n. (isotype of P. mestonii, Nsw). Icon.: Bailey, Compr. Cat. Queensl. Pl. tad. 12, fig. 402. 1913. Dioecious climber. Twigs glabrous, nodes fairly prominent. Lamina moderately petiolate, ovate, ca. 20 cm. long, 14 cm. broad, length: breadth ratio ca. 3:2, glabrous; apex acute to bluntly acuminate; base rounded to cordate at maturity; lateral veins ca. 4 pairs, the distal pair arising alternately ca. 1/3 up the midrib, the 2nd pair slightly above the base, the rest directly from the base. Stipules much shorter than petioles. Jn- florescences shorter than leaves, peduncular stalks glabrous, as long as petioles; males thinner and shorter; bracts orbicular, sessile. Female flowers sessile; stigmas 2-lipped, on long attenuate styles. Jnfructescences ca. 12 cm. long, 1.5 cm. diam., brick red at maturity. Fruits sessile, com- pletely concrescent at maturity. DIstRIBUTION. New Guinea and Australia. West New Guinea. NortHERN. Begowri: Gjellerup 219, June (Bo); Hollan- dia: Kalkmann BW 3420, March (cans, LAE); Koster BW 1175, Jan. (cANB, LAE). SOUTHERN. Noord-rivier. Alkmaar: Versteeg 1533, Aug. (isotype of P. bilobulatum, Bo). N.E. New Guinea. Morose Dist. Umi River. Kalapit: Millar & van Royen NGF 15658, Jan. (BRI, CANB, LAE, NSW). Mapanc Dist. Bili Bili River: Hoog- land 4847, June (BRI, CANB, LAE). SEPIK Dist. Augustafluss: Hollrung 260 (iso- type of P. rueckeri, Bo); Leitre: Millar & Vandenberg NGF 40944, June (LAE). Papua. CENTRAL Div. U-uma River: Brass 1427, May (isotype of P. rhodo- carpum, BRI). WESTERN Div. Fly River: Brass 8276, Nov. (BRI, CANB); Palmer River: Brass 7258, July (srt). Australia. QUEENSLAND. Cairns Dist. Harvey Creek: Flecker 12449, Aug. (Nsw); Meston & Bailey s.n, (isotype of P. mestonii, Nsw). I am confident that this New Guinean species, Piper mestonii, is none other than the Queensland Long Pepper of Australia which Lam (1924) recorded as being very common along rivers in New Guinea. I have ex- amined all the type materials of the names listed above in synonymy and have found that they differ from each other only in lamina form. 14 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The type of Piper rueckeri has ovate leaves, that of P. bilobulatum nar- rowly ovate to small elliptic leaves with very cuneate bases, while P. rhodocarpum has lamina more cordate than P. mestonii. As numerous re- recent collections have been found to contain all these lamina forms on one plant (e.g. Schiefenhovel 107; Hoogland & Craven 10575, & Pullen 2818), it is here concluded that these leaf forms are but manifestations of the various developmental stages in the life history of one species. There is a superficial resemblance between Piper mestonu and P. ver- steegii, particularly in lamina shape, size, and venation. Because of this, De Candolle misidentified Gjellerup’s collection no. 219 as P. versteegii (vide: C.DC. Nova Guinea Bot. 8(6): 1007. 1914). The two species are however quite distinct (see notes under P. versteegii) and are easily told apart by many other characters as was done by De Candolle subsequently in 1923 in his monographic key. Piper mestonii seems close to P. reinwardtianum of the Moluccas from which it differs only in the stigmas being 2-lipped instead of 3- or 4-fid. The Papuans call this species ‘“O-O-O”’! 11. Piper plagiophyllum K. Sch. & Laut. Fl. Schutzgeb. 260. 1900; Bot. Jahrb. 55: 206. 1918; Ridl. Trans. Linn. Soc. Bot. ser. 2.9: 143. 1916. Type. N.E. New Guinea. Oertzen Gebirge: alt. 300 meters, ee bach 2143, May (B). FIGURE 3 Dioecious trees ca. 5 meters high, entirely glabrous. Lamina very shortly petiolate, asymmetrically ovate, ca. 17 cm. long, 10 cm. broad, length: breadth ratio ca. 3:2; apex short acuminate; base asymmetrically rounded to unilaterally cordate, the lower ones properly cordate; lateral veins 4 pairs, the distal pair arising alternately from the midrib, the rest from the base, the lowermost pair usually very short and faint. Stipules 4—5 cm. long, much longer than petioles. /nflorescences of equal length in both sexes, much shorter than adult leaves, but longer than petioles; peduncular stalks very short, less than 1 cm. Male flowers 2-staminate: stamens ca. 1 mm. long; anthers reniform, large, longer than filaments; bracts orbicular, peltate, sessile, hirsute at the base. Female flowers ses- sile; stigmas 3-fid, sessile; bracts as in the males. Fruits sessile, free crowded. DISTRIBUTION: New Guinea. West New Guinea. VocELKop PENINSULA, Arfak Mts. May (LAE). N.E. New Guinea. Morose Dist. Busu River: mages 12187, ise (BRI, ANB, LAE); Busu Road: peel NGF 11912, Feb. (Bo, CANB, LAE, NSW); Buti- bie River: Hartley 9620, Dec. (BRI, CANB, LAE); K aindi: Brass 29542, May (CANB); Rebel: Schlechter 16791 (Bo); Mt. Torricelli: Schlechter 14604 (BO); re geo rscacieg 16252 (Bo MILNE B 0). AY Div. between Agamoia & Ailuluai Mts.: Brass 27128, : Versteegh BW 12665, zs — oS — Va Nae v3 ~ a LLL) / Ficure 3. Piper plagiophyllum. 2 twig, from Henty NGF 11912. 16 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 June (LAE). Goodenough Island: Brass 24735, Oct. (LAE). Kako: Weinland 256 BO, BRI) Another tree species, Piper plagiophyllum is also found to be common, particularly in disturbed habitats or abandoned gardens. It has quite a broad altitudinal range, from sea level to about 2,000 meters, and is es- pecially abundant in damp areas at about the 1,000-meter contour. Schumann and Lauterbach compared this species with Piper betle al- though they admitted that the two species differ markedly from each other in habit of growth and shape of lamina. Our species is actually very close to P. gibbilimbum from which it differs in two characters, one pair of lateral veins issues from the midrib, and the peduncular stalks are extremely short. 12. Piper sclerophloeum C.DC. Denkschr. K. Akad. Wiss. M.-N. Kl. Wien 89: 530. 1914, includ. var. scandens C.DC. l.c. Tyre. Solomon Islands. Buka, prope pagum Jeta, septembri, K. et L. Rechinger 4387 (w Icon.: C.DC. L.c. 530. tab. 4, fig. 6a. 1914. Dioecious, climbing, often erect. Twigs glabrous at maturity. Lamina moderately petiolate, asymmetrically ovate, ca. 23 cm. long, 16 cm. broad, length: breadth ratio ca. 4:3, the upper side glabrous, the lower side hir- tellous on veins, soon glabrous; apex shortly acuminate or acute; base asymmetrically rounded to unilaterally cordate; lateral veins 4 pairs, the distal pair alternate, arising ca. 3 cm. up the midrib, the 2nd pair at ca. 1 cm. and the rest arising from the base, the broader side often with one more vein. Petioles ca. 3 cm. long, usually equalling stipules. [nflorescences shorter than leaves, with peduncular stalks shorter than petioles; the males ca. 10 cm. long, 0.6 cm. diam., the females 7 cm. long, 1.5 cm. broad. Fe- male flowers sessile; stigma bifid, borne on gradually tapering styles. Fruits sessile, not concrescent. DistTRIBUTION: Solomon Islands. Solomon Islands, BoUGAINVILLE. Buin: Craven & Schodde 513, Sept. (BRI, rare 3917, April (LAE), GUADALCANAL. Duidui Area: Fa’arodo BSIP 12116, t. (LAE); Sirute BSIP 10091, June (LAE). SAN CRISTOBAL. Waimamura: Bran oe. 2583, & 3139, Aug.—Sept. (Brt). SANTA YSABEL. Tiratona: Brass 3345, Dec. (Bo, BRI). This species is reported to be common in both primary and secondary forests below 1,000 meters in altitude. Kajewski recorded the plants as robust climbers that often grow into shrubby trees needing no support. For this reason, De Candolle’s variety scandens is not maintained. The shape of the leaf-base varies considerably in this species. While 1972] CHEW, THE GENUS PIPER, 1 17 most of the collections studied have unilaterally cordate bases, a number of them have been observed to have bases which range from asymmetrical- ly rounded to fairly deeply cordate. Our species resembles somewhat P. mestonii from which it differs in: (1) fruits not concrescent, (2) peduncular stalks shorter than petioles, and (3) lamina generally asymmetrical. 13. Piper stenocarpum C.DC. Ann. Cons. Jard. Bot. Genéve 2: 270. 898. Type. Papua. Central Div. Mt. Yule: G. Belford ?, Dec. 1890, MEL 1010379 (MEL). FIGURE 4. Dioecious climber, entirely glabrous. Lamina very shortly petiolate, ovate to narrowly ovate, ca. 13 cm. long, 5 cm. broad, length:breadth ratio ca. 5:2; apex long acuminate; base usually asymmetrical, the shorter side cuneate, the other rounded; lateral veins 2 or 3 pairs, the distal pair aris- ing from the midrib 2-3 cm. from the base, the second also from the midrib very near the base, the basal pair, if present, always from the base and usually very faint. Stipules very small. Inflorescences longer than leaves, very slender, especially the males; peduncular stalks much longer than petioles; bracts sessile, somewhat oblong. Male flowers 2-staminate; filaments much longer than anthers at maturity. Female flowers sessile; stigmas very short and bluntly 2-lipped, on long attenuate styles. Jn- fructescences to 20 cm. long, 1 cm. diam., bright red at maturity. Fruits sessile, completely concrescent at maturity. DIsTRIBUTION. New Guinea. -E. New Guinea, CENTRAL HIGHLANDS. Korofunota: Womersley & Floyd NGF 6934, Nov. (BRI, LAE). EastERN HIGHLANDS. Daulo: Pullen 407, Aug. Finnesterre Mt.: Womersley & Thorne NGF 11891 (prt, LAE). SEPIK Dist. Nerenanip village: Frodin NGF 28509, Sept. (LAE). WESTERN HIGHLANDS, Mt. Kum: Womersley NGF 43611, March (LAE). Papua. CENTRAL Div. Mt. Albert Edward: van Royen NGF 30092, Jan. (srt, LAE); Mt. Yule: MEL 1010379, coll. G. Belford ? (type, MEL). MILNE Bay Dist. Goodenough Island: Brass 25014, Oct. (LAE). NORTHERN Div. above the Gap: Carr 13793 & 13815, Dec. (CANB). SOUTHERN HIGHLANDS. Near Kiburu: Schodde 1397 (Lak). . Piper stenocarpum is clearly a highland relative of P. mestonii, the Queensland Long Pepper now found to be common in the lowlands and foothills of mountains in New Guinea. The characteristics that the two species have in common are the thin and long male inflorescence, the stamens borne on fairly thick and elongated filaments, the stigmas 2-lipped and borne on long attenuate styles, and the fruits completely concrescent at maturity. They are, however, easily distinguished from each other thus: in P. stenocarpum the inflorescences are much longer than the leaves, the peduncular stalks longer than petioles, and the stigma very shallowly 2- 18 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FIGURE 4. Piper stenocarpum. A, 2 twig; B, stigma, from Pullen 407. lipped, whereas in P. mestonii the inflorescences are shorter than the leaves, the peduncular stalks invariably of the same length as the petioles, and the stigmas very deeply 2-lipped. The altitudinal range of our species extends from about 1500 to 3500 meters and overlaps that of Piper mestonii by about 500 meters at its lower limit. It has been reported to be commonest in fagaceous forests at about the 2200-meter contour. The type collection was possibly gathered by George Belford who went 1972] CHEW, THE GENUS PIPER, 1 19 on an expedition to Mount Yule in Dec. 1890 and subsequently sent his material to Baron von Mueller in Melbourne. 14. Piper versteegii C.DC. Nova Guinea Bot. 8(2): 415. 1910; Can- dollea 1: 179. 1923. P. rhizocaule Trelease, Jour. Arnold Arb, 9: 149. 1928, syn. nov. YPE. West New Guinea. Noord-rivier; near Bivak-eiland: Versteeg 1136, March (lectotype in Bo, L). FicureE 5. A ) a — ZS; ee ees IGURE 5. Piper versteegii. A, leaf and 2 twig, from Brass 14042; B, stamen, “Selagie van Royen NGF 16246; a stigma, from Brass 14042. 20 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Dioecious climber, entirely glabrous. Lamina fairly long petiolate, ovate to very broadly ovate, ca, 25 cm. long, 18 cm. broad, length: breadth ratio 2:1 to 1:1; apex acuminate; base symmetrical, rounded to broadly cordate, often very slightly subpeltate; lateral veins 4—5 pairs, one pair arising ca. 1/5 up the midrib, the rest from the base, often another very short pair from the midrib near the apex. Petioles 12-15 cm. long, sheath- ing at the lower third. Inflorescences terminal, usually more than two at the end of side branches, the bases enclosed by bracts ca. 1.5 cm, long. Male inflorescences slender, to 14 cm. long; peduncles to 3 cm. long. Fe- male inflorescences thicker, to 10 cm. long; peduncles ca. 2 cm. long. Male flowers 2-staminate; anthers oblong, 2-valved; filaments longer than anthers; bracts orbicular, shortly peltate. Female flowers sessile; stigmas 2-fid, rarely 3-fid, filiform, borne on short somewhat thick styles. Fruits ca. 2 mm. long, 1 mm. broad, oblong, partially concrescent at maturity. DIsTRIBUTION. New Guinea and Bismarck Archipelago. West New Guinea. NoRTHERN. Idenburg River: Brass 14042, April (prt). SOUTHERN. Noord-rivier; near Bivak-eiland: Versteeg 1136, March (lectotype of P. versteegii, BO, L); Sabang: Versteeg 1768, Sept. (Bo). N.E. inea. Morose Dist. Gurokar: Brass 29399, May (CANB, LAE); Rawlinson: van Royen NGF 16246, March (srl, LAE, NSW); Wagau: Millar NGF 23453, June (LAE). Papua. GuLF Div. Kira: Brass 1115, March (isotype of P. rhizocaule, prt). NorTHERN Div. Kokoda: Carr 16313, March (cANB). WESTERN Div. Fly River: Brass 7238, July (prt). Bismarck Archipelago. NEw Britain. Kandrian: Sayers NGF 21973, March (LAE, NSW). I find no justification for maintaining in specific status Trelease’s Piper rhizocaule. Having compared the type of P. rhizocaule with that of P. versteegi, I conclude that the two are entirely conspecific. Piper versteegii bears a strong superficial resemblance to P. mestonit, particularly in the leaves; but on closer examination it is found to differ from the latter radically by the terminal inflorescences, more than two at each node and bracteate at the bases, by the broadly sheathing bases of petioles, the 2-fid stigmas, and the fruits only partially concrescent at ma- rit Because the inflorescences remain terminal at maturity, this species per- haps forms the link between section PorHomorPHE which is characterized by axillary inflorescences and section Piper in which the inflorescences are leaf-opposed. 15. Piper wichmannii C.DC. Nova Guinea Bot. 8(2): 418. 1910. P. erectum C.DC. Denkschr. K. Akad. Wiss. M.-N. Kl. Wien 89: 528. tab. 4, fig. 5b. 1914, syn. nov. P. schlechteri C.DC. Bot. Jahrb. 55: 204, 1918, syn. nov, P. arbuscula Trelease, Jour. Arnold Arb. 9: 146. 1928, syn. nov, 1972] CHEW, THE GENUS PIPER, 1 21 TypE. West New Guinea. Cote du Nord: Atasrip 39, Exped. Wichmann (BO, L). FIGURE 6. Dioecious shrubs. Lamina moderately petiolate; asymmetrically broad ovate; ca. 26 cm. long, 18 cm. broad, length:breadth ratio ca. 3:2; the upper side glabrous, the lower minutely hirtellous on the veins only; apex shortly acuminate; base asymmetrically deeply cordate; lateral veins 4 or 5 pairs, the distal pair arising from the midrib at ca. 1.5 cm. from the Ficure 6. Piper wichmannii. A, 2? twig; B, ovary; C, stamen. All from type. 22 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 base, the rest directly from the base, the broader side often with 1 or 2 more veins; margin ciliate. Petiole as long as the sinus of the base of the lamina is deep. Stipules ca. 5 cm. long, longer than petioles. Inflorescences up to as long as leaves, the females invariably shorter; peduncular stalks shorter than petioles. Male flowers 2-staminate; stamens 0.5 mm. long; anthers reniform, dehiscing apically; filaments short, broad, and stout. Female flowers sessile; stigmas 3-fid, subsessile; bracts round, peltate, long pedicellate. Fruits sessile, somewhat obconical, free at maturity. DIstriBuTION. New Guinea, Bismarck Archipelago, and Solomon Is- lands. West New Guinea. WANDAMEN. Wondiwoi Mts.: Schram BW 10742, Feb. (LAE). VoGELKop. Isjon River: van Royen & Sleumer 7558, Oct. (CANB, LAE). Sine loc.: cote du Nord: Atasrip 39, Exped. Wichmann (isotype of P. wichmannii, BO). N.E. New Guinea. Moroze Dist. Mt. Rawlinson: Hoogland 9099, June (BRI, LAE). SEPIK Dist. Djamu: Schlechter 17584, April (isotype of P. schlechteri, SING Papua. CENTRAL Div. U-uma River: Brass 1449, May (isotype of P. arbuscula, BRI). Bismarck Archipelago. NEw BRITAIN: Floyd 3525 , Aug. (BO, BRI, CANB, LAE, NSW, SING); Sayers NGF 24209, March (CANB, LAE, NSW, SING Solomon Islands. BouGAINVILLE. Kieta: Kajewski 1538, March (Bo, BRI, SING); Kupei: Kajewski 1725, April (Bri). GUADALCANAL. Tutuve Mts.: Kajewski 2601, April (Bo, Brt). SAN CristopaL. Waimamura: Brass 2589, Aug. (BO, BRI, SING). SanTA YsaseEL. Tiratona: Brass 3319, Dec. (srt). This is perhaps the commonest species of Piper in New Guinea and the Solomon Islands; and has been described as comprising handsome soft- wooded trees often with prop-roots. Its arborescent habit of growth coupled with the characteristically large cordate leaves with long spikes makes it the most distinctive species in the genus. The three species listed in the synonymy above have been found to be entirely conspecific with Piper wichmannii. I have examined all the type materials (excepting P. erectum, of which I have seen an illustration) and I am confident that they are properly placed here. The closest relative of this species is definitely Piper grandispicum, also of New Guinea, from which our species differs in having shorter in- florescences. INDEX TO EXSICCATAE The numbers in parentheses refer to the corresponding species in the text. Argent, NGBF 1007 (10) Beguin 1574 (3) Atasrip 35 (3); 39 (15); 64 (4); 73 Bloembergen 4615, 4702 (6) Boorsma 2 (3); 3 (9); 4 (4); 9 (6); J 1 Beer, BSIP nos. 6390 (4); 6498 (12); Boraule, BSIP nos. 9305 (15); 9478 7259 (9) | (4) 1972] Borgman 259 (7); 383 (13) Bowers 51, 53 (7); 86 (8); 254, 402, 403, 426 (7); 809 (8 Branderhorst 330, 425 (4) Brass 683 (8); 701 (1); 1115 (14); 1427 (10); 1449 (15); 1562 (8): 1644 (4); 2577 (12); 2581 (8); 2583 (14); 7258 (10); 8128 (8); 8276 (10); 11671, 12252 (7); 13243 (15); 13710, 13976 (8); 14042 (14); 19327 (4); 23235 (7); 23576, 23684 27128 (11); 27255, 27285 (8); 28864 (1); 29175 (4); 29201 (1); 29399 (14); 29542 (11); 30814 (13); 30918 (7); 31596, 32089 (8); 32091 C03 82080) aie Linas Buderus, NGF 207 Burn- Murdoch, ies nos. 6964 (15); 7122 (12); 7462 (15) Carr 11454 (8); 11636 (4); 11896, 12436 (8); 12725 (4); 13011 (13); 13199 (15); 13321, 13793, 13815, 13924 (13); 15447, 15793 (8); 16313 1 (14) Clarke, ANU nos. 9 (15); 9509 (7); 9538 (15) eeeioe ae (13) Cole 5 Coode : eect NGF 40353 (7) Corner, RSS nos. 187 (12); 193 (6); 197 (8); 3038 (12) Cowmeadow, BSIP nos. 3281 (15 3283, 3782 (12); 4755 (15); 4762 (4); 4830 (8): 4836 (15); 4838 “we Craven & Schodde 22 (12); 145 (1); 201, 246 (8); 512 (15); 513 (12); 1207 (7) Darbyshire & Hoogland 8124 (5); 8254 (7); 8344 (10) Dennis, BSP nos. 8148 (12); 8532 (15) CHEW, THE GENUS PIPER, 1 23 Docters van Leeuwen 9745 (15) Fa’arodo, BSIP nos. 12116, 12170 1 Flecker 12449 (10) Flenley, ANU nos. 2040 (7); 2314 Floyd, NGF nos. 3525 (15); 5510 (10); 6601 (15 Floyd & Womersley, NGF 6300 (13) Floyd et al., NGF 5286 (4) Foreman, NGF 45627 (12) Foreman et al., NGF nos. 45931 (8); 48004 (4) Frodin, NGF nos. 26434 (11); 26954 (4); 28407 (7); 28509 (13); 32052 (7); .getfe 18 Gafui, BSP nos. 7511 (6); 7512 (15); 12971 (15); 1297 Gjellerup a Boe 599 (5) Gray, NGF Gray & heed. vaacs 8050 (9) Hainsworth 49 (15); 61, 158 (8) Hartley 9620 (11); 9639 (2); 9765, 9839 (8); 10036 (4); 10189 (11); 10355 (15); 10438 (1); 10938 (15); 11345 (8); 11796 (7); 12187 (11); 12376 Hellwig 624 (8) Henty, NGF nos. 10626 (7); 11554 (15); 11912 (11); 11986 (4); 13676 (11); 16925 (4); 16960 (1); 20898 (8); 20922 (15); 29361 (8) Henty & Streimann, NGF 38900 (8) Henty et al., NGF nos. 33063 (15); 38972 (7); 38991 (15) Heyligers 1420 (4) Hollrung 260 (10) Hoogland 3973 (1); 3976 (15); 4165 (4); 8847 (2); 8895 (11); 9099 Hoogland & Craven 10111 (14); - 10114 (4); 10299 (8); 10368 (5); 24 JOURNAL OF THE ARNOLD ARBORETUM 10409, 10410 (8); 10575 (10) Hoogland & MacDonald 3513 (14) Hoogland & Pullen steed ae 5254 (8); 5301 (7); 6163, 6165 (8) Hoogland & Schodde aa oS 6914, 7283 (8) Hooley 27 Hornibrook 93 (7); 118 (15) Hunt, RSS nos. 2124, 2392 (12); 2393 (4) 10507 (4); (15); 2408 (12); 2557 ( Jackson 68, 69 (15); 125 (8) Jones 3024 (4) Kairo & Streimann, NGF nos. 26030 (8); 26031 (4); 35722 (8) Kajewski J538 (15): i985 (12); T5303 (G6): T7a3 -C15)< i782 (4); :2047 (1); 2133 (8); 2175 (1); 2185 (12); 2579 (8); 2598 (12); . 2601 (15); 2646 Kalkman, BW nos. ye 3644 (10); 3782 (4): 4044 (7); 5200, 5252 (8) Kanehira Pe Hatusima 11531 (9); 12643 er . 1175 (10); 4269 (9); 4409, 10775 (4 Kotali, BS7P nos. 11336 (15); 11341 12 Kuswata & Soepadmo 188 (3) Lam 2460, 3630 (3) Lauterbach 1337 (4); 1489 (8) Lavarack & Ridsdale, NGF nos. 31239, 31366 (12); 31475 (1) Lipageto, BSP nos. 3352 (4); 3369 (15) Maenu’u, BSIP nos. 5175 (4); 5186 (12); 6070 (as); 6149 (12) Mann, NGF 431 Mann & Osborn, ‘NOR 43037 (6) Mann & Vandenberg, NGF nos. 43192 (7); 43442 (8) Mauriasi, BS/P nos. 7602 (15); 8106 (4); 8479 (12); 8754, 8836 (15); 9596 (12); 9771 (15); 9794 (8); 11389 (12); 11455 (15); 11456 (4); 11526 (8); 11635 (12); 11636, 11759 (15); 11771 (8); 13223 (4): [voL. 53 13285, 13456 (12); 13457 (8); 13559 (15); 13585 (8); 13641 (12); 13656 (8); 13668, 13878, 13943 (12); 13983 (8); 14116, 14242 (12); 14269, 14364 (15) McKee & Floyd, NGF 6389 (13) Millar, GH 99 (1); NGF nos. 9930, 12228 (1); 14575 (8); 22662 (4); 22764 (7); 23240 (8); 23391 (15); 23453, 35393 (14); 37647 (15); 37693 (6); 37695 (8); 37730 (15) Millar & Holttum, NGF 15824 (7) Millar & Nicholson, NGF 13847 (7) Millar & van Royen, NGF 156294, 15658 (10) Millar & Sayers, NGF, 23747 (8) Millar & Vandenberg, NGF nos. 35212 (8); 37555 (10); 40900 (6); 40944 10 Morrison, BSIP nos. 177 (15); 281 (4) Nakisi, BS7P nos. 7914 (9); 7919 (6); 7920 (4) Native coll., 5512 (8) NGF nos. 4719 (2); Panoff 105 s od (15) Pulle 172, Pullen 407 tay 411 (8); 454 (7 « 2818 (10); 5168, 5224 (8): 5961 (15) Pulsford & Floyd, NGF 5422 (10) Ramo & Ben, BSIP 8132 (4) Ridsdale & Lavarack, NGF 30627 (8) Robbins 122 (15); 468 (8); dots (4) Robinson PI. no h. Amb, 58 (3) Roemer 1268 Royen iso Gay. 5215 (5); 5278 (2): NGF nos. 16246 (14); 18260 (7); 20253 (4); 30092 Royen & Millar, NGF 15689 (13); 17527 (8) Royen & Sleumer 6159, 6191 (8); 7026 (4); 7052 (1); 7551 (8): 7558 (15) Runikera, BSIP nos. 10341 (15); 10369, 10464 (12); 10786 (8): 1972] 10901 (12); 12705 (8); 12731 (12); 12939 (8) Saanan 26 (3) Sayers, NGF 13205 (4); 19603 (1); 19746 (15); 21695 (8); 21973 (14): 24171 (4); 24209 (15): 24224 (8) Schiefenhovel 107 (1 Schlechter 14604, 16252, 16791 (11); 17010 (8); 17584 (15 Schodde 1393 (8); 1397 (13); 2242 (15); 2355 (7); 2433 (10); 2435 (8); 2927 (1) Schodde & Craven 3605 (4); 3771 (1); 3984 (12); 4084 (8); 4331 (9): 5032 (8); 5067 (14) Schram, BW nos. 9339, 10704 (8); 10742 (15); 12924 (5) (7 Sirute, BSP nos. 9668 (8); 9683, 9970, 10091 (12) Sleumer & Vink 4432 (8) Stauffer & Sayers 5577 (8) Straatman, BMF 207 (4) Street 62 (8 Street & Manner 269, 396 (15) Streimann, NGF nos. 35849, 44296 (8); 44396 (1); 44489 (8) Streimann & Kairo, NGF nos. 47552 (8); 27750 (7): apo (15); 35822 Wa 39034 (13); 39188, 39195 196 Susui, perp 8322 (12) Teona, BSIP nos. 4865 (12); 6301 8 Teysmann 1975, 5633, 11717 (3) Lorrie: & Galore, NGF 42063 (4); e acmarbe & Mann, NGF nos. 42195 (8); 42204A (15) 42263 (6); 42265 (4) Vandenberg et al., NGF nos. 39706 (8); 42047 (14): 42049 (15) Versteeg 1021, 1058 (1); 1086 (8); Royar Boranic GARDENS SypNEy, N.S.W. 2000 AUSTRALIA CHEW, THE GENUS PIPER, 1 25 ae (14); 1515 (7); 1533 (10); 8 (14) ade BW nos. 3120 (8); 3935 5 7367 ave 7572 (9); 12665 je 16341, ee 16802 (8); 16839 (1 Vink, BW nos. 11426 (4) 12052 (9); 12155, 15227 (8) Vink & Schram, BW nos. 8593 (7); 8678 (8); 8924 (7) Vorderman 30 (3 165644 (7); 5) Walker, AN s nos. 32, 419 (7) Weinland 256 (11) White, C. 4 255, 354 (8); 454, 518 (15) White, C.-T. et ‘al, (4); 1567 (10) White, K. J., NGF nos. 9521 (14); 10453 (15); ald (8) 10474 (1) Whitmore, BSJP nos. 905 (8); 1861 (15); 2032 (12); "iso (15); 2230 (4); 2257 (8); 2316 (15); 2917, 3110 (12); 3827 (8); 3867, 3917 (12); 3970 (4); 4190 (8); 4269, 5785 (4); 5838 (12) Whitmore, RSS nos. 6109 (8); 6144, 6222 (12 Whitmore & Womersley, BSIP 829 NGF nos. 1554 12 Womersley, NGF nos. 4321 (8); 4322 17848 (14); 19153 (8); 24752 (5); 43611 (13); 43632 (15) Womersley & Floyd, NGF 6934 (13) Womersley & Ingle, NGF 5378 (11) Womersley & Millar, NGF nos. 8483 (13); 8485 (1) Womersley & Simmonds, NGF 7134 , NGF nos. 11860 12711 (14); (8) Womersley & Thorne (7); 11891 (13); 12766 (8) Womersley & Woolliams, NGF nos. 37083, 37085 (8) Womersley et al., NGF nos. 6097 (7); 37258 (4) 6082, 26 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 THE GENERA OF JUGLANDACEAE IN THE SOUTHEASTERN UNITED STATES 1 THomMaAs S. ELIas JUGLANDACEAE A. Richard ex Kunth, Ann. Sci. Nat. 2: 343. 1824, “Juglandeae,” nom. cons. (WaLnut Famity) Trees [rarely shrubs] mainly of temperate regions, the branchlets terete, the w urable, often dark-colored. Leaves deciduous [or evergreen], alternate [or opposite], petiolate, imparipinnate, usually glandular-dotted beneath and aromatic, exstipulate. Flowers imperfect. Staminate flowers borne in pendulous [rarely erect] aments, 3- to several-fasciculate on branches of the previous year or at the base of branches of the year, or solitary and lateral [or borne in a terminal cluster of several pendulous or erect aments with a central carpellate ament]; each staminate flower ith a primary bract,? 2 secondary bracts [sometimes absent], and 4 or fewer perianth lobes, or the perianth sometimes absent (Carya, Platy- ‘Prepared for a generic flora of the southeastern United States, a project of the Arnold Arboretum and the Gray Herbarium of H rvard Universit through the support of the National Science Foundation (Grant GB-6459X, princi- pal investigator Carroll E. Wood, Jr.). Th in the first paper of the series (Jour. Arnold Arb. 39: 296-346. 1958). The area of the flora includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississippi, Arkansas, and Louisiana. The descriptions are based pri 0 brackets. References that I have not seen are marked by an asterisk I am grateful to Dr. Wood for his thorough editing, valuable suggestions in the Scri to the manuscript. Dr. Donald E. Stone has read the manuscript and h ered a number of ful and much- appreciated comments. Mrs, Nancy Dunkly h helped greatly in checking the bibli- terial of Carya kindly sent by Dr. Stone. *The primary bract of Alfaroa and Engelhardia (Engelhardtia) is 3-lobed and is interpreted by Hjelmqvist (1948) as representing the fusion of two lateral secondary bracts with the primary bract. 1972] ELIAS, GENERA OF JUGLANDACEAE 27 primary bract [3-lobed in Alfaroa and Engelhardia| and 2 secondary bracts (occasionally modified or partly or wholly absent) and of a 4- parted perianth; primary and secondary bracts adnate at least to the base of the ovary, one or more perianth lobes occasionally modified or perianth absent (Carya); gynoecium syncarpous, usually of 2, rarely 3 or 4 carpels, the ovary inferior, 1-locular above, 2- or 4[8]-locular near the base, the single ovule erect, basal but at the apex of the incomplete septum(a), orthotropous, with a single integument; stigmas 2 (sometimes deeply 4- parted), sessile on the ovary, generally fleshy and papillate. Fruit drupa- ceous, a nut with a leathery or fibrous, dehiscent or indehiscent husk * [or fruit a nut with a wing or wings formed by the persistent primary and/or secondary bracts]; pericarp generally indurated, intrusive at the base, incompletely 2—4[8]-locular, the wall and septum(a) with or with- out longitudinal lacunae (cf. Ficure 3, i, 1). Seed large, filling the cavity of the pericarp, 2—4[8]-lobed; endosperm absent; seed coat thin; em- bryo oily, occasionally sweet, the large cotyledons variously lobed, twisted, or folded, the short radicle superior. Base chromosome number 16. TYPE GENUS: Juglans L. A family of seven genera (eight if Oreomunnea Oersted is segregated from Engelhardia Leschen. ex Blume*), with approximately 60-65 species distributed primarily in temperate regions of the Northern Hemi- sphere but extending southward in montane areas of tropical America and Asia, The family is not known to be represented in either Africa or Australia, although Engelhardia reaches Java and New Guinea. Platycarya Sieb. & Zucc., Pterocarya Kunth, and Cyclocarya Iljinsk., are restricted to the Old World, Alfaroa Standl., is confined to the New, and Engelhardia, Carya Nutt., and Juglans L. are represented in both hemispheres (but see footnote 4). Carya and Juglans represent the family in the southeastern United States. The family Juglandaceae is considered by Lawrence (1951) to com- prise the monotypic order Juglandales, while Takhtajan (1969) includes that the pollen ultrastructure of Juglandaceae and Rhoipteleaceae is com- patible. Hutchinson (1967) and Cronquist (1968) include both families and the trifoliolate-leaved Picrodendraceae in the Juglandales. This order is closely allied with the Myricales, Betulales, and Fagales. The Juglan- daceae are distinguished from other Juglandales by the unisexual flowers borne in catkins; the drupaceous or nut-like fruit; the solitary basal, *The husk (incorrectly called “exocarp” by some authors) is derived from the involucre and the perianth, while the hard shell of the nut is derived from ovarian “Although Oreomunnea has traditionally been considered to be congeneric with Engelhardia, Stone (Ann. Missouri Bot. Gard., in press) has reéxamined the evidence an inks that the two should be recognized as distinct genera. Engelhardia as then circumscribed is restricted to the Old World, Oreomunnea to the New. 28 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 orthotropous ovule with a single integument borne in a 1-locular, inferior ovary; and the pinnately compound, exstipulate leaves. cheat Pollen studies by Whitehead (1965) have shown that the family in- cludes several types. On the basis of ten characters that presumably indi- cate evolutionary trends within the family (including increase in size, sphericity, presence of pseudocolpi, distinctness of the columellae, irregular thickening of the ektexine, increase in pore number, increase in hetero- polarity, and various ektexinous thickenings) Engelhardia Roxburghiana Wall. is considered to have the least specialized pollen. Pollen of Engel- hardia and of the somewhat more specialized Alfaroa is isopolar and triporate (occasionally tetraporate, rarely pentaporate). Pollen of Platy- carya is also basically triporate (bi- to pentaporate) but the presence of arcuate pseudocolpi indicates increased specialization. Distinctly larger and definitely heteropolar, the pollen grains of Carya are generally tri- porate. The specialized grains found in Juglans and Pterocarya are large, stephanoporate to basically periporate, and heteropolar. In Juglans the pores may vary from 2 to 37 per grain, with 6 to 18 the prevailing num- bers. According to Erdtman, the Juglandaceae have several pollen charac- ters in common with the Betulaceae, Casuarinaceae, Myricaceae, and Rhoipteleaceae. Studies of the secondary xylem have shown that of Alfaroa and Engel- hardia to be the least specialized (Heimsch & Wetmore, 1939). The diffusely distributed vessels of Alfaroa have thin walls and the vessel ele- ments have scalariform perforations. Although similar, the wood of En- gelhardia appears to be more specialized in its ring porosity. Juglans and Pterocarya are closely allied in several common anatomical features, al- though Juglans appears to be slightly more specialized. Anatomically Platycarya is somewhat isolated, having attained a high degree of spe- cialization in the inflorescence, while the wood is relatively unspecialized as in Alfaroa and Engelhardia. The greatest specialization in the secondary xylem, i.e., distinct ring porosity and rounded, thick-walled vessels, has been reached by Carya. Iljinskaya (1953) notes that Cyclocarya has ring- porous wood as in Carya and Platycarya. Since Alfaroa and Engelhardia exhibit less specialized wood structure than the Anacardiaceae, it seems improbable that the Juglandaceae were derived from the Anacardiaceae as suggested by Cronquist (1968). From a morphological study of the inflorescences and flowers of the Juglandaceae, Manning (1938, 1941, 1948) concluded that Carya and Juglans are the most specialized and Engelhardia and Alfaroa the least specialized genera of the family. Evolutionary trends leading to the Jug- landaceae apparently included the development of imperfect-flowered axillary catkins from a terminal, highly branched panicle of perfect flow- ers. In general, the reduction of floral parts in the Juglandaceae supports the conclusions drawn from pollen and anatomical studies. Various authors have grouped the genera of Juglandaceae into two sub- families on the basis of the fruit type (winged, nutlike fruits vs. unwinged, drupaceous ones), but this distinction appears to be an artificial one. 1972] ELIAS, GENERA OF JUGLANDACEAE 29 In addition, there also appears to be no validly published name available for the subfamily that does not include Juglans when such a division is made. The names Nuciferoideae and Drupoideae proposed by Koidzumi (Acta Phytotax. Geobot. 6: 10. 1937) are not based on those of genera; Oreomunnoideae Leroy (Mém. Mus. Hist. Nat. Paris Bot. II. 6: 86. 1955) was published without a Latin description and indication of the type; and the properly formed name Pterocaryoideae used by Melchior (A. Engler’s Syllab. Pflanzenfam. ed. 12. 2: 42. 1964) in place of Nuci- feroideae is also invalid since he neither gave a full and direct reference to Koidzumi’s description (as required by ICBN, Art. 33) nor supplied a Latin description and a type. The extant genera and species of Juglandaceae appear to be the remain- ing stocks of a once large and much more widespread family. If identifi- cations of fossils are correct, the earliest records of Juglans are from the Middle to Upper Cretaceous. Carya has not been attributed to the Cretaceous, although it was abundant in the Tertiary and present in later times. Fossils of Engelhardia and Platycarya are also known, and several extinct genera have been described The Juglandaceae are of considerable economic importance for their excellent wood and nuts (Juglans, Carya, q.v.). Several species of Ptero- carya and the handsome shrubby tree Platycarya strobilacea Sieb. & Zucc. are cultivated as ornamentals in the United States. REFERENCES: BAILLON, H. Juglandacées. Hist. Pl. 11: 401-407. 1892. ; Beck, G. F. Status of the Tertiary woods of ae western states representing ae Juglandaceae. Northwest Sci. 18: 10. 4.* BENTHAM, G., & J. D. Hooker. Juglandeae. Hay Pl. 3: 397-400. 1880. Berry, E. W. Tree ancestors. vi + 270 pp. Baltimore. 1923. [The Walnut and Hickories, 66-80. ] Seakcenere? H, Z., V. Grir, T. Matveyeva, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. Ed. by A. Feporov. (Russian and English pref- aces.) 926 pp. Leningrad. 1969. [Juglandaceae, 358. CaANDOLLE, C. pe. Mémoire sur la famille des Juglandacées. Ann. Sci. Nat. Bot. V. 18: 5-48. pls. 1-6. 1862. , Juglandaceae. DC. Prodr. 16(2): 134-146. 1864. Conpg, L. F., & D. E. Stone. Seedling morphology in = Juglandaceae, the cotyledonary node. Jour. Arnold Arb. 51: 463-477. 197 Cook, O. F. Evolution of compound leaves in walnuts a hickories. Jour. Hered. 14: 77-88. 1923. Croizat, L. Observations on the ovary of the Juglandaceae. Southwest. Nat. 11: 72-117. 1966. Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. 1966. [Juglandaceae, 148, 149. ENGLER, A. ie Se Nat. Pflanzenfam. III. 1: 19-25. 1894. FoweLLs, H. A. Silvics of forest trees of the United States. U.S. Dep. Agr. Forest Serv. Agr. Handb, 271. vi + 762 pp. 1965. [Carya, 110-138; Juglans, 202-210; maps by E. L. Lirt1e, Jr. ] 30 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 53 Gray, J. Temperate pollen genera in the Eocene meta flora, Alabama. Science 132: 808-810. 1960. [Carya, Juglans, 8 Hans, A. S. Chromosome numbers in the saci Jour. Arnold Arb. 51: 534-539. 1970. HarpDIN, J. W. The Juglandaceae and Corylaceae of Tennessee. Castanea 17: 78-89, 1952. HEIMScH, C. H., Jr. Comparative anatomy of the secondary xylem in the “Gruinales”’ and “Terebinthales,” of Mitre with reference to taxonomic grouping. Lilloa 8: 83-198. 17 pls. 1 ———. Alfaroa pollen and generic ations in the Juglandaceae. (Abstr.) Am. er Bot. 31(Suppl.): 3s. 1944. R. H. Wetmore. The significance of wood anatomy in the taxonomy of the Juglandaceae. Am. Jour. Bot. 26: 651-660. 1939. Hemenway, A. F. Studies on the phloem of the dicotyledons. I. Phloem of the Juglandaceae. Bot. Gaz. 51: 131-135. pl. 13. 1911. HJELMoQvist, H. Studies on the floral morphology and phylogeny of the Amen- tiferae. Bot. Not. Suppl. 2: 1-171. 1948. [Juglandaceae, 30-63. HutTcuHINson, J. The genera of flowering plants. Vol. 2. xi + 659 pp. Oxford Univ. Press, London. 1967. [Juglandaceae, 136-141. IxuseE, M. On the pollen grains of some genera of Juglandaceae. (In Japanese, English summary.) Jour. Jap. Bot. 29: 333-335. 1954. [Platycarya, Pterocarya, & Juglans. | ILyinskayA, I. A. Monografiya roda Pterocarya Kunth. Acta Inst. Bot. Acad. Sci. URSS 1. Syst. 10: 7-123. 1953. [Includes Cyclocarya Iljinsk. gen. nov. English translation (Monograph of the genus Pterocarya Kunth) is available from U.S. Dep. ne Israel Program for Scientific Transla- tions, IPST Catalog No. 5622. 1970.] Jacoss, M. Juglandaceae. Jn: C. G. G. J. vAN STEENIS, Fl. Males. I. 6: 143- 154. 1960. [General account; Engelhardia. Jaynes, R. A., ed. Handbook of North American Nut Trees. vii + 421 pp. The Northern Nut Growers Association, Knoxville, Tenn. 1969. [ Juglanda- ceae, 163-223.] KARSTEN, G. Ueber die pve vida, der weiblichen Bliithen bei einigen Juglan- daceen, Flora 90: 316-333. 32. 1902. KorzumI, G. On the classification of Juglandaceae. (In Japanese and Latin.) Acta Phytotax. Geobot. 6: 1-17. 1937. Krigs, D. A. Comparative anatomy of the woods of the Juglandaceae. Trop. Woods 12: 16-21. 1927. Kuprianova, L. A. The palynology of the Amentiferae. (In Russian.) 214 pp. 48 pls. Bot. Inst. Komarovy Akad. Nauk SSSR. Moskva-Leningrad. 1965. [ Juglandales, 94-118 Kurz, H., & R. K. GopFrEy. Trees of northern Florida. xxxiv + 311 pp. Gaines- ville, Fla. 1962. [Juglandaceae, 36—50.] Lanopon, L. M. Ontogenetic and anatomical studies of the flower and fruit of the Fagaceae and Juglandaceae. Bot. Gaz. 101: 301-327, 1939 Leroy, J. F. Contre la théorie généralisée des carpelles-sporophylles. I. Une structure singuliére d’axe invaginé et de placentation caulinaire chez Annamocarya A. Chev. (Juglandaceae). Compt. Rend. Acad. Sci. 232: 432-434. 1951. [Amnamocarya sinensis (Dode) J. F. Leroy.]; II. Phylogénie structurale de la placentation dans le groupe Juglans—Carya (Juglandaceae). Jbid. 1007-1009; III. Discussion et conclusions. /bid. 233: 1214-1216. 1951. 1972] ELIAS, GENERA OF JUGLANDACEAE 31 . Etude sur les Juglandaceae. A la recherche d’une conception morpholo- gique de la fleur femelle et du fruit. Mém. Mus. Hist. Nat. Paris Bot. II. 6: 1-246. 31 pls. 1955. Lirtte, E. L., Jr. Checklist of native and naturalized trees of the United States (including Alaska). U.S. Dep. Agr. Forest Serv. Agr. Handb. 41. 472 pp. 1953. [Carya, 80-92; Juglans, 213-215. ] . Atlas of United States Trees. Vol. 1. Conifers and important hard- woods. U.S. Dep. Agr. For. Serv. Misc. Publ. 1146. vi + 9 pp. + maps 1- 200 + [2 pp.]. 1971. [Carya, maps 111-118; Juglans, maps 133, 134.] Lussock, J. A contribution to our knowledge of seedlings. Vol. 2. ii + 646 pp. London & New York. 1892. [Juglandaceae, 506-522. | MANNING, W. E. The morphology of the flowers of Juglandaceae. I. The in- florescence. Am. Jour. Bot. 25: 407-419. 1938; II. The pistillate flowers and fruits. /bid. 27: 839-852. 1941; III. The staminate flowers. /bid. 35: 606-621. 1948. . Alfaroa and Engelhardtia in the New World. Bull. Torrey Bot. Club 86: 190-198. 1959. . Additional notes on Juglans and Carya in Mexico and Central America. Ibid. 89: 110-113. 1962. MEtcuior, H. Juglandaceae. Jn: A. Engler’s Syllabus der Pflanzenfamilien. ed. 12. 2: 41-43. 1964. Nace, K. Studien iiber die Familie der Juglandaceen. Bot. Jahrb. 50: 459-530. pl. 4. 1914 . Kartographische Darstellung der Verbreitung der Juglandaceen. Ibid. S3k, pls. 5, 6. . Juglandaceae. Jn: W. Joncmans, ed., Fossil. Catal. II. 6: 1-87. 1915. Nakat, T. Juglandaceae Lindley. Fl. Sylv. Koreana. 20: 67-88. pls. 16-18. 1933. Nicotorr, M. T. Sur le type floral et le développement du fruit des Juglandées. Jour. Bot. Morot 18: 134-152; 380-385. 1904; 19 bis: 63-84. pls. 1, 2. 1905. PARMENTIER, P. Recherches anatomiques et taxonomiques sur les Juglandacées. Revue Gén. Bot. 23: 341-364. pls. 8-11. 1911. P’er, C. The Juglandaceae of Eastern China. Bot. Bull. Acad. Sinica 1: 203- ] StanpLey, P. C. The American species of Engelhardtia. Trop. Woods 12: 12- 15. 3927, Stone, D. E. Evolution of cotyledonary and nodal vasculature in the Juglan- daceae. Am. Jour. Bot. 57: 1219-1225. 1970. . New World Juglandaceae, III. A new perspective of the tropical mem- bers with winged fruits. Ann. Missouri Bot. Gard. (In press.) & C. R. Broome. Pollen ultrastructure: evidence for relationship of the Juglandaceae and the Rhoipteleaceae. Pollen Spores. (In press.) , J. Retcu, & S. WHITFIELD. Fine structure of the walls of Juglans and Carya pollen. Ibid. 6: 379-392. 1964. ; : TRELEASE, W. Juglandaceae of the United States. Missouri Bot. Gard. Rep. 7: 25-46. 25 pls. 1896. U. S. DEPARTMENT OF AGRICULTURE, ForEST SERVICE. Woody-plant seed man- 32 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ual. U.S. Dep. Agr. Forest Serv. Misc. Publ. 654. vi + 416 pp. 1948. [Carya, 109-111; Juglans, 201-204.] VERHOOG, H. A contribution towards the developmental gynoecium morphology of Engelhardtia spicata Lechen. ex Blume (Juglandaceae). Acta Bot. Neerl. 17: 137-150. 1968. Vines, R. A. Trees, shrubs and woody vines of the Southwest. xii + 1104 pp. Austin. 1960. [Juglandaceae, 123-137.] WHITEHEAD, D. R. Pollen morphology in the Juglandaceae, I: Pollen size and pore number variation. Jour. Arnold Arb. 44: 101-110. 1963; II. Survey of the family. Jbid. 46: 369-410. 1965. WITHNER, C. L. Stem anatomy and phylogeny of the Rhoipteleaceae. Am. Jour. Bot. 28: 872-878. 1941. Wore, J. A. Tertiary Juglandaceae of western North America. M. A. Thesis. Univ. Calif., Berkeley. 1957.* WoopwortH, R. H. Meiosis of microsporogenesis in the Juglandaceae. Am. Jour. Bot. 17: 863-869. pls. 50, 51. 1930. KEY TO THE GENERA OF JUGLANDACEAE IN THE SOUTHEASTERN UNITED STATES Branchlets with continuous pith; leaflets with involute vernation; staminate aments usually 3-fascicled; stamens to 8 or 10 per flower (usually 4); peri- se eitisyn cal ae, « Kile = Gays Sas Gee ey em ye 1. Carya. Branchlets with chambered pith; leaflets with conduplicate vernation; staminate aments solitary or in pairs, stamens to 40 per flower; perianth 3—6-lobed; car- pellate flowers with a 4-lobed perianth; fruit with an indehiscent husk, the nut. irregularly furrowed, (5 ccc blot a 2. Juglans. 1. Carya Nuttall, Gen. N. Am. Pl. 2: 220. 1818, nom. cons. Trees of varied woodland habitats, with smooth or exfoliating, aromatic, esinous, gray bark; heartwood hard, brown, sapwood pale; branchlets terete, tough, flexible, with continuous pith, the bud scales few and val- vate or numerous and imbricate, the axillary buds® smaller than the terminal bud. Leaves few- to many-foliolate, often glandular dotted be- neath, petiolate, leaf scar large, elevated, oblong or semiorbicular, often 3-lobed, emarginate; leaflets involute in vernation, usually ovate to obo- vate, generally acuminate at apex, mostly oblique at base, serrate [entire], *Paired valvate prophylls, probably protective in function, completely inclose (except for a terminal pore) the inner portions of the axi ary buds. : Ficure 1. Carya section Apocarya. a—h, C. cordiformis: a, staminate flower, 12; b, stamen, 15; c, three-lobed primary bract subtending stamens, 10; d, branchlet with two carpellate flowers, X 1 1/2; e, carpellate flower and involucre, X 8; f, mature fruit with involucre dehiscing — note winged margins ) . : valve removed, X 1; k, seed, X 1; 1, terminal branchlet, winter condition, X 1 m, n, C. myristiciformis: m, mature fruit with involucre dehiscing, X 1; n, ma- ture nut, X 1. 0, C. aquatica: mature fruit with one involucral valve removed. often subcoriaceous, sessile to short-petiolulate, becoming bright yellow in the fall. Flowers vernal, usually appearing after leaves unfold. Stami- nate flowers borne in slender, pendulous aments, these occurring singly or 5-8 in fascicles in the axils of the leaves of the previous year (occasional- ly of the current year) or at the base of the inner scales of the terminal bud; each flower subtended by a persistent, lanceolate bract; perianth absent, the floral envelope usually 2 (3)-lobed, consisting of one primary plus two secondary bracts, the primary one linear to lanceolate, free to 34 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 near the base and usually much longer than secondary bracts, secondary bracts ovate, rounded to acuminate at the apex, fused with the primary bract near the base; stamens 3—10(-15), 2- or 3-seriate, inserted on the slightly thickened receptacle-like inner and lower face of the floral en- velope; the anthers ovate-oblong, emarginate or divided at the apex, yel- low or red, generally pilose or hirsute, as long as or longer than their slender connectives, the filaments abbreviated, free; gynoecium absent. Carpellate flowers borne in 2—10-flowered spikes terminal on branches of the year; perianth absent, each flower surrounded by a 4-lobed slightly 4-ridged involucre that is villose and covered on the outer surface with yellow glandular scales that + persist in fruit; involucre often unequally lobed, composed of a primary bract and a 3-lobed structure probably rep- resenting 3 secondary bracts. Gynoecium bicarpellate, 4-locular at base, 2- and then 1-locular above at level of ovule (cf. Ficurr 2, f-h); stigmas spreading, papillate, + united at the base, often persistent. Fruit often ovoid, globose or pyriform, the involucre of varying thickness, becoming indurated at maturity, 4-valved, promptly or tardily dehiscent to the mid- die or to the base; nut usually oblong, obovate or subglobose, narrowed and usually rounded at base, with 2 abaxial sutures, each valve enclosing half of each cotyledon, usually smooth externally, variously ridged inter- nally, 4-locular at the base, above becoming 2-locular (by an incomplete middle septum separating the basal parts of cotyledons) and then 1-locular. Seed usually filling the nut cavity; testa thin, membranaceous; embryo with large fleshy cotyledons, the cotyledons oily, sweet to bitter, 2-lobed to near midway, the lobes oblong, compressed, with various longitudinal grooves, + concave on the inner surface; radicle short, superior. Base chromosome number 16. (Hicorius Raf., 1817, nom. rejic.; Hicoria Raf., 1838, orth. mut.; Pecania Dochnahl.) Lectotype species: C. tomentosa (Poiret) Nutt. (Juglans tomentosa Poiret), typ. cons. (Ancient Greek name, karya, applied to Juglans regia.) — Hickory. A genus of 15 species mainly of temperate eastern North America and southeastern Asia, represented in the eastern United States and parts of adjoining Mexico by 11 species, absent from western North America, although present there during the Tertiary. Four species occur disjunctly in southeastern Asia. The determination of the species of Carya should be approached with caution, for in addition to occasional hybrids, considerable intraspecific variation has lead to the description of numerous varieties, most of which are without merit. Although fruiting collections are often necessary for accurate identification, careful floral analysis should yield valuable data for distinguishing the species and understanding their biology. Section Apocarya C. DC. (Hicorius subg. Pacania Raf., subg. Drimo- carya Raf.; Hicorius § Apohicoria Dippel; Carya § Pacania (Raf.) Rehder) contains eight species disjunctly distributed between the eastern United States, Mexico, and Asia, The species comprising this section have 5-17 serrate, usually falcate leaflets, 4-6 valvate bud scales (buds pos- 1972] ELIAS, GENERA OF JUGLANDACEAE 35 sibly naked in C. cathayensis) that are hardly or not at all accrescent in spring, 3-8 (usually 4) stamens, a husk with the valves not keeled but frequently narrowly winged along the margins, and lacunose wall parti- tions in the nut. At least all of the American species of the section are diploids, 2x = 32, whereas species of section CarYA are predominantly tetraploid (Stone, 1963). Four species of sect. APocaRYA are found in the southeastern United States. An important component of the deciduous forest of the eastern United States, Carya cordiformis (Wangenh.) K. Koch, bitternut hickory, 2n = 32, occurs from New Hampshire, southernmost Quebec, New York, and southern Ontario, west to Minnesota and southeastern N ebraska, and south to northwestern Florida, eastern Oklahoma, and eastern Texas (see Little, 1971, map 112-E). This species is characterized by scurfy-yellow overwintering buds, lanceolate to lanceolate-ovate, scarcely falcate leaflets that are pubescent beneath, and cylindric to compressed fruit with the valves of the husk narrowly four-winged above the middle. It occurs most commonly in moist sites and reaches its greatest height (over 30 m.) in the rich bottomlands of the lower Ohio River basin. Hybrids with C. il- linoénsis (C. « Brownii Sarg.), C. glabra (C. X Demareei Palmer), and C. ovata (C. Laneyi Sarg.) have been reported. The bitternut hickory is allied to Carya Palmeri Manning, an endemic of Nuevo Leon, Tamaulipas, and San Luis Potosi, Mexico, that differs from C. cordiformis in having tight, smooth bark, in the 11 (9-13) occasionally falcate leaflets, in the more frequently hairy rachises and twigs, and in the glandular peltate scales on the undersurface of the leaflets, On the basis of similarities in leaf margins, stomata, trichomes, bracts protecting the axillary buds, and fruit, Stone (1962) demonstrated close relation- ships between C. cordiformis and C. Palmeri and the Asian C. tonkinensis and C. cathayensis. Stomatal guard-cell size indicates that C. Palmeri is a diploid (Stone, 1961). Abundant only in Arkansas on high bottom-flats of the Mississippi River Delta, Carya myristiciformis (Michx. f.) Nutt. (C. “myristicaeformis,” orth. error), nutmeg hickory, 2” = 32, is also found sparingly in scattered localities in eastern South Carolina, Alabama, and Mississippi, and in southeastern Oklahoma, Louisiana, and eastern Texas, as well as disjunctly in Nuevo Leon and T amaulipas, Mexico (see Little, 1971, map 116-E). Often a tall tree (to ca. 35 m.), C. myristiciformis has a nut with a thick hard shell; ovate-lanceolate to broadly obovate, more or less falcate leaf- lets; and long-pedunculate staminate catkins borne at the base of branch- lets of the year. This species is not known to form hybrids. Largely restricted to bottomlands and swamps, Carya aquatica (Michx. f.) Nutt. (Hicorius aquaticus (Michx. f.) Britton), water hickory or bitter hickory, 2n — 32, is found primarily on the Coastal Plain from southeast- ern Virginia and eastern North Carolina to southern Florida and eastern Texas, and in the Mississippi Embayment to southeastern Oklahoma, western Tennessee, southeastern Missouri and southern Illinois (see Little, 1971, map 111-E). The distinctive characters of this species include the 36 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 exfoliating bark, the reddish-brown, yellow-glandular overwintering buds, the lanceolate-falcate leaflets, and the laterally compressed nut with husk wings extending to the base. Carya aquatica is reported to hybridize with C. texana (C. X ludoviciana (Ashe) Little) and with C. illinoénsis (C. x Lecontei Little). In a study on the hybridity of C. «x Lecontei by an examination of the fatty acids in the oil of the embryo, Stone, Adrouny & Adrouny found that the reputed hybrid contained amounts of fatty acids intermediate with the two probable parents. The pecan, Carya illinoénsis® (Wangenh.) K. Koch (C. olivaeformis (Michx.) Nutt., C. Pecan (Marsh.) Engl. & Graebn., Hicorius Pecan (Marsh.) Britton), appears to be native to the Mississippi River Valley from southern Indiana, northern Illinois, and southeastern Iowa, to Ala- bama, Mississippi, Louisiana, and eastern Texas, but its original range is difficult to determine since it probably was spread by Indians, and it has been widely planted throughout most of the southeastern United States. It also occurs in scattered localities from Coahuila and Nuevo Leon, Mex- ico, southward to Jalisco, Hidalgo, and Oaxaca (see Little, 1971, maps 114-W, 114-E, 114-N). The combination of overwintering buds covered with clusters of yellow hairs; oblong-lanceolate leaflets, the lower pairs falcate; staminate catkins sessile or subsessile near the summit of shoots of the preceding year; and elongate fruits distinguishes the species. In addition to hybrids with C. cordiformis and C. aquatica, C. illinoénsis is reported to hybridize with C. laciniosa (C. < Nussbaumeri Sarg.) and with C. tomentosa (C. * Schneckii Sarg.), of sect. CARYA The remaining species of sect. APocARYA are Asiatic. Carya cathayensis Sarg. has been reported from Chekiang, Kweichow, and Kiangsi provinces, China. The closely related C. tonkinensis Lecomte has been collected in Yunnan, China; Tonkin, North Viet Nam; and Assam, India. Carya Poilanei (A. Chev.) Leroy is also known from Tonkin and from Laos. The only other Asiatic species of Carya, C. sinensis Dode, has been segregated from sect. ApocaryA to form the monotypic sect. RHAMPHO- cARYA Manning & Hjelmqvist (Annamocarya A. Chev.), mainly on the ba- sis of its 7-9 entire leaflets, either hollow or solid pith, few and unequal bud scales, staminate catkins in fascicles of five to eight, 5-15 stamens, 4—6 apically keeled valves of the husk, and lacunae absent in the lower part of the walls of the nut but sometimes present near the apex. The New World section Carya (Carya § Eucarya C. DC.; Hicorius subgen. Glycaria Raf., subgen. Hexacarya Raf.; Hicorius § Euhicorius Dippel, § Eucarya (C. DC.) Sarg.) contains seven species of the eastern United States and northeastern Mexico. Species of this section have 3-9 serrate, nonfalcate leaflets; 6-12 imbricate bud scales that are strongly accrescent in spring; 3-8, but usually 4, stamens; a 4-valved husk with the valves neither strongly keeled nor winged along the margin; and wall ® According to Thieret (1961) the original, and consequently correct, spelling of the epithet is illinoinensis. The Standing ohne gratis on Stabilization appointed by the XI International B otanical Congress (1969) has ruled, however, that illinoinensis is an orthographic error and the spelling illinoénsis should be retained. ze) Sera ate . poe Carya section Carya. carpe flower showing ovule, X 6; g, se midiagramm ot flower (at level of lower arrow in , Suen four-loculate condition - ao) ER SE es. " ste # FIGURE 2. a-—m, branchlet with staminate and _carpellate c fea ts24 eye ! 1 aa Ey C. inflorescences, Kol 3; 38 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 partitions of the nut without lacunae. All the species of this section that have been examined are tetraploid, 2n = 64, with the exception of the diploid C. laciniosa and C. ovata. Common in rich woods, bottomlands, and on slopes, Carya ovata (Mill.) K. Koch (Hicorius ovatus (Mill.) Britt.), 22 = 32, shagbark hickory, is found throughout most of the eastern United States from southeastern Nebraska and southeastern Minnesota to southern Maine, southward to Georgia, Alabama, Mississippi, Louisiana, and eastern Texas, and dis- junctly in Mexico (Tamaulipas and Nuevo Leén south to Hidalgo and Puebla). It is largely absent from the southeastern Coastal Plain (see Little, 1971, maps 118-N, 118-E). Little (1969) recognizes two varieties: var. ovata (including C. mexicana Engelm. ex Hemsl.) and var. australis (Ashe) Little (C. carolinae-septentrionalis (Ashe) Engl. & Graebner, C. australis Ashe), Carolina hickory or southern shagbark hickory. Several varieties described by Sargent do not appear to be worthy of recognition. The fruits usually longer than 3.5 cm., the dark brown or black terminal bud scales, and the generally lanceolate or oblanceolate terminal leaflets of var. australis serve to separate it from var. ovata with its smaller fruits (less than 3.5 cm. long), tan or light brown bud scales, and usually obovate terminal leaflets. Carya ovata is reported to hybridize with C. laciniosa (C. X Dunbarii Sarg.) and C. cordiformis (C. x Laneyi Sarg.). The other diploid member of sect. Carva, C. laciniosa (Michx. f.) Lou- don (Hicorius laciniosus (Michx. f.) Sarg.), shellbark hickory, 2n = 32, is common in rich woods and bottomlands, mainly in the Ohio and Missis- sippi valleys. It reaches the northern limit of its range in western New York and southern Ontario and extends westward to southern Iowa and Missouri, and southward to West Virginia, Kentucky, Tennessee, Arkan- sas, and northeastern Oklahoma, with local occurrences in Virginia, North Carolina, Georgia, Alabama, and Mississippi (see Little, 1971, map 115-E). It is closely related to C. ovata but is distinguished from it by orange- brown or light tan branchlets; by the larger bud scales, glabrous above and accrescent; by oblong-lanceolate to oblong-obovate acuminate leaflets that are soft-pubescent beneath and lack the prominent tufts of hairs lo- cated on the serrate margins of leaves of C. ovata; and by a large, angled, more or less strongly compressed nut with a yellowish- to reddish-brown shell. Hybrids of this species with C. ovata (C. & Dunbarii Sarg.) and with C. illinoénsis (C. X Nussbaumeri Sarg.) have been reported. Carya pallida (Ashe) Engl. & Graebn. (Hicorius pallidus Ashe), sand ov. lev level, X 8; i, mature fruit with involucre beginning to split, x fruit with two involucral lobes removed, < 3/4; k, nut, X 3/4; 1, median cross section of nut — cotyledons of embryo white, sclerified wall of nut hatched, x 3/4; m, terminal bud, winter condition, x 1, » C. lacini flower, X 10; 0, stamen, X 12; p, bracts subtending staminate fl carpellate flower, X 6; r, terminal bud, winter condition, X 1; ling, X 1/6. ary, X 8; h, semidiagrammatic median cross section of carpellate flower el ecg? S M4 ower, X 10; q, S, young seed- 1972] ELIAS, GENERA OF JUGLANDACEAE 39 hickory, is largely confined to the Coastal Plain from southern New Jer- sey to northwestern Florida and Louisiana, and north in the Mississippi Embayment to Tennessee and southeastern Kentucky. Deam (1940) re- ported it from Knox County, Indiana. It is readily distinguished by the combination of silvery-lepidote overwintering buds and staminate inflores- cences; long-attenuate resinous and aromatic leaflets that are silvery- lepidote beneath when expanding; and yellow-lepidote husk that tardily splits to the base (Fernald, 1950). The chromosome number of C. pallida has not been determined, but on the basis of stomatal guard-cell size it is presumed to be tetraploid (cf. Stone, 1961). No hybrids with other species have been reported. Common in most of the eastern United States, generally in dry woods and on slopes, Carya glabra (Mill.) Sweet (Hicorius glabrus (Mill.) Britt.; C. leiodermis Sarg., cf. Little, 1969), 2n = 64, is characterized by close, usually furrowed and rigid bark; acute overwintering terminal buds that are glabrous at first but become silky pubescent in autumn; glabrous petioles, rachis, and leaflets; and obovoid, dark brown, smooth, shining fruit with the husk indehiscent or opening tardily by only one or two sutures. Little (1969) recognized two varieties in this species in addition to var. glabra (C. microcarpa Nutt., H. microcarpus (Nutt.) Britt.), which occurs from Massachusetts and southern New Hampshire to New York, southern Ontario, southern Michigan, Illinois, and north- eastern Kansas, south to Arkansas, Mississippi, Georgia and northwestern Florida. Varietas megacarpa (Sarg.) Sarg. (C. Ashei (Sudw.) Kelsey & Dayton, H. austrinus Small, C. austrina (Small) Murrill), coast pignut hickory, with larger leaflets and fruits, is found from western New York to southern Ohio and Illinois, south to eastern Texas, Louisiana, and cen- tral Florida. Varietas odorata (Marsh.) Little (C. ovalis (Wangenh.) Sarg., C. ovalis var. odorata (Marsh.) Sarg.), red hickory, with a smaller globose or pyriform fruit, the husk of which splits to the base by four sutures, is of scattered occurrence throughout most of the range of var. glabra. (See Little, 1971, map 113-E, for distribution of the species.) A small tree or shrub confined to the scrub of dry sand-ridges from central to northwestern Florida, Carya floridana Sarg. (Hicorius floridanus (Sarg.) Small), Florida or scrub hickory, 2n = 64, is closely related to C. glabra, which in Florida occurs largely in hammocks, and to C. texana, which has a similar rusty pubescence of undulating peltate scales on the young leaves. Kurz & Godfrey (1962) question the taxonomic distinctness of C. floridana, since intergrades with C. glabra can be found in inter- mediate habitats. Carya floridana is separated from C. glabra largely by its habitat and by the characteristic rusty pubescence of the young leaves. Carya texana Buckl. (C. Buckleyi Durand, C. villosa (Sarg.) Schneider, C. arkansana Sarg.), black hickory, 2n = 64, occurs from eastern Kansas, Missouri, southern Illinois, and southern Indiana to eastern Oklahoma, central Texas, and southern Louisiana. Related to both C. ovata and C. floridana, C. texana typically has rusty-pubescent young branchlets and overwintering buds, bud scales hairy-tufted at the apex, leaflets rusty- 40 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 pubescent beneath with peltate scales, and yellow-scurfy fruit with the husk splitting either to the middle or nearly to the base. This species ap- pears to be restricted to dry upland woods or dry sandy or rocky slopes. A hybrid between C. texana and the bottomland diploid C. aquatica (C. > ludoviciana (Ashe) Little) has been reported, but according to Little (1953) additional study is needed to determine its true identity. Laughlin has described a hybrid between C. texana and C. tomentosa (C. X col- lina Laughlin). A woodland species, Carya tomentosa (Poiret) Nutt. (C. alba (Mill.) K. Koch non (L.) Nutt. ex Ell., nom. ambig.), mockernut hickory, 2” = 64, is found from New Hampshire, southernmost Ontario, southern Mich- igan, and southeastern Iowa, south to northern Florida and eastern Texas (see Little, 1971, map 117-E). Carya tomentosa is separated from the phenotypically similar C. laciniosa and C. ovata by its tomentose branch- lets, petioles, rachises, and lower leaf surfaces, with the trichomes usually being curly and fasciculate; by the outer bud scales deciduous in autumn, thus exposing the inner silky scales; by a husk that splits to or below the middle; and by the deeply furrowed but not exfoliating bark. In Illinois and Iowa it is known to hybridize with C. illinoénsis (C. *« Schneckti Sarg.). Most speculations about phylogenetic relationships within Carya have had to rely on morphological observations, but an analysis of the oils from the seeds of the different species has provided some additional insights into interspecific alliances. Stone, Adrouny & Adrouny (1969) found oils from members of sect. APocARYA to have high oleic and low linoleic content, while species of sect. CARYA have a low ratio of oleic to linoleic acid. Ties between the two sections are found in the average oleic and linoleic acid values of C. myristiciformis and C. aquatica, of sect. Apocarya, and C, ovata var. australis (C. carolinae-septentrionalis) and C. laciniosa, both diploids of sect. Carya. Oil data also link C. ovalis with C. aquatica, C. myristiciformis, C. laciniosa, and C. ovata, although C. ovalis has tra- ditionally been considered allied to C. glabra, and Little recently (1969) reduced C. ovalis to a variety (odorata) of C. glabra. The data support the close assemblage of C. floridana, C. texana, C. tomentosa, C. pallida, and C. glabra, and within this group of tetraploids, C. floridana and C. texana appear to be closely allied. A somewhat distant relationship was established between C. illinoénsis and C. cordiformis. Although it was im- possible to test the oils from any of the Asiatic species, it was suggested that ties with them might be via C. myristiciformis and C. aquatica. A perianth is considered by morphologists to be lacking in carpellate flowers of Carya, and the four structures surrounding the gynoecium are considered to represent involucral bracts. Manning (1941) has designated the lower or united basal part of the two stigmas as a “stigmatic disk,” which represents a modified calyx, but he has not published supporting anatomical evidence. Developmental and detailed anatomical studies should be helpful in determining the nature of the accessory structures of the carpellate flowers. 1972] ELIAS, GENERA OF JUGLANDACEAE 41 Flowering of Carya species occurs in the spring (usually in April and May) with the enormous amounts of pollen being dispersed by wind. In a study of commercial cultivars of C. illinoénsis, J. G. Woodroof (1930) found that 26 were proterandrous, 18 proterogynous, and 88 homogamous. Warm, sunny days were found to be optimal for pollen shedding, high humidity or rain retarding or completely stopping anther dehiscence. In C. illinoénsis the Polygonum-type embryo sac is at the four-celled stage at the time of pollination. Fertilization apparently does not occur for five to seven weeks following pollination, although Shuhart (1932) re- rusts only a two-week lag. Embryogeny in C. glabra is of the Asterad Pollination studies of Carya illinoénsis have shown it to be capable of both self- and cross-pollination. Following self-pollination, however, a greater number of immature nuts are dropped, resulting in smaller yields of harvested nuts than from the cross-pollinated trees. The staminate in- florescences are differentiated during the summer and mature the follow- ing spring, while carpellate inflorescences apparently do not differentiate until late winter or early spring. Throughout much of the south-central United States, C. illinoénsis, a prodigious pollen producer, is a major fac- tor in many spring pollen allergie es. The pollen grains of Carya are tectate, usually triporate, suboblate, rounded-triangular to circular in polar view, with finely scabrate sculptur- ing. They are distinctly heteropolar, the pores being slightly off the equator of the grain and the aperture + circular in outline (Whitehead, 1965). It appears from pollen morphological relationships that sect. CaryA is de- rived from sect. APocarYA. Stone (1963) pointed out that most species belonging to sect. Apocarya (all diploids) have pollen grains smaller than 46 yu, while species of sect. CARYA (mainly tetraploids) have pollen grains larger than 46 ». There is also a positive correlation between sto- matal size and ploidy level. Species of Carya were widespread (pantemperate?) during favorable Climates of the Tertiary. Fossils of Carya of Eocene and Miocene age are known from Alaska, the western and northwestern United States, Ver- mont, the southeastern United States, Greenland, Iceland, Spitzbergen, Central Europe, Russia (Kamchatka), and China (Shantung). There are apparently no good Cretaceous records Several species of Carya (C. ovata, C. laciniosa, C. tomentosa, and C. glabra) are the main sources of hickory lumber. The seeds of most species of Carya are high in tannins and consequently are not used as a food source, although they are an important food for wildlife. Nuts pro- duced by Cori ovata and C. laciniosa are reported to be sweet, edible, and of good quality. The pecan, obtained from C. illinoénsis, has become one of the world’s important nut crops. Interest in commercial pecan grow- ing, largely in the early 1900’s, has resulted in the description of numerous horticultural varieties. These cultivars, many of them originally selections rom native stands of C. illinoénsis but later from selective crossing, have formed the basis of the pecan industry of the southeastern and south- 42 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 central United States. Grown best on deep, loose, well-drained, well- aérated soil, C. illinoénsis can properly be considered a multipurpose spe- cies. The trees provide food for man as well as wildlife, they are suitable for landscaping, and the wood is widely used in the manufacture of tools, lumber, and veneers. The seeds are high in caloric value and have large amounts of unsaturated oils (ca. 72 per cent), protein (ca. 11 per cent), and carbohydrates (ca. 13 per cent). REFERENCES: Under family references see BAILLON (pp. 405, 406), Berry (pp. 66-72), Conve & STONE, FowELts (pp. 110-138), HuTcHInson (pp. 138, 139), JAYNES (pp. 163-202), Kurz & Goprrey (pp. 38-50), Litre (1953, pp. 80-92; 1971), SARGENT (1922, pp. 176-200), WHITEHEAD (1963, 1965), and Vines (pp. 127- 137). Aprtance, G. W. Factors influencing fruit setting in the pecan. Bot. Gaz. 91: 144-166. 1931. AUGANBAUGH, J. All about hickory. Am. Forests i 29, 30, 42, 44, 46. 1967.* [Chiefly culture & wood properties of Carya s Barton, L. V. Seedling production in Carya ee (Mill) K. Koch, Juglans cinerea L., and Juglans nigra L. Contr. Boyce Thompson Inst. 8: 1-5. 1936. Berry, E. W. Notes on the geological history of the walnuts and hickories. Pl. World 15: 225-240. 1912. Brison, F. W. Variations in pecans. Jour. Hered. 13: 366-368. 1922 [1923]. Brusu, W. D. Shellbark hickory, Carya laciniosa (Michx. f.) Loud. Am. For- ests 53: 364, 365. 1947. CoLtincwoop, G. H. Pignut hickory, Hicoria glabra (Miller) Sweet. Am. For- ests 43: 546, 547. 1937. . Shagbark hickory, Hicoria ovata (Miller) Britton. Jbid. 298, 299. 1937. Dayton, W. A. Fernow hickory (Hicoria Fernowiana Sudw.). Jour. Forestry 35; 859-864. 1937. [Known from a single cultivated tree “from the South” in Washington, D.C. ]| Duncan, W. H. Preliminary reports on the flora of Georgia — 4. Notes on the distribution of flowering plants including ees new to the state. Castanea 15: 145-159. 1950. [C. laciniosa, 151, FerRNALD, M. L. The a ae basis of the name Carya Pecan. Rhodora 49: 194-196. 1947. [C. illinoénsis.] Foster, A. S. Investigations on the morphology and comparative history of development of foliar organs II. Cataphyll and foliage leaf form and or- ganization in the black hickory (Carya Buckleyi var. arkansana). Am. Jour Bot. 18: 864-887. pls. 61-64. 1931; III. Cataphyll and foliage-leaf ontogeny in the black hickory (Carya Buckleyi var. arkansana). Ibid. 19: 75-99. pls. 2-4, 1932; IV. The prophyll of Carya Buckleyi var. arkansana. Ibid. 710- 728. pls. 50, 51. [C. texana.] —. parative histogenesis of foliar transition forms in Carya. Univ. Calif. Publ. Bot. 19: 159-185. 1935. ———. A histogenetic ced of foliar determination in Carya Buckleyi var. — Am. Jour. Bot. 22: 88-147. 1935. [C. texana Fox, W. B. Notes on na distribution of Carya aquatica in ‘North Carolina. cee Elisha Mitchell Sci. Soc. 64: 237-240. pls. 27, 28. 1948. 1972] ELIAS, GENERA OF JUGLANDACEAE 43 Hyetmovist, H. Notes on some names and combinations within the Amenti- ferae. Bot. Not. 113: 373-380. 1960. [C. integrifoliolata. | Hom, T. Ae tierer ot of Carya alba and Juglans nigra. Bot. Gaz. 72: 375-389, pis... 15, , Kriss, D,. A. The an ¥ jit tonkinensis H. Lecomte. Trop. Woods 16: 50-52. 1928 Lanopon, L. M. Development and vascular ema. of foliar organs of Carya cordiformis. Bot. Gaz. 91: 277-294. pls. 1, 2. 1931. . Embryogeny of Carya and Juglans, a comparative study. Jbid. 96: 93- 117. pl. 1. 1934, LaucHLIN, K. Carya X collina Laughlin, Scarit hickory. Phytologia 16: 343- 453: 1968. [C. texana & C. tomentosa.] Lecomte, H. Une Juglandacée du genre gmt en Indochine. Bull. Mus. Paris 27: 437-440. 1921. [C. tonkinensis Lecom Lirtte, E. L., Jr. Notes on the nomenclature oe Carya Nutt. Am. Midl. Nat. 29: 493-— 508. 1943. . Two varietal transfers in Carya (hickory). Phytologia 19: 186-190. 69. 1 McDanret, J. C. The pollination of Juglandaceae varieties — Illinois seca tions god review of earlier studies. No. Nut Growers Assoc. Rep. 47: 132. nti, Phe pollination of Juglandaceae varieties. II. Notes during 1957. /bid. 48: 89-93, 1957. . Pollination ” the Juglandaceae; 1958 observations on varieties as to flower types. Ibid. 49: 64-66. 1958.* McHarton, T. H. The history, distribution and naming of the pecan (Hicoria pecan). Proc. Southeast. Pecan Growers Assoc. 50: 10-12, 14-25, 27-29, 31-34. cba McKay, J. W. Embryology of pecan. Jour. Agr. Res. 74: 263-283. 1947.* . Interspecific hybridization in soage breeding. Proc. Tex. Pecan Grow- ers’ Assoc. 40: 66-68, 70-73, 76. 1961.* MANNING, W. E. A hybrid between shagbark and bitternut hickory in south- eastern Vermont. Rhodora 50: 60-62. 1948. [C. * Laneyi: C. cordiformis X C. ovata. ] . The status of Hicoria borealis Ashe. Ibid. 51: 85-89. 1949. [C. ovata var. borealis, comb. nov . The genus Carya in Mexico. Jour. Arnold Arb. 30: 425-432. 1949. ————.. A key to the hickories north of Virginia with notes on the two pignuts, Carya glabra and C. ovalis. Rhodora 52: 188-199. 1950. . Additional notes on Juglans and Carya in Mexico and Central America. Bull. Torrey Bot. Club 89: 110-113. 1962. : . Hickories reported in India and Laos with other notes on Carya in Asia. Brittonia 15: 123-125. 1963. H. HyeEtmovist. Annamocarya, Rhamphocarya and Carya sinensis. Bot. Not. 1951: 319-330. 1951. [Carya sect. Rhamphocarya, sect. nov., . Flori 1946. [13 species, including C. magnifloridana, sp. nov. REED, C. A. Hickory species and stock studies at the Plant Industry Station, Beltsville, Md. No. Nut Growers Assoc. Rep. 35: 88-121. 1945. ReHpber, A. Carya alba proposed as nomen ambiguum. Jour. Arnold Arb. 26: 482, 483. 1945, 44 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Risren, E. E. Pollinating the pecan. Jour. Hered. 8: 384. 1917. SARGENT, C. S. Notes on North American trees. II. Carya. Bot. Gaz. 66: 229- 258. 1918. [Includes Carya leiodermis, sp. nov., and many new varieties and combinations. Scott, R. A. Status of an Asiatic member of the Juglandaceae regarded as a ‘living fossil.” Am. Jour. Bot. 40: 666-669. 1953. [C. sinensis is not closely related to the Eocene Juglandicarya. | Sears, P. B. Postglacial migration of five forest genera. Am. Jour. Bot. 29: 684-691. 1942. [Includes Carya.] SHuHaRT, D. V. The formation and development . a pistillate flowers of the pecan. Okla. Agr. Exp. Sta. Bull. 163: 1-7. The morphological differentiation gh the a ah flowers of the pecan. Jour. Agr. Res. 34: 687-696. pls. 1-6. 1927. Pape ea and anatomy of the fruit of Hicoria Pecan. Bot. Gaz. 93: 1- 20. 193 Stone, D. E. oi level and stomatal size in the American hickories. Brit- tonia 13: 293-302. 1961. Affinities of a Mexican endemic, Carya Palmeri, with American and Asian hickories. Am. Jour. Bot. 49: 199-212. 1962. . Pollen size in hickories (Carya). Brittonia 15: 208-214. . New chromosome counts for two species of hickory oor ‘Ibid. 16: 230. 1964. [C. myristiciformis, n = 16; C. carolinae- -septentrionalis, 2n = S2.i , G. A, Aprouny, & S$. Aprouny. Morphological and chemical evidence on the hybrid nature of bitter pecan, Carya < ee Ibid. 17: 97-106. 1965. [Hybrid between C. illinoénsis and C. aquatica.] . H. Fraxe. New World ie pe in II. Hickory nut oils, phenetic similarities, and evolutionary implications in the genus Carya. Jour. Bot. 56: 928-935. 1969. Taytor, L. A. Plants used as curatives by certain southeastern tribes. xi + 88 pp. Botanical Museum of Harvard University. 1940, LC; i shoe 14.] THIERET, J. W. my: specific epithet of the pecan. Rhodora 63: 296. THOR, C. J. B., & C. L. Smirx. A physiological study of seasonal ae in the composition of the pecan during fruit development. Jour. Agr. Res. 50: 97-121. : TRUE, R. H. Notes on the fas history of the pecan in America. Rep. Smithson. Inst. 1917: 435-448. Witson, L. R., & R. oy hd Fossil evidence of wider post-Pleistocene range for butternut and hickory in Wisconsin. Rhodora 44: 409-414. 1942. Wooproor, J. G. The development of pecan buds and the quantitative produc- tion of pollen. Georgia Agr. Exp. Sta. Bull. 144: 134-161. 1924.* LC. linoénsis. | Studies of the staminate res sac and pollen of Hicoria Pecan. Jour. Agr. Res. 40: 1059-1104. 1930. & N. C. Wooproor, Fruit- sd differentiation and subsequent es ment of the flowers in the Hicoria Pecan, Ibid. 33: 677-685. 1926 . The development of the ag nut (Hicoria Pecan) froin flow- er to maturity. Jbid. 34: 1049-1063. 1927. & . Abnormalities in aaa I. Abnormalities in pecan flowers. Jour. Hered. 21: 39-44. 1930 — N. C. Development of the embryo sac and Re ie embryo of Hicoria can. Am. Jour. Bot. 15: 416-421. pls. 23-25. 1 1972] ELIAS, GENERA OF JUGLANDACEAE 45 2. Juglans Linnaeus, Sp. Pl. 2: 997. 1753; Gen. Pl. ed. 5. 431. 1754. Aromatic, resinous trees of temperate or subtropical habitats; bark fur- rowed, scaly; wood durable, dark-colored; branchlets terete, stout, with : " CBE OO 277 oS = j ee v te yn ¥ t , sta wi : carpellate flower, X 3; f, semidiagrammatic bs nce, x 1/2: , : longitudinal section of ca ellate flower showing ovule, perianth, and involucre, f u a Carpellate inflorescenc > J, t : tudinal section to show chambered pith, X 1. k, 1, J. cimerea: k, without involucre, X 3/4; 1, median cross section of nut, cotyledons unshaded, lacunae black, x 3/4. 46 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 chambered pith; terminal buds short to long, usually with 2 pairs of op- posite scales, the inner scales accrescent, the axillary buds + flattened, with 4 ovate scales. Leaves rarely paripinnate, few- to many-foliolate; petioles terete, canaliculate above, often swollen basally, the leaf scars large, conspicuous, elevated, obcordate, 3-lobed; leaflets conduplicate in vernation, usually ovate, acute to acuminate at the apex, generally oblique at the base, membranaceous, serrate or entire, sessile to short-petiolulate, pinnately veined. Flowers vernal, usually opening after the leaves unfold. Staminate flowers sessile or pedicellate, in many-flowered aments borne singly or in pairs from the lower axillary buds of the upper nodes, appear- ing from between the persistent bud scales; perianth 3—6-lobed, adnate to the ovate bract (Ficure 3, b); stamens 8—40, inserted on the perianth in 2 to many rows, the alternately ranked stamens alternate with the lobes, the filaments free, short, the anthers erect, oblong, a conspicuous con- nective usually present; ovary lacking. Carpellate flowers borne in few- flowered terminal spikes on branches of the year, the ovary of each flower inclosed by and adnate to a floral envelope composed of a villous involucre and a perianth, the involucre made up of a primary bract (often free to near the base or fused to the secondary bracts) and two secondary bracts (usu- ally free only at the apex), the bracts often laciniate at the apex, the four perianth lobes longer than the lobes of the involucre, appearing to come from the upper part of the ovary. Gynoecium 2 (rarely 3 or 4)-carpellate, 1-locular at the middle, 4-locular at the very base, a major septum divid- ing the lower part of the locule and a secondary septum (perpendicular to the major septum) bisecting each of these divisions at the very base of the ovarian cavity; stigmas large, spreading, often clavate, fleshy, fim- briately papillate. Fruit globose, ovoid, or pyriform, occasionally ob- scurely 4-angled, the husk (formed mainly from involucral tissue) inde- hiscent, + fleshy, glabrate to hirsute; nut globose to ovoid, often slightly flattened, indurated (the wall formed mainly from the ovarian wall, in our species thick and with lacunae in both wall and septum), longitudinally and irregularly rugose, 2-valved (upon germination), each valve inclosing half of each cotyledon, + ribbed on the one or both sutures. Seed with a thin seed coat covering the large embryo, + laterally compressed, shal- lowly or deeply lobed at the base, each cotyledon 2-lobed, filling the cavity of the nut; hilum minute, basal. Basic chromosome number 16. (Wallia Alef., Regia Loud.) Lectotype spectes: J. regia L.; see Nuttall, Gen. N. Am. Pl. 2: 220. 1818; Britton & Brown, Illus. Fl. No. U.S. ed. 2. 1: 578. 1913. (Classical Latin name for walnut, Juglans regia, derived from Jovis, of Jove or Jupiter, plus glans, acorn.) A genus of 18 to 20 species of temperate and subtropical regions of Europe, Asia, and the Americas. Six species are known from the United States; of these, the two in our area, Juglans nigra L. and J. cinerea L., are the only important forest species. Juglans Hindsii R. E. Sm. and J. californica S. Wats., 2n = 32, are Californian (central California and coastal southern California respectively), while J. major (Torr. ex Sitzgr.) 1972] ELIAS, GENERA OF JUGLANDACEAE 47 Heller occurs from central and southwestern Texas to southwestern New Mexico and central Arizona and (as var. glabrata Manning) south to Guerrero, Mexico, and J. microcarpa Berl., 2n = 32, ranges from western Oklahoma and western and southern Texas to northeastern Mexico. Man- ning (1957, 1962) also recognizes J. pyriformis Liebm.., J. hirsuta Manning, and J. mollis Engelm. in Mexico. Juglans Steyermarkii Manning is en- demic to Guatemala, and J. olanchana Standl. & Williams extends from Nicaragua to Guatemala with a disjunction (var. Standleyi Manning) in Colima and Jalisco, Mexico. The misnamed J. jamaicensis C. DUST = 32, occurs in Cuba, Hispaniola, and Puerto Rico. Juglans recurs in the Andes with some five species distributed from Colombia and Venezuela south to Peru, Bolivia, and Argentina. Juglans regia L. is native from southeastern Europe to India (and perhaps Central Asia) but is widely planted in many other temperate climates of the world. Juglans cathay- ensis Dode, J. cordiformis Maxim. (including J. ailantifolia Cam 7. Sieboldiana Maxim. non Goppert), 2n = 32, J. mandschurica Maxim., nm = 32, and J. stenocarpa Maxim. represent the genus in eastern Asia. In the subgeneric classification proposed by Dode (1909), who recog- nized many more species, Juglans regia constitutes sect. JUGLANS ($ Dios- caryon Dode), and J. cinerea is alone in sect. TRACHYCARYON Dode. All of the other Old World species are placed in sect. CARDIOCARYON Dode and the other New World ones in sect. RHYSOCARYON Dode. The black walnut, Juglans nigra L. (Wallia nigra (L.) Alef.), 2n = 32, occurs from western Massachusetts, central Vermont, and southern Ontario to southern Minnesota and southeastern South Dakota, south through eastern Nebraska to western Oklahoma and eastern Texas, and to northwestern Florida and Georgia (see Little, 1971, map 134-E). A plant of low altitudes, J. nigra grows best on deep well-drained soils, in- cluding good bottomlands. It can be found in mixed mesic forest, but rare- ly in pure stands. This species is easily distinguished by the lower sur- face of the leaflets and petioles being minutely downy with solitary or paired hairs, by the notched leaf scars that lack a fringe of hairs, and by the spherical fruit. Varietas oblonga (Marsh.) Fern. differs only in having ellipsoidal fruits. F lowering in early spring as the leaves appear or slightly afterward, the trees are apparently self-fertile, but individual trees have varying degrees of proterandry or proterogyny. Cross polli- nation usually insures better seed set and higher fruit yields. Distributed over much of the northeastern United States, Juglans cine- rea L. (Wallia cinerea (L.) Alef.), butternut, 27 = 32, occurs from western New Brunswick, New England, southern Quebec, and southern Ontario to Wisconsin and Minnesota, south to northern New Jersey, Vir- ginia, North Carolina, northwestern South Carolina, northern Georgia, Alabama, and Mississippi, Arkansas, and Missouri (see Little, 1971, map 133-E). Apparently growing best on streambanks and on well-drained gravelly or rocky soils, this species, like J. nigra, occurs sparingly through- out its range. It is found in many mesic forests, but it is better adapted to limestone soils than J. nigra is. Juglans cinerea is characterized by its 48 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 and petiole downy with glandular hairs; by the leaf scars with a hairy fringe along the upper margin; and by the ellipsoidal fruit. Juglans appears to be a homoploid genus with 2m = 32 (eleven species). Juglans nigra and J. cinerea are not known to hybridize, but hybrids of J. nigra with J. Hindsii and J. regia and of J. cinerea with J. cordiformis and J. regia have occurred under cultivation or have been produced arti- ficially. Hybrids of J. regia with J. Hindsii and J. cordiformis are also leaflets, downy with fascicled hairs, especially beneath; by the branchlets The characteristic chambering of the pith in Juglans was shown by Harlow (1930) in J. nigra and J. cinerea to be formed at the end of the growing season. In spring the young twigs of the year still have homo- geneous pith. Chambering in J. nigra begins in late August, but in J. cinerea not until late September or early October. The first sign of cham- bering is a change in the color of the pith, followed by the formation of several cavities near the base of the year’s growth (after elongation has ceased) and then throughout the twig. Both prodigious pollen producers, J. nigra and J. cinerea were suspected of contributing to spring pollen allergies, but since neither occurs abun- dantly anywhere, it is doubtful that either is of importance in this con- nection. The pollen grains are 2-celled when shed. The pollen is rela- tively variable but is basically tectate, periporate, and heteropolar with the (2—)6-18(-37) pores restricted to one polar hemisphere and occa- sionally overlapping (Whitehead, 1965). An increase in pollen size and in pore number, as seen in J. regia and J. nigra may be indicative of specialization within the genus. Unlike genera of the Fagaceae and Betulaceae, in Juglans the ovule is well developed at the time of pollination, with fertilization occurring 4-8 days later. Nuclear endosperm has been reported in J. regia (Nast, 1935, 1941) with approximately 1000 free nuclei present before cell wall for- mation occurs. Embryogeny in J. mandschurica is of the Asterad type (Langdon, 1934). Polyembryony has been reported in J. nigra. After flowering the fruits of our species mature in September or Octo- ber of the same season, dropping shortly after the leaves fall. Hypogeous germination usually occurs the following spring with a strong taproot produced by the seedling. Although Berry (1923) suggested that Juglans leaves have been found in Middle Cretaceous deposits, verifiable remains occur only from the Upper Cretaceous to the present. Apparently Juglans was much more widely distributed over the Northern Hemisphere in the past, especially during the Eocene, Miocene, and Pliocene. It has long been known that certain plant species have reduced growth rates and/or reduced reproductive success when growing under or close to plants of certain species of Juglans, e.g., J. nigra and J, regia, but not J. cinerea. The inhibitory substance produced by these species has been determined to be juglone, CioH¢O2, a red crystalline compound, that is produced by oxidation of hydrojuglone, a water soluble substance pro- 1972] ELIAS, GENERA OF JUGLANDACEAE 49 duced in the leaves, fruit, husk, bark, and perhaps the roots. This sub- stance apparently can be washed from the leaves into the ground by rain. This general phenomenon, allelopathy, the inhibition of one plant by a substance or substances secreted by another has only relatively recently received serious attention from ecologists (cf. Muller, 1969). Other ex- amples have been discovered in several unrelated families, including the Gramineae, Simaroubaceae, Myrtaceae, Labiatae, and Compositae. The beautifully grained brownish wood of Juglans nigra is perhaps the foremost cabinet wood of North America; that of J. cinerea is considered to be less valuable. Selected strains and cultivars of J. nigra (more than 100) have been developed, mostly for nuts with larger, thinner shells that will crack easily and for low seed-abortion rates. The seeds are used largely in ice cream and confectionery and as a flavoring. The nuts of J. cinerea are occasionally used locally but are not commercially im- portant. A few cultivars, based on foliage characters, have been selected as ornamentals. The European J. regia is the principal source of walnut lumber and nuts in the Old World. It is successfully grown in various parts of the United States and is extensively cultivated in the mild climates of California. REFERENCES: Under family references see BAILLON (p. 405), Berry (pp. 73-80), Conve & STONE, FoweELts (pp. 202-210), HuTCHINSON (p. 139), JayNEs (pp. 203-211, 212-223), Kurz & Goprrey (pp. 36-38), LITTLE (1953, pp. 213-215; 1971), SARGENT (1922, pp. 169-176), WHITEHEAD (1963, 1965), and Vines (pp. 123- 127). AKIMOCHKIN, N. G. A natural interspecific hybrid from the genus Juglans. (In Russian.) Izv. Glavn. Bot. Sada SSSR (Bull. Jard. Bot. URSS) 44: 11-16. 1961. [J. regia X J. cinerea.] Bascock, E. B. Walnut-oak hybrid experiments. Am. Breed. Assoc. Rep. 6: 138-140. 1911. . Studies in Juglans I. Study of a new form of Juglans californica Wat- son. Univ. Calif. Publ. Agr. Sci. 2: 1-46. pls. 1-12. 1913; II. Further ob- Servations on a new variety of Juglans californica Watson and on certain supposed walnut-oak hybrids. /bid. 47-70. pls. 13-19. 1914; III. (1) Further evidence that the oak-like walnut originates by mutation. (2) A parallel mutation in Juglans Hindsii (Jepson) Sargent. bid. 71-80. pls. 20, 21. 1916. . A new walnut. Jour. Hered. 6: 40-45. 1915. [J. californica var. quercina. | - Walnut mutant investigations. Proc. Natl. Acad. Sci. U. S. A. 1: 535- S37 1015. BENsoN, M., & E. J. WeLSForp. The morphology of the ovule and female flower of Juglans regia and of a few allied genera. Ann. Bot. 23: 623-633. 1909. Berry, E. W. Notes on the geological history of the walnuts and hickories. Pl. World 15: 225-240, 1912. - A petrified walnut from the Miocene of Nevada. Jour. Wash. Acad. Sci. 18: 158-160. 1928. [J. nevadensis, sp. nov. ] 50 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ———. A walnut in the Pleistocene at ee Oklahoma, Jbid. 19: 84-86. 1929. [J. rupestris var. pleistocenica, var. no . A walnut from the Chesapeake Miocene. Did. 24: 227-229. 1934. [J. calvertiana, sp. nov Brooks, M. G. Effect of black walnut trees and their products on other vege- tation. W. Va. Agr. Expt. Sta. Bull. 347. 41 pp. 1951.* [See also E. J. SCHREINER, Morris Arb. Bull. 4: 94-96. 1949. 554-556. 1946.* Cook, O. F. Evolution of compound leaves in walnuts and hickories. Jour. Hered. 14: — 1923. Davey, A. J. Note on the structure of the epicotyl in Juglans nigra. New Phytol. 34: sets 1935. Dope, L. A. Contribution a l’étude du genre Juglans. Bull. Soc. Dendrol. France 2: 67-98. 1906; 11: 22-50. 1909; 13: 165-2 1909. Enpbo, S. The American white walnut or butternut, Juglans cinerea L., from the Upper Pliocene of Japan. Jour. Wash. Acad. Sci. 23: 305-308 - 1933. GorBunov, M. G. New species of Juglans from the Tertiary chain of West- ern Siberia. (In Russian.) Bot. Zhur. 41: 658-666. 1956. Gray, J. Temperate pollen genera in the Eocene (Claiborne) flora, Alabama. Science 132: 808-810. 1960. [Juglans, 810. Hatt, H. M. Walnut pollen as a cause of hay fever. Science 47: 516, 517. 1918. UJ. Hindsiz in California.] Hartow, W. M. The formation of chambered pith in the twigs of butternut and black walnut. Jour. Forestry 28: 739-741. 1930. [J. cinerea, J. nigra.] HEnroTIN, M. Anatomie de la tige et des bourgeons axillaires sériés dans le i 1939. Hom, T. Morphological es of Carya alba and Jugions nigra. Bot. Gaz. 72: 375-389, pls. 15, 16. 1921. Inizecx, J. 8S. The pepe walnuts. Am. Forestry 27: 699-704. 1921. KRAJICEK, J. E. Some factors affecting oak and black walnut reproduction. Iowa State Coll. Jour. Sci. 34: 631-634. 1960. [Germination of Quercus spp. and J. nigra.] Lamps, W. H. A tricarpellary walnut. Torreya 12: 290, 291. 1912. LANGDON, L. M. maces! a Carya and Juglans, a comparative study. Bot. Gaz. 96: 93-117. pl. 1. McDermott F. A. A ieiineateanlaey walnut. el 13: 137-139. 1913. Tri- and tetracarpellary walnuts. Jbid. 14: 27. 1914, McVaucu, R. Suggested phylogeny of Prunus rita and other wide-ranging phy lads in North America. Brittonia 7: 317-346. 1952. [J. microcarpa, 332 McKay, J. W., & H. H. McKay. eh ie 9 in Juglans intermedia Carr. (Abstr.) Am, Jour. Bot. 28(Suppl.): 4s. 1941. MANNING, W. E. The genus Juglans in Mees and Central America. Jour. Arnold Arb. 38: 121-150, 1957. . The genus Juglans in South America and the West Indies. Brittonia 12: 1-26. 1960. a aad A. B. Antagonism of the walnuts (Juglans nigra L. 8 of cinerea L.) certain plant associations, Phytopathology 15: 773-784. 1 Mocoraae. . & I. G. Motorxovsxr. Formation of Sheen blossoms in walnuts during the second oni. (In Russian.) Moskov. Obshch. Isp. Pri. Biol. 57: 74-78. 1952. 1972] ELIAS, GENERA OF JUGLANDACEAE 51 MULter, C. H. Allelopathy as a — in ecological process. Vegetatio 18: 348- 357. 1969. [Includes Juglan Nast, C. G. Morphological sored of the fruit of Juglans regia. Hilgardia >: shins 1935. 4 embryogeny Hive seedling morphology of Juglans regia L. Lilloa 6: 163- pots pls. 1-17. Pierce, N. B. A new nee Science 37: 613, 614. 1913. [J/. quercifolia, sp. nov., supposedly a hybrid of Juglans californica and Quercus sp., but see BABCOCK Porsr, J. 1: ‘Nameércs cromosémicos y algunas observaciones bioldgicas sobre he especies americanas del género Juglans. Darwiniana 14: 156-160. 966. [J. australis, J. boliviana, & J. nigra, n = 16. AE Te S. J. Walnut woods — true and false. Trop. Woods 18: 4-29. 1929. REHDER, A. Abnormal fruits of Juglans nigra. Jour. Arnold Arb, 9: 27-29. 1928. Scuaap, N. W., & E. E. Witson, Structure and seasonal development of sec- ondary phloem of Juglans regia. Canad. Jour. Bot. 48: 1049-1053. pls. 1- Seuabienae J. H. Key to Ohio walnuts based on twig characters. Ohio Nat. 5: 307. 1905 ScHUSTER, C. E. Differences in heterosis of walnut hybrids. Jour. Hered. 28: 216, 217. 1937. [J. regia * J. nigra SHCHEPO?’ Ev, F. L. On metaxenia in black = (Juglans nigra). (In Rus- sian.) Priroda Leningrad 39: 59-61. 1950. y Bieexval blossom of the walnut tree. (In Russian.) /bid. 43: 92-94. Suet D. V. Inner icra of the fruit of Juglandaceae. Proc. Okla. Acad. men: Let t7, THomsen, H. H. tee Hee ae central California black walnut, native or introduced? Madrofio 17: 1-10. 1963. WItson, L. R., & R. M. Eilicer Fossil evidence of wider post-Pleistocene range for butternut and hickory in Wisconsin. Rhodora 44: 409-414. 1942. [J. cinerea & Carya cordiformis. | ARNOLD ARBORETUM Present address: HARVARD UNIVERSITY THE CAry ARBORETUM CAMBRIDGE, MASSACHUSETTS 02138 THE NEw YorkK BOTANICAL GARDEN MILLBROOK, NEw YorK 12545 52 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 REVISION OF STENOPETALUM (CRUCIFERAE) ELIzABETH A. SHAW OF THE SEVERAL GENERA OF Cruciferae endemic in Australia, the most striking is Stenopetalum, a genus of eight species characterized by flowers in which the calyx is erect and closed and the narrow petals are drawn out into a long, slender and often twisted tip. Flowers of this general sort are seen in some American genera such as Schizopetalon and Lyrocarpa, but it is a type not common in the Cruciferae and among Australian genera these elongated petals are so distinctive that each species of Stenopetalum was initially described in that genus and none has undergone the nomen- clatural vicissitudes of other Australian members of this family. Similar flowers occur in Carinavalva glauca Ising, but this rare species, seemingly restricted to a small area in the far north of South Australia, has many- seeded obcompressed siliques, a fruit type not found in Stenopetalum. Stenopetalum lineare was the first Australian species of this family known to science, the specimens brought back by Robert Brown who served from 1801 to 1803 as naturalist under Matthew Flinders on H. M. S. Investigator, Flinders’ charge, politically motivated and intended to forestall French activity in Australian waters, was to complete the sur- vey of the coast of New Holland and it is interesting that his orders from the Lords Commissioners for the Admiralty included detailed instruc- tions for the construction on the quarter deck of the Investigator of a “plant cabin,” intended for “such plants, trees, shrubs, &c., as they [the naturalist and the gardener] may think suitable for the Royal Gardens at Kew . . . .” After one and one-half years of exploration the Investigator was pronounced unfit for the return voyage to England so the living plants collected by Brown and Peter Good, the gardener, were transferred at Sydney to H. M.S. Porpoise. Fortunately Brown and his artist, Ferdi- nand Bauer, remained in New South Wales, retaining there the bulk of the dried collections, for the Porpoise and the living plants were lost — wrecked on the Great Barrier Reef. Brown returned to England in 1805 and the collections became the basis for his Prodromus Florae Novae Hollandiae. During Brown’s years in Australia he became familiar with some of the most spectacular plants of the flora and it is perhaps understandable that he was not impressed by Australian Cruciferae. His published work makes only the briefest reference to the family, and although Brown rec- ognized S. lineare as representing a distinct genus, he did nothing with the material and finally passed it on to De Candolle for description in Systema Naturale. The next species described, S. robustum of Western Australia, was correctly placed by Endlicher in Stenopetalum as were the species subsequently described by Mueller and by Bentham. However, 1972] SHAW, REVISION OF STENOPETALUM 53 of three species figured during the 1840’s in Jcones Plantarum and de- scribed by one or the other of the Hookers in Stenopetalum, two are now referred to Menkea Lehm. and one is the widespread halophyte Hutchinsia procumbens. MORPHOLOGY AND BIOLOGY Most species of Stenopetalum are annuals, although S. decipiens of central Australia is a woody perennial and a coastal form of the wide- spread and variable S. lineare is woody and perennial. The genus is rep- resented in all states, including Tasmania, and in the Northern Territory and is predominantly Eremaean although S. robustum of Western Australia is coastal in habitat and there is the coastal form of S. lineare just men- tioned. With the exception of S. robustum, in which the valves are com- pressed parallel to the septum, all species have short siliques which are globose to (ob-)ovoid and round in section. The pedicels are variable in length and orientation and provide some useful diagnostic characters. The seeds of all species are strongly mucose when moistened, the thick perisperm immediately producing a broad band of radiate mucus. Ex- cept in the tuberculate or verrucose seeds of S. nutans and S. velutinum, the testa is smooth and there are otherwise no characters of real diagnostic or taxonomic value observable in the seeds. Pubescence is quite variable. Stenopetalum decipiens, S. velutinum, and S. lineare are pubescent with irregularly branched trichomes. Those of the first two species are nearly sessile with very long, horizontally spreading rays which in S. decipiens are almost inextricably interwoven into a dense felt. Trichomes of S. lineare are similar but smaller and less elaborately branched. Stenopetalum nutans is unique in the genus in be- ing sparsely pubescent with malpighiaceous trichomes. The other four species are apparently glabrous, papillose, or pubescent with simple tri- chomes. This represents a difference of degree rather than of kind and on plants of S. robustum and S. filifolium there can be both papillae and short, simple, rather crinkled trichomes. The cauline leaves are generally entire or dentate in S. velutinum, S. decipiens, S. robustum, S. filifolium, and S. nutans; S. anfractum, S. lineare, and S. sphaerocarpum have lower cauline leaves which are pinnatisect or trifid as in some species of Arabi- della, although in all three the upper leaves may be entire and the coastal form of S. lineare includes plants with all the cauline leaves entire. Although all species have elongated petals there is a marked variation in size, from the 3 to 4 cm. long petals of S. mutans and of some forms of S. robustum to the 5 mm. long petals of S. sphaerocarpum. Stenopetalum robustum and S. filifolium, both of Western Australia, have white or brightly colored yellow to orange petals, but the other six species are un- usual in the family in that the petals are dully colored and range from olive-green through purplish to chocolate-brown. In all species the elon- gated tips are circinnately ‘coiled in the bud. As anthesis proceeds the un- coiling of the petals forces open the sepals at the apex of the bud, al- 54 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 though they never spread widely and the calyx remains closed. In no species are the nectaries particularly well developed and the glandular tissue never completely encircles the ovary and stamens. It usually is present only around the single stamens although there is sometimes a lobe between the members of the paired stamens. This occurs occasionally in most species. Plants from the northern part of the range of S. robustum do show a peculiar development of the glands so that on either side of each single stamen is a flat protruding lobe of tissue. This is conspicuous both in flower and fruit and can be useful in distinguishing S. robustum from S. filifolium when the plants are only in flower. In S. filifolium there is development of glandular tissue above the point of insertion of the single stamens; this is shelf-like and although sometimes bilobed is never the “‘horseshoe” seen in S. robustum. Examination with the scanning electron microscope of pollen of four species has shown it to be subglobose to ovoid, tricolpate and coarsely reticulate, the lumina large and the muri correspondingly narrow. It is hoped to examine the pollen of remaining species in the course of a more comprehensive survey of pollen types in the family. Six species of Stenopetalum have been grown under greenhouse condi- tions, S. nutans at the Botanic Garden, Adelaide and S. Sphaerocarpum, S. velutinum, S. lineare, S. anfractum, and S. decipiens at Harvard Uni- versity. Of these species, all except S. decipiens are self-compatible, and S. lineare and S. sphaerocarpum sometimes produce cleistogamous flowers. In normal flowers, the anthers dehisce just as the uncoiling of the petals begins and at the time that the apex of the calyx has been forced open, pollen has been shed. I have seen in the field flowering plants of all species except S. decipiens, but I have never seen them visited by insects. The only species with scented flowers is S. robustum, the odor described some as “‘vanilla-like,” by others as “musty.” Self-compatibility is advantageous to plants of an environment so inhospitable as central Australia where the capacity for seed production and dispersal from small populations or from a single plant during favorable seasons, not of regu- lar occurrence, is of prime importance. It is true also that self-compati- bility and the potential for autogamy greatly increases the chance that a species will reach suitable habitats more and more peripheral from its center of origin. In central Australia such habitats are isolated and the capacity for autogamy may be associated with the distribution of species such as S. nutans, S. anfractum, and S. sphaerocarpum, sparsely spread across South Australia and the Northern Territory into some very arid parts of Western Australia. INTERSPECIFIC RELATIONSHIPS Relationships among the species must be considered in the light both of present distribution and of post-Pleistocene climates in Australia. It is generally agreed that at some time between the late Pleistocene and the mid-Recent, Australian biota were greatly affected by the onset of 1972] SHAW, REVISION OF STENOPETALUM 55 a period of severe aridity, during which the vegetation of central and south-central Australia was hard hit and, perhaps, nearly wiped out, some species retreating into southeastern Australia, others surviving in refugia in the ranges of central Australia (Crocker & Wood, 1947). This “Great Arid Period” was followed by two or three pluvial/arid (but less severely so) cycles during which there was a gradual advance of vegeta- tion from less severely affected areas and from the refugia into present- day arid Australia. Chromosome counts (see TABLE 1) indicate two haploid numbers, 4 TABLE 1. Chromosome Numbers of Stenopetalum species COLLECTION NUMBER SPECIES N 2N AND LOCALITY SELF-COMPATIBLE S. anfractum 10?,12? Shaw 654 yes Western Australia S. decipiens & Nelson 1814 no Northern Territory S. filifolium SII+I1* Turner 5367 * Western Australia (as S. croceum Bunge) S. lineare 5 * 10 Turner 5297 * Western Australia Beauglehole 20701 yes Northern Territory S. sphaerocarpum 5 10 Shaw 664 yes Western Australia S. velutinum 4 8 Eichler 17563 yes South Australia * Turner (1967). and 5, in the genus. The number in S. anfractum probably is 5, although repeated attempts to obtain a clear figure in root tip cells were unsuccessful and the somatic count is given as questionably 10 or 12. The counts of “n = 8 for S. decipiens and S. velutinum do corroborate the morphological evidence which sets apart these species from the others. These two dense- ly pubescent species, S. velutinum and S. decipiens, have a center of dis- tribution in the central Australian ranges and the present distribution is not inconsistent with advance from a refugium in the MacDonnell — Kirchauff — George Gill system, although both species show a disjunc- tion between south-central Queensland and the area north and west of the Simpson Desert dune systems. The annual and self-compatible S. velutinum is far more widely spread than is the perennial and apparently self-incompatible S. decipiens and is found as far south as Ooldea in outh Australia and westward at least to the vicinity of Laverton in Western Australia. Another cluster of related species is formed by the 56 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Eremaean species, S. lineare, S. anfractum, and S. filifolium (all probably with a base number of 5). The last species differs in having brightly colored petals and I include it here with some hesitation for the floral structure is much like that of some forms of S. robustum. However, these three species do have in common more or less oblong siliques on short, erect or spreading pedicels. Stenopetalum lineare which now is widely spread and common in central Australia is represented also in South Australia, Tasmania, and Victoria by a coastal form which may have been geographically distinct since the Recent aridity. Stenopetalum robustum of the Southwestern Province of Western Australia is the only species to have compressed siliques, while S. sphaerocarpum, although represented by populations scattered across southern Australia, is neither coastal nor truly Eremaean and differs morphologically with very small flowers and globose siliques on rigidly spreading horizontal pedicels. Stenopetalum nutans, a large-flowered Eremaean species with pendent obovoid siliques and malpighiaceous trichomes, also stands somewhat isolated morpholog- ically. The discontinuities in morphology among these species and groups of species support the view that they are the remnants of a more highly developed pattern of speciation, disrupted during the arid period or periods, which cannot now be satisfactorily reconstructed. INTERGENERIC RELATIONSHIPS The genus is basically Eremaean and thus belongs to an Australian floristic element of uncertain origin. Burbidge (1960) suggested that en- demic genera among the Eremaean Cruciferae, Chenopodiaceae, Composi- tae, and Aizoaceae are derived from a littoral or sand dune flora present dur- ing the Cretaceous when Australia was dissected by epeiric seas, elements of which flora might have reached Australia from the northern hemisphere by migration along the shores of the Tethys Sea. In addition to weak- nesses in this theory pointed out by Burbidge herself, the most recent re- construction of Pangaea (Dietz & Holden, 1970) indicates that even by the end of the Jurassic, land breakup in the Southern Hemisphere was so far advanced that the Australian landmass was attached only to 1972] SHAW, REVISION OF STENOPETALUM 57 Antarctica. Prantl (1891) had placed Stenopetalum, along with Menkea, in a subtribe of the Schizopetaleae, associated with American genera such as Mancoa, Tropidocarpum, Matthewsia, and Schizopetalon. The other sub- tribe, Physariinae, recognized by Prantl is entirely American including Lesquerella, Physaria, and Synthlipsis. Although much of Prantl’s system is fantasy, part of it described by a later author as “ein phylogenetisches Unding,” it may be that he was in this case nearer the truth than was Schulz. However, many of the more than 370 genera in this family are the termini of otherwise extinct evolutionary lines, a situation which makes so difficult the delimitation of tribes in the Cruciferae. In the ab- sence of evidence to the contrary, it is best to regard Stenopetalum, geo- graphically isolated in Australia, as one such endpoint, the nearest related genus being, perhaps, Menkea. COMMON NAMES AND USES Stenopetalum is of no economic significance save that in favorable sea- sons in central Australia some species such as S. Jineare occur in such abundance that they become part of the diet of the aboriginal population. The central Australian species are highly flavored and quite hot and probably are used both for their nutritive value and mustardy taste. Cle- land & Tindale (1954) report that seeds of S. ineare were eaten at Yalata on the Great Australian Bight, while Cleland & Johnston (1937) noted that in the Musgrave Ranges of South Australia S. velutinum, called “unmuta” by the Pitjanjara, is cooked on hot stones with grass, the leaves then stripped off and eaten. However, natives at Haast’s Bluff in the Northern Territory informed Cleland & Tindale (1959) that both S. lineare and S. nutans are fit only for emus and are too strong for human consumption; S. lineare “makes you giddy if you try to eat it.” However, it may be S. anfractum here referred to; both species occur at Haast’s Bluff and are mixed, under the name S. Jineare, in Cleland’s collections. The name for S. lineare, S. nutans, and S. velutinum used by members of the Pitjanjara, Pintubi, and Aranda tribes is “unmuta,” “ ‘enmerta” or something similar, although Johnston & Cleland (1942) recorded “karra garra” and “kara-kara” for S. lineare at Ooldea in South Australia. ACKNOWLEDGEMENTS I wish to thank the Nuffield Foundation (Australia) for financial as- sistance which made possible field work in Western Australia during 1966. I should like also to thank Mrs. Lily Riidenberg who so carefully made chromosome counts from the plants grown at Harvard University and Mrs. Maria Szent-Ivany of Adelaide who prepared the illustrations. The herbaria housing the specimens seen are cited by the abbreviations of Lanjouw & Stafleu (Regnum Veg. 31: 1964); I am grateful to the cura- tors of these institutions for the privilege of examining this material. For each species only a representative sampling of specimens is cited. 58 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 SYSTEMATIC TREATMENT Stenopetalum R. Br. ex DC. Syst. Nat. 2: 513. 1821. Plants erect glabrous or pubescent annuals or perennials; basal leaves, if present, pinnately divided, but often lost early in the life of the plant; cauline leaves pinnatisect or trisect or entire; inflorescences ebracteate, the buds usually ellipsoid, sometimes globose; aestivation circinnate; sepals erect (the calyx closed), the lateral saccate; petals white to orange or brown to olive-green, unguiculate, the blade elongated and often pro- duced into a long narrow tip; stamens erect, tetradynamous; glandular tissue variable; ovules 4—20 per locule; siliques completely dehiscent and bilocular, globose, ellipsoid or (ob-)ovoid, round or compressed parallel to the septum, glabrous on the interior and exterior; septum entire; styles very short or obsolete; stigmas depressed-capitate; seeds oblong to ellipsoid and flattened, copiously mucose when moistened; embryo no- torrhizal or pleurorrhizal. TYPE SPECIES: Stenopetalum lineare R. Br. ex DC., Syst. Nat. 2: 513. 1821 KEY TO THE SPECIES A. Plants congitehen papillose or, if pubescent, the trichomes simple or malpi- ghiaceo B. Pedicels erect or divaricately spreading at maturity. C. Main stems clearly anfractuose; cauline leaves pinnatisect with lin- ear segments; petals purplish or olive-green; — 4-10(15) mm. MU eo a ae ae . S. anfractum. C. Main stems nearly straight; all cauline leaves Bi and filiform or gs dentate; petals bright yellow-orange; gre 3-4 mm. lon by Sa ee ae eae a a ge ee S. filifolium. rownish. E. Pedicels 3-8 mm. long and stout, usually resected spreading ; plants glabrous or papillose; siliques Mohbete eee ety Sse tm es Se ee De Be ree ea ra. Sphaerocarpum. E. Pedicels 8-15(20) mm. long and slender, sd rea pu- bescent with malpighiaceous trichomes; siliques obov ee ee ee . nu D. Siliques orbicular to ellipsoid or obovoid, compressed parallel to the septum; petals white, yellow or orange. ............ 6. S. robustum. A. Plants pubescent with irregularly branched or stellate trichomes. F. All parts, including leaves, densely pubescent; all leaves entire or re- motely dentate; sepals with gibbous thickening at the tip. G. Plants annual; ovules 10-20 per locule; siliques 4-7 long, seeds tuberculate. . 8. S. velutinum. G. Plants perennial, woody at the base; ovules 3— 10 per locule: siliques 7-11 mm. long; seeds smoo i pe ie deci cipiens F. Upper parts of plant glabrous or sparsely gears: lower cauline leaves pinnatisect to trisect; sepals not thickened at the ti 1. S. lineare. 1972 | SHAW, REVISION OF STENOPETALUM 59 1. Stenopetalum lineare R. Br. ex DC. Syst. Nat. 2: 513. 1821. Map 1 HototyPe. Victoria? Hab. in ora australi Novae-Hollandiae, R. Brown (G—pc); isotypes? (BM, E, MEL, P). S. lineare var. canescens Bentham, Flora Austral. 1: 78. 1863. Holotype. Vic- toria. Port Phillip, F. Mueller (x). Plants annuals or, in coastal sites, often perennial, subglabrous to dense- ly pubescent; trichomes sessile or shortly stipitate, stellate or irregularly branched, densest toward the base of the stems; stems to at least 5 dm. high, but the height most variable, usually branched, erect or, especially in coastal plants, sometimes decumbent with erect laterals; basal leaves usually less than 5 cm. long, exceptionally to 10 cm., pinnatipartite with linear to narrowly obovate segments, these sometimes finely dentate or, in coastal plants, often entire; cauline leaves to 6 or 7 cm. long, exceptionally to 10 cm., generally pinnatisect or trisect, the uppermost sometimes entire and filiform; inflorescences rather loose, elongating before anthesis, the buds obovoid or ellipsoid; sepals 3-6 mm. long, elliptic-oblong to narrowly obovate, the lateral strongly saccate at the base; petals (4)6-13 mm. long, chocolate-brown to olive-green, the linear claw expanded into an oblong blade roughly U-shaped around the single stamens (open on the outer side); infructescences loose and elongated (to 30 cm.); pedicels less than 5 mm. long, usually nearly erect, rarely horizontally spreading or even reflexed; siliques 3.5-7.5 mm. long, oblong and terete; stigmas capitate and slightly elongated; ovules 8-14(18) per locule; seeds 0.8— 1.5 mm. long, oval to oblong and compressed, the testa red- or orange- brown, smooth. 2 = 5, 2m = 10 DIsTRIBUTION. Widely spread in the interior of southern Australia, but also found in Tasmania and on the coasts of South Australia and Victoria. Queensland. Gilruth Plains, Cunnamulla, Allen 626, 666 (CANB); Nocka- tunga Station, Blake 11820 (Bri); Birdsville, Blake 12211 (BRI). New South Wales. Milparinka, Collier in 1910 (Nsw). Wilcannia, Kennedy in 1885 (MEL); West Wyalong, Ingram in 1958 (NE); Broken Hill, Morris 2332 (prt, K); Bal- ranald, Lucas in 1878 (MEL). Victoria. Kulkyne National Park, Beauglehole 1137 (mex); East Whipstick near Bendigo, Robbins in 1948 (MEL); Werribee, Fullager (MEL); Murrayville, Williamson in 1928 (c); Mt. Arapiles, Hicks in 1960 (ap), Tasmania. Bellerive, Lucas in 1923 (Nsw); Ralph’s Bay, Comber 1611 (£, K); R. Jordan, Rodway in 1898 (nsw). South Australia. 17 miles north of Mt. Gason bore, Lothian 355 (ap); Frome River, 2 km. north of Marree, Shaw 191 (ap); Koonamore, Osborn in 1928 (syp); 26 km. south The Gums H. S., Shaw 175 (ap); Port Noarlunga, Griffith in 1912 (ap); Beachport, Tate in 1882 (ap); Ross River near Ludgate’s Well, Shaw 492 (ap); 11 km. north of Mable Creek, Shaw 522 (ap); Wudinna, /sing in 1938 (ap); Bitter Well, Coo- maba, Cleland in 1929 (ap); between Tomkinson and Mann Ranges, Clela } in 1954 (ap); Fowler’s Bay, Richards (AD, MEL). Northern Territory. 35.9 miles 60 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 s. w. Tobermory H. S., Chippendale (NT); Atnarpa Station, Beauglehole 20701 (mEL); 3 km. north of Charlotte Waters, Shaw 483 (a, AD); Haast’s Bluff Re- serve, Soe in 1957 (AD); 40 km. west of Erldunda, Cleland in 1935 (AD). Western Australia. 5 miles south of Giles, Kuchel 191 (AD); 3 km. north of Old Eucla, Wilson 1647 (ap); Kalgoorlie, 5 miles along Coolgardie Road, Burbidge 1875 (BRI, CANB); Israelite Bay, Brooke in 1885 (MEL); Stirling Range, Mor- rison in 1902 (xk). The type locality probably is Port Phillip Bay, visited by Brown in April, 1802, and that is the locality given on those specimens cited as possible isotypes. However, there is in the herbarium of the British Museum (Natural History) a specimen, labelled “Bass Strait’? which may have been collected in the Kent Group by Brown during his stay following Flinders’ departure. Stenopetalum lineare is both the most common and the most widely spread species in the genus and is the only one known from Tasmania. Although in inland parts of Australia this species is always an herbaceous annual, plants from coastal parts of South Australia, Victoria, and Tas- mania often are woody, the main stems decumbent with erect laterals, and perennial. Bentham’s var. canescens, described as ‘‘A low branching more robust form, the young shoots slightly hoary with a minute stellate pubescence, and the leaves rather thicker,” is based upon a typical plant of the coastal form, woody at the base and with the leaves mostly entire and probably quite fleshy. Although De Candolle described Stenopetalum (S. lineare) as glabrous, I have never seen plants which are completely without trichomes. The density of pubescence is, however, quite variable and older plants may have trichomes only in the leaf axils. The original (and incorrect) description of S. lineare as glabrous led urs to describe, in fact, to re-describe, the littoral form as var. canescens, a name which has been wrongly applied to plants from the interior, the result being confusion of S. velutinum with S. lineare. From Tasmania and the mainland coast toward the interior can be seen a gradual change in fruit shape. Plants of the coastal form have narrow cylindrical fruit; the average ratio of silique length to width (across the replum) in the Tasmanian plants I have seen is 4.7:1. In the interior, S. lineare has fruit more nearly ellipsoid. The average ratio of length to width is 3.0:1 in plants of the Flinders Range and Olary Spur in South Australia and 2.4:1 in plants from the Lake Eyre basin. These plants have seeds which are larger and exude mucus more rapidly than do those of coastal plants. To a plant of S. lineare goes the distinction of having been the first plant collected in Australia by Ferdinand Mueller. Mueller landed at Port Adelaide early in December, 1847 and the plant (MEL 10781) is labelled in his hand, “In poeinsulae Levevre oris [Port Adelaide] /Dec. 1847/F. Mueller/prima fecit planta a me in terra Australasia collecta; ex hac causa nomen triviale ‘gratulatorium’ adposis.” Although Mueller often used the name “Stenopetalum gratulatorium” (or gratulorum) on labels, it it was never published. 1972] SHAW, REVISION OF STENOPETALUM 61 Map 1. Stenopetalum lineare (dots); S. robustum var. robustum (triangles), var. pedicellare (squares). 2. Stenopetalum anfractum E. Shaw, sp. nov. Figure 1; Map 4. Plantae annuae, glabrae vel sparse papillosae; caules rigiduli et multum ramosi, caulo principali anfractuoso-flexuoso; folia caulinea pinnatisecta; petala olivacea vel obscure purpurea, lamina longe et anguste producta; pedicelli fructiferi brevissimi crassique; siliquae oblongae vel ellipsoideae, teretes; ovulae 8-15 in loculo. 2n = 10? vel 12? Hoxortype. South Australia. Ca. 25 km. west of Welbourne Hill, Wib- ley 927 (ap), isotype (GH). Plants slender annuals, glabrous or sparsely papillose; stems 1—4 dm. long, rigid, usually much branched and thus anfractuose, terete to sharply angled; basal leaves rosulate and apparently similar to the cauline leaves, but rarely present on flowering or fruiting plants; cauline leaves 1-6 cm. long, pinnatisect with 1 to 3 narrow and nearly linear lobes per side, the upper leaves sometimes entire and filiform; inflorescences rather loose and few-flowered, the buds ellipsoid; sepals 3.5—5.5 mm. long, oblong to narrowly obovate, not saccate; petals 8-13 mm. long, olive-green to 62 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Ficure 1. Stenopetalum anfractum. Habit, < 1/2. purplish, the claw narrow but expanded into the suborbicular to elliptic blade, this elongated into a narrow tip more than half the total length of the petal; stamens 2.5—4 mm. long; glandular tissue + square around the single stamens and subtending the adjacent paired stamens; infructescences loose, the pedicels 4-10(15) mm. long, short and stout, appressed or 1972] SHAW, REVISION OF STENOPETALUM 63 slightly spreading; siliques 4—6(7) mm. long, oblong to ellipsoid and nearly terete; stigmas capitate and sometimes bilobed; ovules 8-15 per locule. DIstrIBUTION. On sandy soils in the southern part of the Northern Territory, in South Australia to the west of Lake Eyre, and in western Australia to the longitude of Cue. South Australia, 34 km. north of Old Lambina H. S. ruins, Shaw 447 (A, aD); 22 km. south of Mt. Willoughby, Shaw 516 (AD); 26 miles north of Woorong Creek, Beauglehole 20167 (MEL); De Rose Hill Station, Lothian 858a/54 (AD); Wild Cat Bore, 25 km. s. w. of Everard Park H. S., Whibley 1119 (ap); Erna- bella, Cleland in 1933 (ap). Northern Territory. 32 miles n. e. of McDonald Downs Station, Lazarides 5925 (AD, CANB, MEL, NSW, NT, PERTH); 5 miles e. of Coniston H. S., Chippendale (ap, NT); 35 miles s. w. of Alice Springs Township, Perry 5477 (AD, BRI, CANB, NT); the “Pudding” area, Mt. Olga complex, Schodde 435 (AD, CANB). Western Australia. Giles, Hill 1377 (ap); Mt. Squires, Barrow Range, Helms in 1891 (AD, MEL, NSW); 22 miles west of Browne Range, Gun- barrel Highway, George 5408 (PERTH); 80 miles east of Meekatharra, Gardner 7874 (PERTH); 50 miles s. e. of Gascoyne Junction, Turner 5382c (TEX); 48 km. south of Mt. Magnet, Shaw 648 (ap); Cue, Andrews in 1903 (PERTH); 67 km. S. w. of Payne’s Find, Shaw 663 (AD). Stenopetalum anfractum is most easily recognized by the rigid stems which are repeatedly branched, producing a distinct zig-zag effect. At the point of branching the main stem is usually D-shaped in section, with definite ridges at the angles, a character which can be of help in distin- guishing S. anfractum from the rather similar and, in part, sympatric species, S. lineare. At one time I thought that this taxon was best treated as a subspecies of the Western Australian S. filifolium, so that material seen before 1968 was annotated as “S, filifolium ssp. anfractum,” a name never published. However, I now believe that it is better treated as a distinct species which is most closely related to S. lineare. 3. Stenopetalum filifolium Bentham, Flora Austral. 1: 78. 1863. Map 3 Hootypr. Western Australia. Swan River, Drummond s.n. (K). Plants slender annuals, glabrous or sparsely pubescent; trichomes when present, usually at the base of the stems, simple, short and more or less crinkled; stems to 40 or 50 cm. high, erect and often very fine, some- times quite elaborately branched; basal leaves unknown; cauline leaves 1-4(7) cm. long, 1-3(5) mm. wide, rather few and scattered, entire and often filiform, rarely remotely dentate; inflorescences loose, the flowering pedicels very short and the flowers apparently subsessile, buds ellipsoid; sepals 3.54.5 mm. long, elliptic to oblong, the lateral saccate; petals 10- 15 mm. long, bright yellow-orange, the claw linear and expanded into an oblong blade, this drawn out into an acuminate tip; stamens 2.5-3.5 mm. long; glandular tissue forming a narrow shelf on the inner side of each single stamen; infructescences usually loose, but variable; pedicels 64 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Map 2. Stenopetalum nutans. Map 3. Stenopetalum velutinum (dots); S. filifolium (triangles). 3-4 mm. long, stout and usually spreading from the axis; siliques 4.5—6 mm. long, ellipsoid and slightly compressed or terete; stigmas bilobed; ovules 6-10 per locule; well developed seeds not seen. 27 = 5 II + I DIsTRIBUTION, Western Australia, in the districts of Austin, Coolgardie, Irwin, Avon, and Darling. Western Australia. Lake Violet Station, 75 miles e. of Wiluna, Bennett 71 (PERTH); 8 miles south of Menzies, George 2724 (PERTH); Pioneer, Davies 269 (PERTH); 5 miles west of Meekatharra, Speck 579 (CANB, PERTH); 107 miles north of Mullewa, Turner 5367 (TEx); Yalgoo, Gardner 7756 (PERTH); 18 miles s. w. of Warriedar H. S., Shaw 666 (ap); Merridin, Koch 2845 (x, Nsw); 77 1972] SHAW, REVISION OF STENOPETALUM 65 miles north of Geraldton, Shaw 617 (ap); Northampton, Maiden in 1909 (Nsw); near Smith’s Mill, 20 miles e. of Perth, Andrews in 1902 (Nsw). Stenopetalum filifolium is an attractive little plant which seems to be quite widely spread in the southern half of Western Australia. The known range overlaps, in part, those of several other species of Stenopetalum, but the only one with which this species might be confused in S. anfractum. Stenopetalum anfractum has leaves usually at least trisect and petals of a dull brownish or purplish color, while S. filifolivm has narrow entire leaves and petals which are a bright, clear orange-yellow. I have col- lected S. filifolium near the Murchison River Crossing on the North- western Coastal Highway and at several places southwest of Mt. Magnet. In the latter area it often occurs with S. anfractum, but the two species could always be easily distinguished. | | ae ‘ t s ai \ \ 5 et qa , es r Map 4. Stenopetalum anfractum. Map 5. Stenopetalum decipiens (dots); S. iadentae, (triangles). 4. Stenopetalum sphaerocarpum F. Muell. Hook. Lond. Jour. Bot. 8: - 1850. Map 5. 66 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 LectotyPe. South Australia. Towards the entrance of the Murray River, F, Mueller in 1848 (mE 10847) Plants slender annuals, glabrous or sparsely papillose; stems 1-3 dm. long, often straggling, and wiry and leafless when in fruit; basal leaves to 5 cm. long, pinnatifid with 2 or 3 lobes per side, narrowing gradually to the slender petiole; cauline leaves 0.5—4 cm. long, few and scattered, entire, usually trifid with the central lobe slightly longer than the laterals, tapering to a slender petiole; inflorescences few-flowered and elongate, the buds ellipsoid; sepals 2—2.5 mm. long, oblong or obovate to elliptic, the lateral ones markedly saccate; petals 3.5-6 mm. long, 0.6—1 mm. wide, olive-green or purplish, the claw suddenly expanded into the round or elliptic blade, this drawn out into a slender tip; stamens 2—2.5 mm. long; glandular tissue poorly developed, consisting of 1 or 2 small lobes on either side of each single stamen; infructescences loose, the pedicels 3—7 (11) mm. long, stout and rigid, horizontal or slightly recurved and the siliques often pendent; siliques 3—-5.5 mm. long, globose or broadly obovate, slightly depressed at the base of the style; stigmas depressed and capitate; ovules (4)6—-10 per locule; seeds 1.5—2 mm. long, oblong to ellipsoid and flattened, yellow- or orange-brown, the testa smooth. » = a, an 10. DISTRIBUTION. From southern New South Wales and northwestern Victoria across southern South Australia into Western Australia. South Wales. 120 miles south of Menindee, G. J. White in 1963 (Nsw). Victoria. Lendrook Plain, Kulkyne National Park, Beauglehole 5283 (MEL); about 10 miles east of Kiamal in the Ouyen district, T. B. Muir 1106 (MEL); Underbool, Malone in 1915 (a, NSw); vicinity of Rainbow, C. S. Sutton 1633 (MEL); near Dimboola, F. M. Reader in 1898 (MEL). South Australia. Cane- grass Station, nd in 1937 (Ap); Kringin Reserve, ca. 32 km. north of Pinnaroo, Wilson in 1961 (ap); Karoonda, Winkler in 1934 (ap); 5 km. north of junction of Chauncey’ s Line and Monarto South road, Eichler 15097 (ap); Crystal Brook, Mueller in 1851 (mex); Ardrossan, Tepper 292 (AD, MEL); Port Lincoln, Muel- ler [?] (EL); Cunyarie Hills, Wilson 1884 (ap); Wudinna, /sing in 1938 (AD) ; 8 miles north of Kingoonya, "Beauglehole in 1968 (AD); Koonibba, Cleland in 1928 (aD); Ooldea Soak, Cleland in 1939 (aD). Western Australia. Mt. Magnet, Fitzgerald in 1903 (Nsw); Lake Monger, southwest of Warriedar H. S., Shaw 664 (A, AD); Morawa, Gardner 7518 (PERTH); Swan River, Drummond 687 (BM, E, K, MEL, P, W); basaltic ridges north of Stirlings Range, Mueller in 1851 (MEL); Boulder, Staer in 1905 (£); Camp 54, Elder Exploring Expedition [80 miles s. Victoria Springs], Helms in 1891 (aD, MEL). Stenopetalum sphaerocarpum is morphologically uniform throughout the range, the small flowers and globose siliques on rigid horizontally spreading pedicels at once distinguishing this from other species of Stenopetalum. The flowers are the smallest of any species in this genus, for the petals although always drawn out into the characteristic elongated tip are no more than 6 mm, long. This species and S. lineare are the only species of Stenopetalum known from Victoria and southern South Australia. 1972] SHAW, REVISION OF STENOPETALUM 67 5. Stenopetalum nutans F. Muell. Fragm. Phytogr. Austral. 3: 27. 1862. Map 2. Ho.otyPe. Queensland. Between Stokes Range and Coopers Creek, F. Wheeler (met 10789). Plants annual or, under favorable conditions, biennial, sparsely pubes- cent; trichomes malpighiaceous; stems usually to 3 dm., exceptionally to 5 dm. high, single to several, erect or decumbent and often much branched; basal leaves to about 6 cm. long, entire to remotely dentate or pinnatifid, narrowing to a slender petiole about as long as the blade; cauline leaves 3-9 cm. long, 4-12 mm. wide, linear to narrowly elliptic, entire or dentate toward the base; inflorescences loose and few-flowered, the buds obovoid or ellipsoid and purplish at the tip; petals 15-40 mm. long, the blade 0.8- 1.5 mm. wide, yellowish green or brown, the claw 2 to 3 mm. long, ex- panded into an elliptic blade drawn out into a long, slender and often twisted tip; stamens 3—5.5 mm. long; glandular tissue poorly developed, consisting of a bilobed piece of tissue on the inner side of each single stamen; infructescences loose and often secund, the pedicels 7-12(17) mm. long, at maturity recurved and the siliques nutant, occasionally hori- zontally spreading; siliques 6-10(13) mm. long, obovoid and sometimes slightly compressed; stigmas capitate and depressed; ovules 6-12 per locule; seeds 1.5-2 mm. long, reniform, orange-brown, the testa tubercu- late or verrucose. DistRIBUTION. Stenopetalum nutans is widely spread across central Australia, from Roma, Queensland, in the east to the Hamersley Range in Western Australia. It is common in the southern half of the Northern Territory and in the adjacent part of South Australia, especially to the northwest of Lake Eyre. The area encompassed by the scattered records from Western Australia suggests that S. mutans is widespread in the Eremaean of that state. Queensland. Roma, anon. (MEL); Charleville, Clemens in 1945 (srt); “Gil- - Tenham Station, 25 miles s.s.e. South Australia. Abminga, Jsing in 1933 (aD); 11 km. southeast of new Lam- 22745 (MEL); Kenmore Park, Cleland in 1950 (aD); 15 km. west of Everard Park homestead, Eichler 17422 (ap); N. Musgrave Ranges, Basedow 257 (AD, K, Nsw); Piltardi, Wilson 2536 (AD). Northern Territory. 3 miles west of Arga- dargada H. S., Swinbourne 570 (NT); Tarlton Downs, Mathew in 1955 (aD, CANB, BRI, MEL, NSw, NT); 3 km. north of Charlotte Waters, Shaw 482 (ap); 17 miles east of Harts Range Police Depot, Lazarides 5207 (aD, CANB, MEL, NSW, NT, PERTH); 5 miles south Taylor River Crossing, Stuart Highway, Nelson 1250 AD, NT); 40 miles south Aileron Township, Perry 5366 (AD, BRI, CANB, MEL, NSW, 68 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Stenopetalum nutans, with large pendent siliques and long green petals, these with twisted tips, is the most distinctive species of the genus, im- mediately attracting one’s attention in the field. 6. Stenopetalum robustum Endl. Enum. PI. Hugel. 4. 1837. Plants annuals, papillose or subglabrous, variable in size and habit; stems (of fruiting plants) 3 cm. to 1 m, long, short and robust or loose, straggling, and much branched; basal leaves to 4(8) cm. long, entire to deeply pinnatisect, usually lost early; cauline leaves to 3(6) cm. long, few and scattered, entire and linear or oblanceolate or coarsely lobed or dentate; inflorescences few-flowered and (in southern plants) often sub- umbellate, toward the northern part of the range, laxer; sepals 2.5—5 mm. long, elliptic-oblong to obovate, the laterals markedly saccate, all gibbous at the tip; petals 6.5-20 mm. long, variable (see discussion); stamens 2.5—5 mm. long, the anthers sometimes with a small appendage at the base of each cell; glandular tissue white to reddish, consisting of a pro- tuberant lobe on each side of each single stamen; infructescences loose; pedicels 2-4(6) mm. long, usually recurved [7-20 mm. long in var. pedicellare|; siliques (4.5)5.5-9 mm. long, elliptic to suborbicular or obovoid and compressed; stigmas obscurely bilobed; ovules 6-12 per loc- ule; seeds about 2 mm. long, oblong to suborbicular and flattened, reddish- brown. KEY TO THE VARIETIES Pedicels 2-4 mm., exceptionally to 6 mm. Me var. robustum. Pedicels 7-20 mm. long, and very slender.........__.. var. pedicellare. 6a. Stenopetalum robustum var. robustum. Map 1. Horotype. Western Australia. King George’s Sound, Hiigel (w). S. croceum Bunge in Lehmann, Plantae Preissianae 1: 258. 1844-45. Holo- type. Western Australia. “In arenosis sylvae haud longe ab oppidulo Perth,” Preiss 1939 (Pp), isotypes (MEL, P, w). S. gracile Bunge, Lehmann, Plantae Preissianae 1: 257. 1844-45. Holotype. Western Australia. “In arenosis umbrosis vallis haud longe ab ora mari- tima Perth,” Preiss 1938 (p), isotypes (MEL, P, w). S. robustum var. gracile (Bunge) Ostenf. Kgl. Danske Vidensk. Selsk. Biol. Medd, 3(2): 65. 1921. ‘S. minus Bunge, Lehmann, Plantae Preissianae 1: 258. 1844-45. Holotype. 1972] SHAW, REVISION OF STENOPETALUM 69 Western Australia. “Ad Princess Royal Harbour,” Preiss 1936 (P), isotypes (MEL). S. album E. Pritzel, Feddes Repert. 10: 133. 1911. Holotype. Western Aus- tralia. “In partibus australibus prope oppidulum Busselton, haud procul ab ora marina,” Max Koch 1969 (Bp), isotypes (MEL, NSW). Western Australia. Murchison’s River, anon, (MEL); Yandanooka, A. Morrison in 1904 (£); Nungarin, Stoward 343 (pm); Moora, Ashby 159 (AD, PERTH); 2 miles east of Forrestdale, George 3039 (PERTH); 4 km. west of the Mandurah Road on the road to White Hill, Shaw 710 (ap); Bunbury, E. Ashby in 1930 AD, BM); Busselton, Lucas in 1928 (Nsw); on limestone outcrops below the Cape Naturaliste lighthouse, Shaw 702 (aD); 1.5 km. east of Hamelin Bay, Shaw 688 (aD); (Lake Muir), Tourbay [Torbay, near Albany], Mueller (MEL). Stenopetalum robustum is, as the synonymy suggests, the most variable species of the genus. The most conspicuous variation, that responsible for the several synonyms, is in the petals — in their color, in the absolute length, and in the relative lengths of the blade and its acuminate tip. Plants from the southernmost localities, such as the type of S. robustum from King George Sound, and those in the vicinity of Cape Leeuwin and Augusta have white or cream-colored petals with short and rather broad tips which are about 1 to 3 times the blade in length, or have petals which show no clear demarkation between blade and tip and are trullate. At the other extreme are the populations occurring northward from the vicinity of Mandurah (about 50 miles south of Perth). These plants have yellow to deep orange petals with very long and slender tips which may be 10 times or more the blade in length. Between the extremes occur all intermediates. A sample taken near Cape Naturaliste included plants with the short white petals of the type, plants with longer (but white) petals, plants with longer petals (only the tips orange or yellow), and some with petals entirely yellow or orange, but the tips not so long as those seen on plants growing north of Mandurah. Plants from the southern part of the range — those with short white petals — tend to be shorter and more robust, and to have denser in- florescences, and this is generally true of those in coastal areas. Those growing inland and north of Perth, are usually large and more branched. Some plants in a collection made near Yanchep (Shaw 529, an, GH) were 1 meter in height and very loose and straggling. Also noticeable in pro- ceeding from Cape Leeuwin northward is a gradual increase in laxness of inflorescences and infructescences. Plants with compact, but slightly elongated inflorescences, look remarkably like flowering plants of Stack- housia. The holotype of S. robustum, collected by Hiigel at King George Sound (Albany, on the south coast of Western Australia) has short white petals, described by Endlicher as “Petala sepalis duplo longiora, . . . ” that is, about 2 1/2 lines long. The next collections, those of Preiss, were made at Princess Royal Harbour, the western arm of King George Sound (Preiss 1936), and near Perth (Preiss 1938, 1939). From these three collections Bunge described three species of Stenopetalum. That from Princess Royal 70 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Harbour is S. — described as having “‘petalis sepala duplo pe Sabah (croceis) . and ‘Caules . florentes 3-4 pollicares.” This is rather see uri material of the fori collected by Hiigel and the ref- erence to the petals as yellow may have been a misinterpretation of dried material. Oddly enough, Bunge made no outright comparison of this “species” with S. robustum and he may not have realized that the lo- calities are essentially the same. The two collections made near Perth be- came the types of S. gracile and S. croceum. The nye is based on white-flowered xo the petals described as “ . sepala 3plo super- antibus (albis), . . . ” and as 6 lines in length. Bo ecaerns croceum Bunge deserihed as having petals “ . . . sepala quadroplo superantibus (croceis) . . . ” and 8 lines long with a long slender tip. Comments in the protologues of these three do show that Bunge understood that the plants before him were closely related to Endlicher’s species and Bunge cannot be faulted for his failure to realize that the isolated collections from Perth and Albany do belong to a single species. A fifth name be- longing here is S. album E. Pritzel, based on a collection from Busselton on Geographe Bay. The type is quite ordinary white-flowered material much like the type of S. robustum. 6b. Stenopetalum udneaany var. pedicellare (F. Muell. ex oe E. Shaw, comb. no Map 1 S. pedicellare F. Muell. ex Bentham, Flora Austral. 1: 79. 1863. Holotype. Western Australia. Murchison River, Oldfield (x). Western Australia. 45 miles s.e. of Carnarvon, Beauglehole 11754 (MEL); be- tween Carnarvon and Shark Bay, Berston in 1962 (PERTH); 103 miles s.e. of Carnarvon, Beauglehole 11783 (MEL); Lake Violet, Speck 1301 (aD, CANB, PERTH); Twin Peaks Station, Royce 5721 (PERTH); Champion Bay [Gerald- ton], Gray 54 (MeL); Yandanooka, Morrison in 1904 (£); Glenorn Station, Mal- colm, Burbidge in 1938 (PERTH). Although described by Bentham as a distinct species, these slender plants differ from the northern form of var. robustum only by the more straggling habit and very long pedicels. The petals, described by Ben- tham as “apparently yellow” are a bright orange-yellow, much as in S. filifolium. 7. Stenopetalum decipiens E. Shaw, sp. nov. FicureE 2, Map 5. Plantae perennes, suffruticosae, saepe ramosissimae, dense pubescentes trichomatibus ramosis, his in tomento coacto implexis; petala liguliformes hepatica vel olivacea; infructescentiae laxae, pedicellis divaricatis; siliquae ellipsoidea vel anguste obovatae; ovulae 5—10 in loculo; semina laeves por- phyrea vel obscure aurantia. 2” = 8. Hoxotyre. Northern Territory. 44 miles w.n.w. of Hartz Range Po- lice Station, Lazarides 5889 (Ap), isotypes (BRI, CANB, MEL, NSW, NT). Ficure 2. Stenopetalum decipiens, Habit, X 1/4; detail of pubescence, X 2. 72 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 53 Plants perennial, usually straggling subshrubs, densely pubescent; trichomes sessile or subsessile, much and irregularly branched, the rays in- terlocked into a dense felt-like tomentum, this on older plants often abraded by wind and sand; stems to at least 70 cm. long, woody and often much branched; basal leaves none; cauline leaves 1—5 cm. long, more or less linear (2—3 mm. wide), entire or finely and remotely dentate, rounded at the tip, inflorescences loose even before anthesis, the buds large and ellipsoid; sepals 3.5—4.5 mm. long, oblong to elliptic, the lateral sepals saccate, all with a gibbous thickening at the tip; petals 8.5-13 mm. long, chocolate-brown or purplish-brown to olive-green, the claw linear and expanded into a ligulate blade, this elongate but not acuminate; stamens 3.5—5 mm. long; glandular tissue roughly U-shaped on the inside of each single stamen; infructescences loose; pedicels (3.5)4.5-9 mm. long, usually divaricately spreading; siliques 7-10.5 mm. long, ellipsoid to narrowly obovate, terete to slightly compressed; stigmas small and capitate; ovules 5—10 per locule; seeds ca. 1.5 mm. long, oblong to oval and slightly flattened, the testa red-brown to dull orange and smooth. DisTRIBUTION. Common in the central Australian ranges, especially in the Harts and George Gill ranges; also in the Barrow Range in Western Australia and in central Queensland. Occasionally on sandy soils but most often on stony hillsides and in crevices on rock faces and ledges. Queensland. 70 miles north of Longreach, Davidson in 1953 (BRIU); near Eromanga, S. T. Blake 11894 (srt). Northern Territory. Tobermory, Nathans Hole paddock, Chippendale in 1954 (Nsw, NT); Harts Range, 6 miles south of the Police Station, Chippendale in 1957 (AD, CANB, NSW, NT, PERTH); sand plain near Barrow Creek, Gauba in 1950 (w); Mt. Undoolya, 18 miles east of Alice Springs, Chippendale in 1956 (BRI, CANB, NT); Emily Gap, 9 miles e.s.e. Alice Springs, Nelson 1814 (GH, NT); 1/2 mile Ss. Dope is Rockhole, Ooraminna Range, J. Must 481 (GH, NT); Yuendumu, Cleland in 1951 (a, ap); Haast’s Bluff Reserve, Cleland in 1957 (ap); George Gill Range and Kathlenn Spring, Beauglehole 26961 (ap); s.e. side of Mt. Connor, Beauglehole 22838 (ap). West- ern Australia. Mt. Squires, Barrow Range, Helms in 1891 (Ap, Nsw). Stenopetalum decipiens owes the epithet (‘deceiving’) to its similarity to, and frequent confusion with, S. velutinum, a species with which it is in part sympatric, and to which name it keys in floristic works dealing with central Australia. Under central Australian conditions, S. velutinum seems always to be annual, but S. decipiens is clearly perennial and very woody at the base. Plants of the latter (from seed of Nelson 1814) grown in the greenhouses of Harvard University were, at three months, clearly woody. Both species are densely pubescent with irregularly branched trichomes, but those of S. decipiens are interwoven into a felted tomentum and, with a little care, strips of this tomentum four or five millimeters in length can be peeled from the stems. These interlocked trichomes are especially sub- ject to the abrasive action of wind-blown sand and as a result of such abrasion and excoriation of the thin “‘bark’”’ mature plants may appear to be quite glabrous. Remnants of the indumentum usually can be found in 1972] SHAW, REVISION OF STENOPETALUM 73 leaf axils and stem axils. Stenopetalum decipiens has large siliques, on longer pedicels, than does S. velutinum, and smooth rather than tubercu- late seeds. In both growth form and habitat, S. decipiens is much like Arabidella filifolia, also a straggling perennial, which often is found on rock faces in the Flinders Ranges of South Australia (Shaw, 1965). 8. Stenopetalum velutinum F. Muell. Plants Indig. Col. Vict. 1: 49. 1862. Map 3. HototypPe. Victoria or South Australia. Murray Desert, Mueller [?| (MEL 10866). Plants erect annuals, densely pubescent; trichomes sessile or subsessile, very irregularly branched; stems 1—4(6.5) dm. long, usually branched near the base; basal leaves to ca. 7 cm. long, lanceolate, entire or re- motely dentate, withering early and usually lost from fruiting plants; cauline leaves 1-7 cm. long, 1-9 mm. wide, lanceolate to linear, entire or remotely dentate, sessile or shortly petiolate; inflorescences initially dense but elongating at anthesis, buds ellipsoid; sepals 3.5—5.5 mm. long, ob- long or elliptic, the lateral saccate, the median cucullate with a callose thickening at the tip; petals 6-21 mm. long, 0.7—1.2 mm. wide, yellow- green to brown, the claw linear and expanded into an elliptic blade, this usually extended into an elongate tip; stamens 3-6 mm. long; glandular tissue a crescentic or rhomboid lobe on each side of each single stamen; infructescences loose, to 20 cm. long; pedicels 2-10 mm. long, slightly spreading or erect; siliques 4.5-6.5(8) mm. long, subglobose to oblong or obovoid; stigmas depressed-capitate and rather fleshy; ovules 10-20 per locule; seeds 1-1.5 mm. long, oblong or elliptic and flattened, orange- or red-brown and tuberculate. n = 4, 2n = 8 Distr1BuTIoN. Widely spread throughout central Australia, also in western Victoria, and in southwestern New South Wales. Although a com- mon species, S. velutinum seems never to occur in great abundance. Queensland. 20 miles west of Miles, S. T. Blake 5905 (prt); Armadilla, be- tween the Warrego and Maranoa, Barton 231 (MEL, NSw); between Stokes Boorman in 1903 (sm, Nsw); Wanganella, Officer in 1903 (ysw). Victoria. Nhill, St. Eloy d’Alton 8 (met). South Australia. Pedirka, Zsing in 1932 (aD); 74 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Stenopetalum velutinum sometimes is confused with S. lineare, the re- sult of that species having been described by De Candolle as glabrous with a subsequent misapplication of the name S. lineare var. canescens, as in Black’s Flora of South Australia, to the inland form of S. lineare. Steno- petalum velutinum has all parts of the plants, even the axes of the in- florescences and the sepals, densely pubescent with appressed trichomes, while plants of S. lineare, although often quite densely pubescent towards the base, have the upper parts of the stems glabrous or nearly so. Most closely related to S. velutinum and sympatric with it is S. decipiens which can be distinguished by the perennial habit and very dense felt-like to- mentum. EXCLUDED SPECIES STENOPETALUM DRABOIDES Hooker, Icones Plantarum ¢. 617. 1844. Holotype. Western Australia. =e! smb Drummond (“Crucif. n. 3”) (kK) = Menkea draboides (Hook.) Bent STENOPETALUM INCISIFOLIUM ae f., Icones Plantarum ¢. 276. 1840. Holo- type. Tasmania. “Hab. Blackman’s River, on the Road to Hobart Town, Van Dieman’s Land,’ Gunn 644 (xk), isotype (GH) = Hutchinsia pro- cumbens (L.) Desv. STENOPETALUM PROCUMBENS Hooker, Icones Plantarum ?¢. 610. 1844. Hole estern Australia. Swan River, Drummond 3 (K) = Menkea a hm e STENOPETALUM TRISECTUM Tate, Handb. Flora Extratrop. S. Austral. 17 [in sched.]. 1890. Holotype. South Australia. Innamincka, McLeod in 1884 MEL). The type seems to be a diseased specimen of S. lineare LITERATURE CITED BurpincE, N. T. The phytogeography of the Australian region. Austral. Jour. Bot. 8: 75-211. 1960. CLELAND, J. B., & T. H. JOHNSTON. Notes on native Care vet uses of plants in the Musgrave Ranges region. Oceania 8: 208-215. & N. B. TrnDALE. Ecological surroundings of the N oem natives in Central Australia and native names and uses of plants. Trans. Roy. Soc. S. Austral. 77: 81-86. 1954. The native names and uses of plants at Haast Bluff, Central Australia. Ibid. 82: 123-140. 1959. Crocker, R. L., oop. Some historical influences on the development of the South Australian vegetation communities and their bearing on con- cepts and classification in ecology. Trans. Roy. Soc. S. Austral. 71: 91-136. 1947. Dietz, R. S., & J. C. Hotpen. The breakup of Pangaea. Scientific American 223(4): 30-41. 1970. Jounston, T. H., & J. B. CLetanp. Aboriginal names and uses of plants in the Ooldea Region, South Australia. Trans. Roy. Soc. S. Austral. 66: 93- 103. 1942. PRANTL, K. Cruciferae. Nat. Pflanzenfam. III. 2: 145-206. 1891. 1972] SHAW, REVISION OF STENOPETALUM 75 Scuutz, O. E. Cruciferae. Nat. Pflanzenfam. ed. 2. 17b: 227-658. 1936. Suaw, E. A. A critical revision of some endemic Australian genera of Cruciferae. Trans. Roy. Soc. S. Austral. 89: 145-253. 1965. TurNeER, B. L. Jn IOPB chromosome number reports IX. Taxon 16: 62-66. 1967. ADDENDUM (see comment under S. lineare): In Margaret Willis’ By Their Fruits (Sydney, 1949) there is a photograph of a specimen of Helichrysum collected by Mueller on Lefevre Peninsula on 16 December, 1847, the day after he and his sisters landed. The sheet was an- notated by James H. Willis of the National Herbarium of Victoria with the suggestion that this may have been Mueller’s first botanical specimen from Australia. However, the comment on the label of the Stenopetalum leaves, I think, no doubt that it was crucifer rather than composite which first attracted the Baron’s botanical attention. Gray HERBARIUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 76 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 53 A REVISION OF THE NEOTROPICAL GENUS LISIANTHIUS (GENTIANACEAE) RicHarp E. WEAvER, JR.1 THE GENUs Lisianthius as here interpreted comprises 27 species and 2 varieties of woody or semiwoody Gentianaceae distributed throughout the Greater Antilles and continental America, from south-central Mexico to northwestern Colombia. Although the plants are frequently found on road- side banks, and the flowers are usually large and showy, the genus has been rather poorly collected. in the field, at which time seeds and preserved material for cytological and anatomical investigation were obtained, and 10 species were grown in the greenhouse. HISTORY OF THE GENUS , who kindly loaned the specimens es- sential to my work: a, BM, DS, DUKE, F, GH, HAL, IJ, LL, MEXU, MICH, MO, NY, P, 1972] WEAVER, REVISION OF LISIANTHIUS ee Browne (1756). He described two species, both from Jamaica: Lisianthius erectus, foliis lanceolatis floribus singularibus terminalibus; and Lisianthius foliis cordato-acuminatis, petiolis brevibus, floribus terminalibus quandoque geminatis. Browne placed the generic description following that of the first species, as was his custom (W. T. Stearn, in litt.). Linnaeus (1767), in taking up Browne’s genus, changed the spelling of the name to “Lisian- thus.” He assigned a binomial, L. longifolius, to Browne’s first species, and L. cordifolius, to the second. Linnaeus’s spelling of the name has been the one most commonly used. In fact, Rothmaler (Fedde, Repert. Sp. Nov. 53: 178. 1944.) proposed conservation of Lisianthus L. over Lisian- thius P. Browne. However, the proposal was rejected (Taxon 3: 242. 1954). Only Kuntze (1891) and more recently Weaver (1969, 1970) and Robyns and Nilsson (1970) have used the correct and original spell- ing “Lisianthius.” Aublet (1775) described four species from French Guiana, using the name Lisyanthus. Aublet’s plants are allied to but not congeneric with Lisianthius P. Browne, according to the present interpretation. It is dif- ficult to believe that Aublet meant to establish a new genus, in view of the close relationship of his plants to Lisianthius P. Browne and the nearly identical spelling of the names. Yet, according to W. T. Stearn (in litt.), when Aublet (doc. cit.) furnished a detailed generic description, as he did in the case of Lisyanthus, his intent was to establish a new genus. For previously established genera, he merely cited the generic name. If Lisyanthus Aubl. is indeed to be interpreted as a genus distinct from Lisianthius P. Browne, under the present International Code the name must be treated as an orthographic variant of Lisianthius P. Browne and therefore a later homonym (Art. 75) to be rejected (Art. 64). The younger Linnaeus (1781) added two species to his father’s Listan- thus: L. chelonoides and L. glaber, neither of which are referable to Lisianthius P. Browne. He cited “Lisianthus alatus Aubl.” as a synonym of L. chelonoides, evidently considering Lisyanthus Aubl. as merely an orthographic variant of Lisianthus Sprengel (1825) included, in what he called “Lisianthus P. Browne,” species described by Linnaeus (1767), Aublet (1775), Linnaeus filius (1781), Swartz (1788), Kunth (1818), and others. Most of these are not referable to Lisianthius P. Browne as interpreted here. Grisebach (1839) gave to Lisianthius P. Browne the new but nomen- claturally superfluous name Leianthus and divided the genus into two sections, LisiaNtHIus (P. Browne) Griseb. and OmpHatosticmA. In ad- dition ‘ep accepted Lisyanthus Aubl., which he divided into five sections: MACROCARPAEA, CHORIOPHYLLUM, CHELONANTHUS, Heia_ (Mart.) Griseb., and Carotisvanrave: Later Grisebach (1845), while maintain- ing Leianthus, elevated Leianthus sect. OMPHALOSTIGMA to generic rank as Petasostylis. At the same time he used the name Lisianthus, attributed to Aublet (1775), for his earlier Lisyanthus, but left the content of the genus essentially unchanged. Bentham (1876) maintained Leianthus in the sense of Grisebach 78 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 (1839), including Petasostylis as a synonym, and Lisianthus, which he attributed to Aublet (1775). While accepting the sections of Grisebach (loc. cit.) he added three new sections to the latter genus: SY MBOLANTHUS (G. Don) Benth. & Hook. including Letothamnus Griseb., ADENOLISIAN- THUS (Progel) Benth. & Hook., and Irtpacuia (Mart.) Benth. & Hook. Kuntze (1891) correctly applied Lisianthius P. Browne and treated Lisianthus L. and Leianthus Griseb. as synonyms. Doubting that Aublet (1775) meant Lisyanthus to be a new genus, Kuntze adopted the name Helia Mart. for Lisyanthus (Lisianthus) Aubl. sensu Grisebach (1839, 1845) and Bentham (1876). Gilg (1895), whose system has been employed by most recent authors, continued the taxonomic separation made by his immediate predecessors, but used the name Lisianthus L. for the plants here referred to Lisian- thius P. Browne. Gilg employed a rather narrow generic concept based for the most part on pollen morphology. He elevated to generic status all of the groups which Bentham (1876) had previously treated as sections of ‘“‘Lisianthus Aubl.” INTERGENERIC RELATIONSHIPS In Gilg’s (1895) treatment of the family Lisianthius (as Lisianthus) was associated with seven other genera, most of them small, and all but one, Tachiadenus Griseb., restricted to the New World, in the tribe Gentianeae subtribe Tachiinae. Gilg’s fundamental taxonomy of the family was based almost entirely on the morphology of the pollen grain. As previously pointed out (Weaver, 1969), Gilg’s treatment is unsat- isfactory in a number of respects. Anatomically the Tachiinae resemble the plants in Gilg’s tribe Helieae more than they do those in other subtribes of the Gentianeae. Perrot (1898) reported that only the Helieae and Gentianeae-Tachiinae are characterized by having mechanical elements in the cortical area of the stem. Lindsey (1937, 1940) reported that the anatomical features of the Helieae are strikingly similar to those of the Tachiinae, but are more advanced. The two groups of genera are characterized by the nearly con- stant occurrence of glandular tissue at the base of the ovary, a character found only sporadically in other groups of the family. Lisianthius is similar to the Helieae in having fused calyx-lateral traces, found in none of the other generic groups. Macrocarpaea glabra (L.f.) Gilg and Zonanthus cubensis Griseb., both placed by Gilg in the Tachiinae, do not have fused calyx-laterals, but in other respects the flowers of these species are anatomically identical to those of the Helieae (Lindsey, 1937). Lindsey, in fact suggested, that although the Helieae and the Tachiinae are suf- ficiently different to be placed in distinct groups ‘‘of some grade,” the similarities are erg that the Helieae be best reduced to subtribal status in the Gentia According me Gilg (1895) the Helieae are unique in the family in pos- sessing pollen that is held in tetrads or polyads. Since Macrocarpaea 1972] WEAVER, REVISION OF LISIANTHIUS 79 glabra and Zonanthus cubensis have single pollen grains but are anatomical- ly nearly identical to Gilg’s Helieae, the value of pollen in tetrads as a character for delimiting a group as large as a tribe seems questionable here. As explained earlier, Lisianthius has been broadly construed in the past, and the name, usually as Lisianthus or Lisyanthus, has been applied to a rather diverse group of plants. Nilsson (1970) recently surveyed the pollen morphology of these lisianthioid gentians, and his results are at variance with those of Gilg. He reported that, while the pollen types in Lisianthius (as delimited in this study), Macrocarpaea (Griseb.) Gilg, Rusbyanthus Gilg, and several anomalous species are distinctive, the pollen grains of Adenolisianthus (Prog.) Gilg, Calolisianthus (Griseb.) Gilg, Chelonanthus (Griseb.) Gilg, Helia Mart., Irlbachia Mart., Lehma- niella Gilg, Pagaea Griseb., Purdieanthus Gilg, and Symbolanthus G. Don all have approximately “the same basic morphology and some even have features which are transitional within the range of variation of pollen types in the genera.” The taxonomy of these lisianthioid gentians is in a state of chaos. Macrocarpaea is the only genus of any size monographed during this cen- tury. The genus, as treated by Ewan (1948) is apparently heterogeneous and hence unnatural. Nilsson (1968) found three pollen types in Macro- carpaea: a distinctive one; a type identical to that of Rusbyanthus, a monotypic genus placed by Gilg in a separate tribe between the Gen- tianeae and the Helieae; and a final type, with pollen in tetrads, similar to that found in Chelonanthus, a genus placed by Gilg in the Helieae. Those species of Macrocarpaea with Rusbyanthus-type pollen do not deviate significantly in other respects from others in the genus. Neither for that matter does Rusbyanthus cinchoniifolius Gilg, and the two gen- era should probably be combined. Those species of Macrocarpaea with pollen in tetrads resemble Chelonanthus morphologically, and probably should be combined with it. The remaining lisianthioid genera, all placed by Gilg in the tribe He- lieae, are in general difficult to characterize morphologically. Lagenanthus Gilg, Lehmaniella, and Purdieanthus differ from the remaining genera in having tubular corollas, but, contrary to Ewan’s (1948b, 1952) findings, I have been unable to recognize reliable morphological criteria for sep- arating them. Helia is distinctive in having a salverform corolla, but in other respects it closely resembles Chelonanthus. Intergradation between Adenolisianthus, Calolisianthus, Chelonanthus, Irlbachia, and Pagaea completely obscures generic boundaries. Several species are anomalous in both pollen and gross morphology and appear not to fit into any of the existing genera. These include Iribachia caerulescens (Aubl.) Gilg, Calolisianthus imthurnianus (Oliver) Gleason, Lisianthus chimantensis Steyerm. & Maguire, and Calolisianthus frigidus. The last species, however, according to Nilsson (1970) has pollen ap- proaching that of Symbolanthus. Study of this entire group of gentians has been hampered not only by the pitifully small number of existing specimens but also by the 80 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 wretched condition of many of them. Cytological studies may prove of considerable taxonomic value as indicated by the few chromosome counts reported by Weaver (1969). At any rate these plants are in need of con- centrated study. Nilsson’s pollen work is a good starting point. However, the study, when and if it is made, must be a more comprehensive one than that of Gilg. Gilg’s work shows quite well the confusion that often re- sults from reliance on a single character in the classification of a large group of plants. A comprehensive study of the lisianthioid gentians is beyond the scope of this paper and no formal taxonomic decisions con- cerning them will be made at present. Instead, the following key has been prepared as an aid in distinguishing, wherever possible, the lisian- thioid gentians as they now stand (Irlbachia caerulescens, Lisianthus chi- mantensis, and Calolisianthus imthurnianus are not included; those species of Macrocarpaea with pollen in tetrads key out under C helonanthus). KEY TO THE GENERA OF LISIANTHIOID GENTIANS Oe ee ee a ge ee Lisianthius. Stigmas bifurcate, the lobes linear or flattened; ovary unilocular or bilocular, the placentae conspicuously inrolled, projecting strongly into the lumen of the ovary and nearly filling it; old placentae not visible on the mature cap- sules, the margins adjacent to the sutures smooth; corolla tube without bundles of metaxylary fibers; main stem usually 4-angled. 2. Pollen grains single; inflorescences branched systems of apparently sim- ple dichasia, the divisions subtended by conspicuous foliar bracts; fo- liage leaves with many pairs of principal lateral veins... <..... . ...cc¢-os043 oe en ee en ae ee acrocarpaea, Rusbyanthus. - Pollen grains in tetrads or polyads; inflorescence an apparently simple dichasium or much more commonly a compound dichasium, this once or — rf i) than 6 cm. long, the corolla funnelform. ............ Symbolanthus. Bases of the stamens connected by a carnose cylinder of tissue or not, but the cylinder never with triangular teeth; connectives of the anthers included or short-exserted; inflorescences long-pedunculate ; flowers usually campanulate, or rarely tubular, salverform, or funnel- f 4. Corolla campanulate, rarely salverform or funne ‘ 5. Corolla salverform; foliage leaves concentrated in the lower half of the stem. Helia. orm. 4. Corolla tubular. ...... Lagenanthus, Lehmaniella, Purdieanthus. lform 1972] WEAVER, REVISION OF LISIANTHIUS 81 5. Corolla campanulate, rarely funnelform; foliage leaves evenly distributed along the stem. 6. Leaves usually palminerved, sessile, coriaceous or sub- coriaceous, rarely penninerved; lobes of the stigma much broader than the style; inflorescence an apparently simple or a compound dichasium, the lateral branches never ap- pearing racemose, or often reduced to a single flower. .... ee oe eee ee Calolisianthus. Leaves, at least the lower ones, penninerved and petiolate, membranaceous or very rarely coriaceous; lobes of the stigma usually but barely broader than the style; inflores- cence a once or twice bifurcately branched dichasium, the ultimate branches usually secund or complanate and ap- pearing racemose, or very rarely the inflorescences reduced to a single, terminal flower. 7. Ultimate branches of the inflorescences equal, usually secund or complanate; style conspicuously longer than the ovary; anthers longer than broad; corolla often gibbous; pollen in tetrads or polyads; coarse annuals or percaninis: | 322: Chelonanthus, Adenolisianthus, Macrocarpaea arborea, M. cerronis, M. quelchii, M. salicifolia. . Ultimate branches of the inflorescences often unequal, seldom secund; style usually shorter than the ovary or equalling it; anthers usually about as broad as long; corolla never gibbous; pollen in polyads; annuals. .... Pagaea, Iribachia. ~sI Lisianthius is easily separated from the other lisianthioid genera by the characters outlined in the key. The capitate or peltate, indistinctly lobed stigmas, the weakly projecting placentae, the presence of floral fibers, as well as the characteristic pollen morphology mentioned earlier, clearly set it apart as a distinct genus. Therefore I would disagree with Steyer- mark (1953) who argued for the inclusion of several of the lisianthioid genera, e.g. Chelonanthus, Calolisianthus, Irlbachia, and Pagaea, within Lisianthius. MORPHOLOGY Habit. The species of Lisianthius have been described as annual or perennial herbs, shrubs, or trees. Lisianthius capitatus, for example, is re- ferred to as an herb by some collectors and a slender tree by others. Field observations, coupled with herbarium studies and observations of culti- vated plants have led me to conclude that most of the species are peren- nials, although some of them may be rather short-lived. Plants of sect. ListANTHTUS subsect. HERBACEI appear to be mostly perennial, suffrutescent herbs. Only the primary roots and the basal 1 dm. of stem are distinctly woody. In the field, plants usually have both a green flowering shoot and a dead shoot with old empty capsules present at the same time. Cultivated plants of L. migrescens die back to the base 82 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 after flowering and eventually new shoots are produced from the lower nodes. All plants of the subsection are monopodial. The branching is ex- current and generally opposite, and all of the shoots are determinate. The species of section OmpnHatosticmMa are similar in aspect and growth form to those of sect. ListaNTHrus subsect. HerBAce!; they are coarse monopodial herbs with opposite, excurrent branching. Lisianthius meianthus, however, is definitely an annual. The only population of this species seen in the field, in eastern Guatemala, was made up of old dead individuals and young vegetative ones. Living and dead shoots were never present on the same plant. The plants of sect. ListaNTHIUs subsect. FruTicosi are all perennials. Most species are subshrubs. Even though the aérial portions of these plants persist more than one season, they are not distinctly woody except at the base. Many of the species flower the first or second year from seed, when they are still predominantly herbaceous. Lisianthius exsertus, L. jefensis, L. skinneri, and sometimes L. peduncularis are true shrubs, while L. capitatus and L. umbellatus become slender trees with stems to 6 meters tall and 8 cm. in diameter. The plants of sect. Listanturus subsect. FrutTicosi are generally not monopodial; the branching is opposite or alternate and is not excurrent. These typically have a single main trunk which is branched above. How- ever, in most species, certain individuals may be entirely unbranched and spindly, resembling herbs. The main shoots are indeterminate, and the inflorescences are borne on determinate, lateral shoots. The genus may be divided into two artificial groups on the basis of habit. The species of sect. OMPHALOSTIGMA and sect. LISIANTHIUS sub- sect. HERBACEI are annual or perennial suffrutescent herbs with determinate main axes. The species of sect. LIsIANTHIUS subsect. FRUTICOSI are sub- shrubs, shrubs, or slender trees with indeterminate main axes. This dif- ference in habit is one of the primary characters used here to divide sect. LISIANTHIUS into two subsections. Unfortunately, this character is often difficult to interpret from herbarium specimens. Most species, even the herbaceous ones, are rather tall and coarse and since it may be impos- sible to fit an entire plant on an herbarium sheet, often only lateral branches are collected. These are determinate in both subsections of sect. LISIANTHIUs and may easily be mistaken for the main axes. Indument. Most species of Lisianthius lack an indument of any sort. However, a very fine pubescence is present in L. cordifolius and L. meian- thus, and in some collections of L. glandulosus, L. longifolius, and L. saponarioides. Although the pubescence may be rather dense, only in L. cordifolius and L. longifolius is it readily evident without magnification. The individual hairs (or excrescences) are unicellular, 1 mm. or (often considerably) less in length, and the pubescence is best described by the terms “puberulous” or “spiculate.” Stems. Decurrent from the leaf bases on opposite sides of the her- 1972] WEAVER, REVISION OF LISIANTHIUS 83 baceous portions of the stems are pairs of raised parallel ridges, which run the length of each internode and are usually inconspicuous without magnification. The ridges of each pair are generally close together, ap- pearing as one. In some species, however, the ridges of each pair are widely separated, the stem then appearing indistinctly 4-angled. The angles may be expanded into narrow, scarious wings, which in L. cuspida- tus are minutely serrulate. Leaves. The genus is also easily divided into two artificial groups on the basis of whether or not the foliage leaves are truly sessile. The species of sect. LislANTHIUsS subsect. HERBACEI are characterized by sessile, usu- ally clasping foliage leaves. In these species the lamina of the blade is confluent with the basal sheath, and the petiole is entirely lacking at least in the lower leaves; the upper leaves may be short-petiolate. In the species of sect. OMPHALOSTIGMA and subsect. Fruticosi of sect. LIsIANTHIUS, the leaves are narrowed to a petiole which may be winged and very short, causing the leaves to appear subsessile. The leaves on the determinate lateral branches of all species, as well as those leaves subtending the lateral branches in sect. OMPHALOSTIGMA and sect. ListaNTHIUS subsect. HERBACEI, are smaller and of different shape than the foliage leaves. When the foliage leaves are elliptic, obo- vate, or oblanceolate, the upper leaves and those on the lateral branches are ovate or lanceolate; when the foliage leaves are ovate or lanceolate, the upper leaves and those on the lateral branches are lanceolate or linear. In sect. OMPHALOSTIGMA and subsect. HERBACEI of sect. ListaNTHIUs the transition is gradual. In subsect. Fruticost of sect. LIsIANTHIUS the transition is abrupt; the difference between the two types of leaves is striking, especially in such species as L. capitatus, L. cordifolius, and L. umbellatus. The presence of two types of leaves on one plant makes the interpreta- tion of leaf size and shape rather difficult, especially when only a lateral branch has been preserved. In this study “foliage leaves” refers to those leaves on the main shoots of plants of sect. LIsIANTHIUS subsect. FRUTI- Cosi, or those leaves on the main shoot which do not subtend a lateral shoot in sect. OMPHALOSTIGMA and subsect. HERBACEI of sect. LISIAN- THIUs. The leaves of sect. ListaNTHIUS subsect. FRUTICOSI are evergreen, but Inflorescences. The inflorescences of Lisianthius are rather difficult to interpret, and as a result have been variously described as “cymose, 84 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 1972] WEAVER, REVISION OF LISIANTHIUS 85 “thyrsoid,” or “dichasial.” According to my interpretation, the inflores- cences in the genus are basically compound dichasia. There have been, however, many modifications on the basic scheme. The dichasia may be 1 to 5 times compound, or they may be reduced (1) to the point where they appear simple, or (2) to single flowers. The subunits of the dichasia may be 3-flowered, or one flower in each may abort, so that the lateral branches appear scorpioid. The various modifications of the dichasium, as found in Lisianthius, with a possible sequence of these modifications, are illustrated in Ficure 1 Ficure 1-A represents a normal (twice-) compound dichasium, the basic type found in the species of sect. OMPHALosTIGMA and most of those in sect. ListANtHIUs subsect. Herpacel. This type of dichasium is found in many gentianaceous genera, including Centaurium, Curtia, Sabatia, and Schultesia, and is probably the generalized one in the fam- ily. Each of the subunits is subtended by a pair of bracts, the lowermost often foliaceous, the upper ones becoming scarious. The lateral pedicels of each of the ultimate subunits are subtended by a pair of scarious brac- teoles, in the axils of which aborted buds are frequently found, indicating that the lateral flowers themselves represent reduced subunits. From Type 1-A, the modifications have probably proceeded in two directions. Ficure 1-C represents the type of modified compound di- chasium found typically in Lisianthius brevidentatus and not infrequently in the remaining species, except for L. cuspidatus, of sect. LIstlANTHIUs sub- sect. HERBACEI, as well as L. seemanii (sect. LISIANTHIUS subsect. FRu- TICosI). This type, in which the lateral branches appear scorpioid, is easily derived from Type 1-A, as shown in Ficure 1-B, by the loss of branches or flowers on one side (often but not always the same side) of each trichotomy. Although it may at first appear to be a simple dichasium, Ficure 1-E actually represents a reduced compound type. The lateral axes are di- chasial branches terminated by a solitary, pedicellate flower. T hat they are not merely simple, bracteolate pedicels is borne out by two observa- tions: (1) aborted buds are frequently found in the axils of the brac- teoles; and (2) the morphology of the axis is different above and below the bracteoles in that the raised parallel ridges mentioned before are absent above them. Type 1-E is the basic type of dichasium found in most of the species of sect. LisiANTHIUs subsect. Fruticos. In addition this type is occasionally found in the other species of sect. ListANTHIUS subsect. HERBACEI and regularly in L. cuspidatus. It is easily derived from Type 1-A, as shown in Ficure 1-D, by loss of the upper subunits. In this study, Type 1-E will be referred to as an “apparently simple” dichasium. Ficure 1-G represents the ultimate in the reduction of the dichasium as found in Lisianthius. The lateral branches of Type 1-E, as shown in Ficure 1 (opposite).. The probable sequen . e in the evolution of the dichasium in Lisianthius. See text above for explanation. 86 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FicureE 1-F, have been lost by reduction and a single flower remains. This type is found typically in L. axillaris and not infrequently in most of the other species of sect. ListANrTHIus subsect. Fruticos!. Types 1-A and 1-C then, are the ones typical of most species of sect. OMPHALOSTIGMA and sect. LISIANTHIUS subsect. HERBACEI. Types 1-E and 1-G are the ones typical of most of the species of subsect. Fruticosi of sect. LIstan- THIUS. The various arrangements of the inflorescences in the genus, with one possible sequence in their evolution, are shown schematically in FIGURE 2. Types 2-A, 2-B, and 2-C are similar, except for the modifications in the dichasia which were discussed earlier. In these three types, the dichasia are terminal and axillary on the primary and secondary shoots, all of which are determinate. The secondary shoots are frequently leafy below the first division and are basically similar to the upper portion of the primary shoot. In some species, such as Lisianthius brevidentatus (ar- rangement as in Ficure 2-C) and L. saponarioides (arrangement as in Ficure 2-A) the dichasia are frequently sessile or nearly so, and the up- per internodes very short, so that the inflorescences appear compacted. Types 2-A to 2-C are characteristic of the species in sect. OMPHALO- STIGMA and subsect. HERBACE! of sect. LISIANTHIUS. The remaining types shown in FicureE 2 are characteristic of the species in sect. LISIANTHIUS subsect. Fruticost. In these species the primary axes are indeterminate. The dichasia are axillary or, more commonly, terminal and axillary on determinate lateral shoots, which in this study will be referred to as “flowering branches.” Except for Lisianthius and a few closely related genera, the Gentian- aceae are herbaceous with terminal inflorescences. The condition in sect. LIsIANTHIUS subsect. FRuTIcosi, in which the plants are woody with indeterminate axes, must certainly be a derived one. Since the species of subsect. Fruticosi of sect. LIstANTHIUS are characterized almost entirely by having apparently simple dichasia, the evolution from determinate to indeterminate main axes has probably been through the type repre- sented by Ficure 2-B. Flowering branches are probably homologous with the lower lateral shoots of Types 2-A to 2-C. The basic type is shown in Ficurr 2-D. The branching is opposite and each division is terminated by an apparently simple dichasium. One to three pairs of reduced leaves are present below the lowermost trichotomy, but rarely are any present above. Rather, the upper divisions are subtended by somewhat foliaceous or scarious bracts. This type is found occasionally in most species of sect. ListANTHIUS sub- sect. Fruticosi and regularly in L. latifolius and the L. skinneri complex. Specialization has progressed from Type 2-D in two directions. Through Types 2-E (L. longifolius) to 2-F (L. longifolius) and 2-G (L. cordifolius) the flowering branches become more complex and freely branched, with reduced leaves found above the first division and on the secondary. axes. The dichasia are mostly reduced to single flowers. The branching is op- The arrangement of the inflorescences : pecaupae: and the probable sequence in the Fic woletion ot this arrangement. See text for explanati [ZL6T SAIH.LNVISIT AO NOISIATY “WAAVAM 88 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 posite, alternate or bifurcate in Types 2-E and 2-F, and nearly always bifurcate in Type 2- From Type 2-D through Types 2-H to 2-K, the flowering branches be- come simpler. The leaves are reduced in number to le pair sub- apparently simple dichasium (FicuRE 2-J), or, ultimately, a single flow- er (FicurE 2-K). Type 2-H is typical in all of the species of series Longi- folit (except for L. saaighe!: a" series Exserti of sect. LISIANTHIUS sub- sect. Fruticosr. Types o 2-K are found occasionally in the above- mentioned series, and ope 4 Ki is typical in L. axillaris. gf -L, which is characteristic of the two species in sect. LISIAN- THIUS subsect. Fruticos1 series Umbellati, closely resembles Type 2-H A Ficure 3. Cross-sections of the oo eee of representative Lisianthius spe- . longifolius, lobe: rus, lobes seid ocatina tls C. L. cordifolius, lobes erate: D. - auratus, lobes ridged. 1972] WEAVER, REVISION OF LISIANTHIUS 89 when dissected, and is clearly derived from it. The internodes above the lowermost division and the stalks and branches of the dichasia are greatly shortened, the whole affair resembling a capitulum or an umbel. The bracts subtending the two lowest divisions are large and foliaceous and appear involucrate. Flowers. In most species the calyx lobes are carinate abaxially to some degree. The width of the keel varies from species to species, from a slight, almost imperceptible ridge along the midline, to a conspicuous wing 4 mm. broad. In a few species the lobes are completely ecarinate. The terms used in this study to describe the abaxial surface of the calyx lobes are defined below. Lobes, drawn in cross section, representing each of the types are shown in Ficure 3 Ecarinate: the abaxial surface of the lobes uniformly convex (3-B). Ridged: the lobes thickened along the midline, the sides in cross section straight or slightly concave (3-D). Cc G Ficure 4. Flowers of representative species of Lisianthius, showing the vari- ous corolla shapes and the posture of the corolla lobes, X 1. A. L. longifolius, Bok: adamsii, corolla tubular-fun- nelform, inflated, lobes somewhat recurved; C. L. brevidentatus, corolla tubu- lar, lobes spreading; D. L. skinneri, corolla tubular, inflated, lo bes recurved at e tips; E. L. m nigrescens, corolla funnelform, lobes spreading; F. L. e aor corolla tubular-campanulate, lobes erect; G. L. sapo sariiles, saltiie salver- form, lobes flaring. 90 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Carinate: the lobes strongly thickened along the midline, the spat por- tion slightly projecting; in cross section the sides strongly concave (3-C) the thickened portion along the midline of the lobes expanded into a thin, flat, membranous and strongly projecting wing (3-A). Although the last three calyx-lobe types intergrade from species to species, they are generally constant in a given species. The structure of the corolla of all the species is basically the same. However, the proportions of the sectors vary, resulting sometimes in strikingly different shapes. The basal sector of the corolla tube, usually the lower third or sometimes the lower half, that portion enclosing the ovary and in turn enclosed by the calyx, is constricted into a narrow tube with parallel sides. The stamens are inserted near the distal end of this constriction. Above the insertion of the stamens, the corolla tube expands, often abruptly but sometimes gradually, into a somewhat broader tube which is terminated in the limb. The terms used in this study to describe the various corolla shapes are defined below. Tubular: the upper, expanded portion of the tube long and relatively narrow, with more or less parallel sides or somewhat inflated; the limb short and erect or gradually spreading (Ficures 4-C, 4-D). Tubular-funnelform: differing from the tubular only in that the limb is rela- tively long and recurved (Ficures 4-A, 4 Funnelform: the upper expanded neviitin of the tube long and narrow, taper- ing gradually to the throat; the limb relatively long and gradually spreading (FIGURE Di Tubslercombaneidis: the upper, expanded portion of the tube short and broad, tapering slightly to the throat; the limb short and erect or gradually spreading (Ficure 4-F). Salverform: the tube narrowly constricted for nearly the entire length, with the sides more or less parallel; the upper, expanded portion of the tube nearly obsolete, and the limb at right angles to the tube (Ficure 4-G). Pressing and drying frequently alter the shape of the corolla. The lower, constricted portion of the tube is greatly exaggerated in dried speci- mens, and the posture of the limb, erect, spreading or recurved, is fre- quently difficult to interpret. The corollas of all species are marcescent to some degree, that is the corolla withers after anthesis but does not absciss for some time. In some species, particularly those of sect. ListAaNTHTUS subsect. HERBACEI, the corolla may persist for months after the capsule matures and the seeds are shed. In most species of sect. LIsStantHrus, however, the corol- la does absciss before the capsule matures. The degree to which the corolla persists appears to be correlated with the amount of sclerification of the vascular traces to the corolla. Those species in which the fibrous bundles are most conspicuous and best developed are the species in which the corolla persists on the mature capsules. The stamens are inserted on the corolla tube ca. 1/2—1/3 of the distance 1972] WEAVER, REVISION OF LISIANTHIUS 91 from the base to the apex in sect. LIsIANTHIUs or just below the apex in sect. OMPHALOSTIGMA. In dried flowers, the point at which the filaments become free from the corolla is usually visible externally as a slight swelling on the lower constricted portion of the corolla tube. In sect. OMPHALOSTIGMA the filaments are of equal length in a given flower; in sect. List!ANTHIUs the filaments are of unequal length in a given flower, usually two short, two of intermediate length, and one long. From observations of greenhouse plants, it appears that the filaments are at their full length when the flower is open and the anthers have dehisced. The filaments are exserted in all species; they protrude from the throat of the corolla tube but do not necessarily surpass the lobes. The style is filiform and always exserted, in some species surpassing the corolla lobes in pressed flowers. As is the case with the filaments, the style appears to elongate little after the flower is mature and the anthers have dehisced. In the majority of species the style always sur- passes the filaments, the anthers never coming into direct contact with the stigma; in the species of sect. OMPHALOSTIGMA, the style is always over- topped by the filaments; in a few species of sect. LisIANTHIUs, the relative position of the two is variable. The stigma is either capitate or peltate, and in contrast to most gentians, only indistinctly bilobed (Ficure 5). The peltate type is more common in the genus, and it is probably the generalized type in that it resembles more closely the basic condition in the family. Most genera of Gentiana- ceae are characterized by having a deeply bilobed stigma with linear or lamellar lobes. The peltate type of stigma in Lisianthius resembles this generalized gentianaceous type, but the lobes are very small and some- what connate on the adaxial surface. In addition the margins are reflexed, obscuring the apex of the style and causing the lobes to appear hemispher- ical. The capitate stigma is easily derived from the peltate type. In its extreme form, as found in L. longifolius and L. cordifolius, the stigmatic A B Cc oer io’ as ; na Ficure 5. Stigmas of Lisianthius, X 15. A. L. longifolius, stigma capitate, showing the sahieas 3 B. The same, another view; C. L. nigrescens, stigma peltate. 92 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 lobes are further reduced in size and are entirely connate along their adaxial surfaces; the margins are not at all reflexed. The stigmas of the closely interrelated L. glandulosus, L. laxiflorus, and L. troyanus, al- though capitate, are somewhat intermediate to the peltate type. In these species the lobes are small with barely reflexed margins, but they are not entirely connate. Grisebach (1839) divided his genus Leianthus (= Lisianthius) into two sections based on their supposedly distinctive stigmatic types: sect. OMPHALOSTIGMA, with peltate stigmas, and sect. ListANTHIUS, with capitate stigmas. Later (1845) Grisebach elevated sect. OMPHALOSTIGMA to generic rank, as Petasostylis, without changing its definition, and in- cluded within it L. saponarioides and L. nigrescens. At the same time he included in the genus Leianthus (his old section Lisianthius), L. cordifolius, L. cuspidatus, L. exsertus, L. latifolius, L. longifolius, and L. umbellatus. The genera were separated primarily by the type of stigma, as the sections were earlier. Actually, L. nigrescens and L. saponarioides are the species in which the form of the peltate stigma is most extreme, However, except for L. longifolius and L. cordifolius, all of the species included in Leianthus (sensu Grisebach, 1845), have definitely peltate stigmas. Lisianthius nigrescens, according to my interpretation, is more closely related to all of the species included by Grisebach in Leianthus (sensu 1845) than it is to L. saponarioides. In fact it is so similar to L. cuspida- tus that most authors have considered the two synonymous. I have found the type of stigma to be nearly useless as a criterion in determining relationships in Lisianthius. The peltate and capitate types intergrade from species to species and sometimes, particularly in L. auratus and L. seemannii, within a species. Both types, in distinct form, are found in the Jamaican L. longifolius. Petasostylis has not been accepted by any authors since Grisebach, and there appears to be no reason whatever for maintaining it. Fruits. The fruit is a semiwoody, bivalvate, loculicidal capsule. Upon dehiscence the placentae are visible as whitish, erose bands along the edge of the valves adjacent to the sutures. The nearly mature capsule is generally flattened slightly on a plane perpendicular to the suture and is oblong-ellipsoid or fusiform in shape. However a few species in sect. ListaNTHIUS subsect. HERBACEI differ in that the capsules are ovoid in shape and not flattened. The two types are illustrated in Ficure 6. CHROMOSOME NUMBERS The chromosome number of nine species was recently reported as ” = 18 (Weaver, 1969). During the present study, the nine species listed below were investigated, and not a single deviation from the earlier re- ported number has been found: L. adamsii (Weaver 2342) ss Oo ~I to Le WEAVER, REVISION OF LISIANTHIUS 93 auratus (Weaver & Wilbur 2253) axillaris (Weaver & Wilbur 2252) brevidentatus var. brevidentatus (Weaver 2184) laxiflorus (Weaver & Howard 2349) nigrescens var. nigrescens (Weaver 2134) nigrescens var. chiapensis (Weaver 2168) oreopolus (Weaver 2173 peduncularis (Weaver & Wilbur 2247) quichensis (Weaver 21 viscidiflorus (Weaver 2179) ae a oP a a Ta ee eee A. L. nigrescens, capsule broadly Ficure 6. Capsules of Lisianthius, X 2. A ovoid; B. L. cordifolius, capsule oblong-ellipsoid. Flower buds were fixed in modified Carnoy’s solution and the pollen mother cells were subsequently squashed in acetocarmine. Voucher speci- mens have been deposited at DUKE, except for L. laxiflorus which is at A. With the exception of Gentiana L., Gentianella Moench, and Sabatia Adans. little is known of the chromosome numbers in the Gentianaceae. However from a perusal of the summaries of chromosome numbers in the Gentianaceae (Rork, 1949, and Live, 1953) and the chromosome indices, it appears that most of the larger genera are characterized by having more than one chromosome number. Sabatia, in which there is a wide array of numbers (Perry, 1967), is outstanding in this respect. Ha- lenia Borkh., on the other hand is similar to Lisianthius in that all of the six species investigated, including representatives from both the Old and New World, are characterized by having the same chromosome num- er. Although not all of the species have been investigated, a notable gap being the two species in sect. OMPHALOSTIGMA, change in chromosome number appears to have played no evolutionary role whatever within 94 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Lisianthius. Nor has chromosome number been helpful in determining relationships within the genus. In the genera supposedly most closely re- lated to Lisianthius, comprising the Gentianeae-Tachiinae of Gilg (1895), only two species, Eustoma russellianum G. Don, with nm = 36 (Rork, 1949), and Macrocarpaea thamnoides (Griseb.) Gilg, with mn = 21 (Weaver, 1969), have been investigated. Although these counts are too scanty to give more than a hint as to the relationships between Lisianthius and its supposed relatives, as previously pointed out by Weaver (doc. cit.), they do suggest a possible relationship between Lisianthius and Eustoma. FLORAL ANATOMY Various aspects of the floral anatomy of Lisianthius have been described by previous investigators. Perrot (1898) included three species, L. ex- sertus, L. nigrescens, and L. longifolius in his investigation of the anatomy of the Gentianaceae but limited his discussion of floral anatomy to the ovary of L. nigrescens. Woodson (1936) described the floral fibers of L. brevidentatus. Lindsey (1937, 1940) described in detail the vasculature of the flower of L. longifolius. The anatomy of the flowers of 21 species was investigated during the course of the present study, with emphasis on the floral fibers and placen- tation. Flowers were infiltrated with paraffin using the tertiary butyl al- cohol series suggested by Johansen (1940). The material was sectioned on a rotary microtome at 15 » and stained in 1% Safranin in 95% EtOH and 1% Fast Green in 95% EtOH (Johansen, loc. cit.). Flowers of most of the species investigated were collected in the field and preserved in FAA. The flowers of the remaining species were taken from herbarium specimens and boiled in water immediately before infiltration. The vascular traces to the calyx, corolla, and androecium of Lisianthius are variously supplied with heavily lignified elements. These elements were described by Woodson and illustrated (1936) from what he called L. brevidentatus. Woodson’s material was collected by W. A. Schipp in British Honduras and was almost certainly L. brevidentatus var. collinus, since the typical variety of L. brevidentatus is not known from that country and Schipp 1205 is the type and only collection of var. collinus. According to Woodson (1936), these heavily lignified elements are primarily fibers of metaxylary origin with a few phloem or pericyclic fibers around the periphery of the bundles. They are most conspicuous in the calycine and corolline traces but are also found in the staminal traces when the stamens are still adnate to the corolla tube. The present study has shown that there is considerable interspecific variation in the degree to which the fibers are developed, both in the same organ and among the various organs. Only in a very few species [L. oreopolus, L. brevidentatus, L. viscidiflorus (sect. ListaNTHIUS sub- sect. HERBACE!), L. skinneri, L. longifolius, L. glandulosus (sect. LISIAN- THIUs subsect. Fruticost), and L. meianthus (sect. OMPHALOSTIGMA) | are they well developed in the calyx, and then always more so in the tube 1972] WEAVER, REVISION OF LISIANTHIUS 95 than in the lobes. In general, the fibers are poorly developed in the staminal traces. Only in L. silenifolius, L. brevidentatus, and L. longi- folius are they at all conspicuous. It is in the corolline traces that the fibers are most widespread in occur- rence in the genus; they are well-developed in 15 of the 21 species in- vestigated. In those species in which they occur, the fibers are poorly developed at the base of the corolla in fresh, newly opened flowers; how- ever, in older flowers the fibers are well developed nearly to the base of the corolla. The development thus appears to be basipetal, as would be neceiy in a determinate organ. The fibers never extend into the corolla obes. The following species are well supplied with fibers in their corolline traces: Sect. LISIANTHIUS Sect. OMPHALOSTIGMA Subsect. HERBACEI Subsect. FRUTICOSI L. brevidentatus L. adamsii L. meianthus L. nigrescens L. auratus L. oreopolus L. axillaris L. quichensis L. cordifolius L. silenifolius L. glandulosus L. laxiflorus L. longifolius L. skinneri In the following species the fibers are poorly or not at all papheresats ta Ne corolline traces: Sect. LISIANTHIUS Subsect. HERBACEI Subsect. FRUTICOSI L, viscidiflorus L. capitatus L. exsertus L. latifolius L. troyanus L. umbellatus In all of the latter group of species, the marcescent corolla abscisses before the capsule matures and the seeds are shed. In the former group of species, except for L. skinneri, the corolla is more or less persistent on the mature capsule. Well developed fibers in the corolline traces obviously give sup- port to the marcescent corolla and determine to a large extent the degree to which it persists. Similar floral fibers were reported by Woodson (1936) to be present in Coutoubea spicata Aubl. (Helieae), Canscora Lam., Centaurium Gilib. (Gentianeae-Erythraeinae), and Zygostigma Griseb. (Gentianeae-Tachii- nae) but not in those genera, e.g. Chelonanthus Gilg, Macrocarpaea Gilg, Calolisianthus Gilg, etc., of the Helieae and Gentianeae-Tachiinae includ- ed in Lisianthius by some authors. According to Gilg, the Gentianeae-Tachiinae, of which Lisianthius is a member, are characterized by having unilocular ovaries. He illustrated 96 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 the ovaries of L. nigrescens (p. 92, Fig. 41-J) and L. saponarioides (p. 92, Fig, 41-E, actually L. acuminatus Perk., as will be explained later) as being unilocular. Perrot (1898) also described the ovary of L. nigrescens as being unilocular. However the present study has shown that all stages toward bilocularity, including complete bilocularity, are found in the Ovaries of the various species. The ovary of Lisianthius is bilocular at the base in all of the flowers examined. In most of the species, however, the septum becomes incomplete for 15 to 500 p» at the base of the placentae or below. As the placentae form, they abut and fuse and the ovaries appear bilocular again in the mid-portions. In five of the six species of sect. LisIANTHIUS subsect. Herpacet (L. brevidentatus being the exception) examined, however, the placentae do not abut except in the upper portions, and the ovaries are unilocular for most of their length. Three of the species examined, L. axillaris, L. skinneri, and L. latifolius, in subsect. FRUTICOSI, are com- pletely bilocular; L. troyanus and L. laxiflorus are bilocular except for a portion 15 » in length where the septum is incomplete. In most of the species then, the ovaries are unilocular with parietal placentation but show varying tendencies toward bilocularity. A few species are bilocular with axile placentation. Buds of Lisianthius umbellatus, a species characterized by having an almost completely bilocular ovary when mature, were sectioned at sev- eral stages in their development. In younger buds the septum is incomplete and the placentae completely separate. The arms of the septum and the inner surface of the placentae are covered with a distinct epidermis. In older buds the placentae abut as they enlarge; the epidermis then breaks down and the placentae fuse. In those species where it is bilocular or nearly so, the ovary is flattened along an axis perpendicular to the septum, and the placentae are brought into close physical contact thus allowing fusion to take place. In the species of sect. LisIANTHIUS subsect. HERBACEI where it is unilocular for most of its length, the ovary is not flattened and the placentae are widely separated. A few miscellaneous aspects of the floral anatomy of Lisianthius should be noted. Large crystals (druses) are found around the periphery of the basal portion of the ovary of L. troyanus. Crystals were not found in any other species of Lisianthius. Surrounding the basal portion of the ovary of all species investigated is a 5-lobed ring of densely staining, non- vascularized, glandular tissue. According to Lindsey (loc. cit.), ovaries glandular at the base are found in most of the genera associated with Lisianthius in the Gentianeae-Tachiinae by Gilg as well as those in the anatomically similar Helieae. BREEDING SYSTEMS The breeding systems of Lisianthius have not been studied in the field; the following observations were made from plants grown in the green- 1972] WEAVER, REVISION OF LISIANTHIUS 97 houses of Duke University. All six species studied, Lisianthius adamsit, L. capitatus, L. cordifolius, L. longifolius, L. nigrescens (var. chiapensis) and L, troyanus, are self-compatible. In all the species cited above, except Lisianthius nigrescens, the styles are considerably longer than the filaments, and the anthers do not come into physical contact with the stigmas. The flowers of these species are proterandrous and thus appear to be adapted for outcrossing. These plants, in fact, seldom set seed unless the flowers were artificially pol- linated. In most populations of L. nigrescens var. chiapensis, on the other hand, at least some of the filaments in each flower are as long as the styles, and some of the anthers do come into direct contact with the stigmas. The flowers of members of this taxon are characterized by hav- ing the stigmas receptive as the anthers dehisce and thus are adapted for self-pollination. These plants invariably set seed profusely in the greenhouse without artificial pollination. All species of Lisianthius have relatively showy flowers which produce copious amounts of nectar from glands at the base of the ovary, and al- though they are odorless, would seem to be quite attractive to animal pollinators. Little is known, however, of the pollinators of Lisianthius. No insects were observed on the plants in the field. Judging from the strikingly different colors of some of the flowers and the wide variation in flower form, particularly the shape and size of the corolla and the de- gree to which the styles and filaments are exserted, the various species are probably adapted to pollination by a wide variety of animals. In fact, specialization leading to pollination by different vectors has probably been a major evolutionary force within the genus. Smaller bees of the genus Bombus were frequent visitors during the summer months to flowers of cultivated Lisianthius adamsu, L. cordi- folius, L. longifolius, and L. troyanus. In all of these species the corolla tubes are relatively wide and the styles and filaments are well exserted. In addition the corolla lobes are widely spreading or even recurved, further exposing the anthers and stigmas. Bees invariably landed on the filaments and crawled into the corolla tube in search of the abundant nectar; they never came into contact with the anthers or the stigmas. Al- though Bombus is perhaps not a pollinator of the Lisianthius species in their native habitats, its behavior on cultivated plants strongly suggests that nectar-seeking bees are not effective pollinators of those species with well exposed anthers and stigmas. These species are probably pollinated by various lepidopterans, pollen gathering bees, and/or hummingbirds. Lisianthius saponarioides, with erect flowers and short, narrow corolla tubes seems well suited to pollina- tion by butterflies. Species in which the corolla tube is relatively broad and the anthers and stigmas held within the erect or slightly spreading corolla lobes are possibly pollinated by nectar-seeking bees. It appears that closely sympatric species of Lisianthius always differ strikingly in corolla and/or flower form. 98 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 POLLEN MORPHOLOGY Nilsson (1970) has recently described the morphology of pollen grains of what he called “Lisianthius s. lat.” His study, based on acetolyzed grains observed both with the light microscope and scanning and transmis- sion electron microscopes, included a large number of plants not referable to Lisianthius P. Browne. Only Nilsson’s “Longifolius-type” is found in that genus; the following description is condensed from Nilsson’s (loc. cit., p. 4). Pollen grains in monads, 3-colporate, subprolate to prolate, 28-50 xX 23-40 ». Colpi meridional, 2/3-3/4 of polar axis. Exine 3—5 p thick. Sexine as thick as nexine, or thicker, reticulate or occasionally + smooth. Lumina angular, ca. 0.5—7 » in diameter. Muri ca. 0.5—1 » wide. According to Nilsson (1970), the plants here included in Lisianthius are homogeneous in their pollen type. Of the species examined, 21 are characterized by having grains with a finely reticulate sexine pattern; the coarseness of the reticulum varies intraspecifically as well as interspecifical- ly and therefore appears to be of no taxonomic significance. Only L. ex- sertus varies significantly in its pollen morphology, and only in the sexine pattern which is nearly smooth when observed with the light microscope, but shows an intricate, very closely interwoven pattern in scanning pic- tures. The sexine of the grains of L. exsertus is basically the same as that in the other species, except that the lumina of the reticula are greatly re- duced (Nilsson, in litt.) According to Nilsson (1970) the pollen type found in Lisianthius (his Longifolius-type) is distinct from and not easily correlated with the pol- len types of the other plants included in his study. (These plants and their relationship to Lisianthius has already been discussed.) It differs from the only other single-grain types in its prolate or subprolate grains with a fine reticulate pattern, relatively narrow muri, and more or less angular lumina, and in the absence of verrucoid processes (Nilsson, loc. cit.). Nilsson (1970) found that Lisyanthus loranthoides Griseb., a Peruvian species, has pollen of his Longifolius-type. Grisebach (1839, 1845) placed this species in a monotypic section, CHoRIoPHYLLUM, of Lisyanthus Aubl. (not the same as Lisianthius P. Browne), but suggested that perhaps it belonged in a separate genus. Macbride (1959), although not having seen any specimens of this taxon, suggested that ex char. it might be the same as Macrocarpaea chlorantha Gilg. An isotype of L. loranthoides (Mat- thews 1315, ox¥) has been seen. With its conspicuously bifurcate stig- mas, 4-angled stems, and calyx, with obtuse lobes, fused for nearly 2/3 its length, this species clearly is not referable to Lisianthius as here in- terpreted. Morphologically it appears to be somewhat intermediate be- tween it and Macrocarpaea, and probably should, as Grisebach suggested, be placed in a distinct genus. 1972] WEAVER, REVISION OF LISIANTHIUS 99 INTRAGENERIC RELATIONSHIPS _ Chromosome number, floral anatomy, and pollen morphology have yielded few clues for determining relationships between the species of Lisianthius, Therefore a reliance on gross morphological features is nec- essary. Since there are few morphological trends within the genus, an infrageneric classification based on them must of necessity be highly sub- jective. Lisianthius longifolius, the lectotype species of the genus, and the other species also characterized by basically tubular or funnelform corollas and unequal filaments inserted in the lower half of the corolla tube, form a logical group which I am calling section ListantHIuS, The remaining two species, L. meianthus and L. saponarioides, which differ strikingly by their salverform corollas and their equal filaments inserted near the apex of the corolla tube, seem to form a second group of equal rank to be called sec- tion OMPHALOSTIGMA. Section LisIANTHIUS is comprised of two subsections, the HERBACEI with 7 species, monopodial, suffrutescent herbs with determinate main axes, usually compound dichasia, and sessile leaves, and the Fruticost with 18 species, subshrubs, shrubs, or even small trees with indeterminate main axes, usually apparently simple dichasia, and petiolate leaves. Section ListantHIus subsect. Fruticos1 is rather heterogeneous in a number of respects, and three subgroups are readily recognizable. Two Jamaican species, Lisianthius umbellatus and L. capitatus differ striking- ly from all other members of the subsection in that their inflorescences appear capitate or umbellate owing to extreme reduction of all the branches. These two species comprise the series Umbellati, The remain- ing species are characterized by having diffuse inflorescences which are obviously dichasioid. However, one of these, the Jamaican L. exsertus, Longifolii Exserti Umbeilati / FRUTICOSI HERBACE! Lisianthius FicurE 7. Probable relationships of the species groups in Lisianthius, «, 100 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 has flowers resembling L. capitatus in series Umbellati and is therefore intermediate between the species of that series and the remaining species in the subsection. In addition, the pollen of LZ. exsertus differs from that of all other members of the genus in having pollen with a nearly smooth rather than a distinctly reticulate sexine. On the basis of its somewhat intermediate Stihte and its unique pollen, L. exsertus has been placed monotypic series, Exserti. The remaining fifteen species, still a rather eS assemblage, but all with diffuse inflorescences, tu- bular or funnelform corollas, and pollen with a reticulate sexine pattern, comprise the series Longifolit. A suggestion as to the possible “phyletic” relationships between the various species groups is presented in FicURE 7. This scheme is based al- most entirely on gross morphological criteria and is highly conjectural. The generalized gentianaceous plant would appear to be somewhat as follows: a perennial, monopodial herb with a determinate main axis, in- florescences of compound dichasia, sessile leaves, rotate or salverform corollas, stamens of equal length in a given flower, and unilocular ovaries. If this type could be taken for the generalized one in Lisianthius, evolu- tion has proceeded as follows: (1) from perennial herbs to annual herbs (rarely) in one direction and to subshrubs, shrubs or trees in the other; (2) from a determinate main axis to an indeterminate one; (3) from compound dichasia to apparently simple dichasia to solitary flowers; (4) from sessile leaves to distinctly petiolate ones; (5) from a salverform corolla to a tubular-funnelform one, with variations; (6) from stamens of equal length in a given flower to stamens of unequal length; and (7) from a unilocular to a bilocular ovary. None of the extant species or species groups perfectly fits the hypo- thetical generalized type. The species in sect. OMPHALOSTIGMA approach it but differ in having petiolate leaves and nearly bilocular ovaries. Like- wise, the plants of subsect. HERBACEI also approach the hypothetical pro- totype, but they differ in having tubular or funnelform corollas and filaments of unequal length. Therefore, both sect. OMPHALOSTIGMA and sub- sect. HERBACEI appear to be early offshoots from the primitive type. The plants of subsect. FrutTicosi1 resemble those of subsect. HERBACEI in flow- er form, but differ in being fruticose, with indeterminate main axes, in- florescences of apparently simple dichasia, and petiolate leaves, all char- acters which deviate from the hypothetical prototype. Subsect. FruTicosi has very likely been derived from subsect. HERBACEI or from the line lead- ing to it. The series Exserti and Umbellati are both offshoots from the main line in subsect. Fruticost, leading to series Longifolii. The Exsertt shares characters with both of the other series and presumably arose from the line leading to the Umbellati. esd =} [To be continued | 1972] ROBERTSON, MALPIGHIACEAE 101 THE MALPIGHIACEAE IN THE SOUTHEASTERN UNITED STATES * KENNETH R. ROBERTSON MALPIGHIACEAE A. L. de Jussieu, Gen. Pl. 252. 1789, ““Malpighiae,” nom. cons. (MALPIGHIA FaMILy) *Prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University made possi p Mississippi, Arkansas, and Louisiana. The descriptions are based primarily on the plants in this area, with additional information from extraterritorial taxa in brackets. References I have not seen a arked by an asterisk. gestions ; to Dr. Bruce MacBryde (Pontificia Universidad Caté Saint Louis University, and the Missouri Botanical Nancy Dunkly for assistance with bibliographic problems and for typing the manu- from material collected by George N. Avery. 102 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Fruit a drupe [berry, nut, samara, schizocarp, or capsule]; seeds with little or no endosperm; embryo circinate [or prevailingly straight to curved.] Type Genus: Malpighia L. A family of 55-65 genera and 700-1300 species largely confined to the tropics and subtropics with the greatest diversification in tropical America. Only Heteropteris HBK. is represented in both hemispheres, with about 80 species in the Americas and H. leona (Cav.) Exell in West Africa. Byrsonima lucida (Miller) DC. is the only species of the family in our area, but Malpighia glabra L., Janusia gracilis Gray, Aspicarpa hyssopi- folia Gray, A. longipes Gray, A. humilis (Bentham) A. Juss., and Galph- imia angustifolia Bentham (Thryallis angustifolia (Bentham) Kuntze; ” closely related to, or perhaps better included within, G. brasiliensis (L.) A. Juss.; cf. MacBryde) occur in the southwestern United States. The Malpighiaceae are often divided into two subfamilies, Malpi- ghioideae (Planitorae Niedz.) and Gaudichaudioideae A. Juss. (Pyra- midotorae Niedz., Hiraeoideae Scholz in Melchior, nom. inval.), which are subdivided into six tribes based primarily on fruit morphology. Only tribe Malpighieae (fruits drupes or 1—3-locular nuts) is represented in our area. The Malpighioideae are neotropical shrubs or trees with flat floral receptacles and unwinged fruits, while the Gaudichaudioideae are pantropical, mostly lianas, usually with pyramidal receptacles and winged or crested fruits. MacBryde, however, has reported 3-angled pyramidal receptacles in Verrucularia glaucophylla A. Juss., of the Malpighioideae- Galphimieae, and has suggested that a reévaluation of the traditional subfamilial division is needed. The current system of classification is that of N iedenzu, which was based on the system proposed by A. Jussieu. As noted by C. V. Morton, “Although Niedenzu’s taxonomic work is detailed and careful, he ap- parently made up his rules of nomenclature as he went along. . . . Per- haps more than any other taxonomist, Niedenzu found it necessary to sub- divide groups, and so in the larger genera there are numerous subgenera, each with sections, subsections, and series and subseries. It is doubtful that this plethora of names is really necessary.” Because of the methodol- ogy of Niedenzu, and that of various other authors, the nomenclature of much of the family is chaotic, although many problems have been clarified by the recent works of Morton, MacBryde, and Cuatrecasas. Generally aligned with families included by Engler & Prantl in the Geraniales, the Malpighiaceae, Polygalaceae, Trigoniaceae, and Vochysia- ceae were included by Scholz, Cronquist, and Thorne in their Rutales, Polygalales, and Geraniales, respectively. Others ally the Malpighiaceae with the Linaceae, Geraniaceae, Zygophyllaceae, Erythroxylaceae, and re- *Thryallis Martius (1829; type species: T. longifolia Mart.) has been conserved over Thryallis Linnaeus (1762; type species: T. brasiliensis L.) for a small Brazilian genus. Thus, the name Galphimia Cav. (1799; type species: G. glauca Cav.) can be used without conservation for an American genus of about ten species, even thou it is a taxonomic synonym of Thryallis L. (cf. Taxon 16: 76, 77. 1967; 17: 328. 1968). 1972] ROBERTSON, MALPIGHIACEAE 103 lated families. Takhtajan places this latter group of five families in his Geraniales (derived from the Rutales, Saxifragales, and Dilleniales), while Hutchinson includes them in the Malpighiales (derived from the Tiliales, Bixales, and Dilleniales). Cronquist considers the origin of the Geraniales, Linales, and Polygalales to be from the Rosales via the Sapin- dales. A distinctive feature of the family, although by no means confined to it, is the “Malpighian” type of trichome, which is unicellular, two-armed, and attached by a more or less medifixed stalk. Unbranched, stellate, glandular, and stinging trichomes also occur in the family. Complex, somewhat fleshy, paired glands visible to the naked eye frequently occur on the outside of the sepals and on the leaf margins, lower leaf surfaces, or petioles. The stomata are paracytic, and the wood is commonly diffuse- porous with a tendency for the simple-perforated vessels to be aggregated into radial chains or clusters. The parenchyma is prevailingly paratracheal, and the multiseriate rays are heterogeneous. Although the anatomy of many lianous species is normal, some have highly complicated stelar structures, The known chromosome numbers (of about 30 species in 14 genera) sug- gest that considerable diversification through polyploidy and aneuploidy has occurred. Reported sporophytic chromosome numbers are 12, 18, 20, 22, 24, 34, 38, 40, 42, 54, 56, 58, 72, ca. 84. More than one chromosome number has been reported in Byrsonima crassifolia, Galphimia angusti- folia, possibly G. glauca, Hiptage Madablota, Malpighia glabra, and Stigmaphylion ciliatum. Two basic pollen types are found in the Malpighiaceae. In the first the grains are 3(4, 5)-colporate, with the colpal furrows distinct, united at the poles (syncolporate) or bibranched (parasyncolporate), while grains of the second pollen type are 4- to ca. 24-porate, often with com- plex geometric shapes correlated with the number of apertures, and with vestigial furrows forming surface patterns. In her study of African and Madagascan genera (including four American genera for comparison), Lobreau found correlations between pollen morphology and certain an- droecial and gynoecial features and indications that Malpighiacean pol- len is similar to that of the Hypericaceae and Caryocaraceae. Erdtman found palynological similarities with the Humiriaceae, Tremandraceae, Trigoniaceae, and Zygophyllaceae. A thorough study of New World gen- era is needed for confirmation of a correlation of pollen morphology and the established subfamilial classification. phe Although actual pollination has been observed only in Malbpighia, Bunchosia L. C. Rich. ex HBK., and Hiptage Gaertner, the flowers of most genera appear to be “bee flowers.” There are striking similarities between the flowers of certain Malpighiaceae and those of some species of Oncidium (Orchidaceae), but little has been published about this re- markable convergence. The floral mechanism of Hiptage Madablota is said to resemble that of Aesculus (Hippocastanaceae). Dimorphic to highly modified cleistogamous flowers occur in certain genera. 104 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The embryo sac is usually tetrasporic, 16-nucleate and of the Penaea type. Galphimia glauca and G. gracilis have bisporic, eight-nucleate, Al- lium-type embryo sacs. Species of Heteropteris HBK., Stigmaphyllon A. Juss., Banisteria, Hiptage, and Aspicarpa L. C. Rich. that have been in- vestigated show nucellar polyembryony with fertilization evidently not occurring. (Fertilization has been reported only in Malpighia glabra, but it probably occurs in Tristellateia australis.) Further studies are needed to ascertain the extent and effects of apomixis in the family. Fossil materials, mostly leaves and fruits, allegedly representing Mal- pighiaceae have been reported from the Middle Cretaceous to the Pliocene. Unlike members of numerous other families that are primarily tropical today, the Malpighiaceae evidently did not occur in the paleoflora of the present-day Arctic Zone, although the family is recorded from the Tertiary flora of Europe. The economic uses of members of the Malpighiaceae are few, consider- ing the family’s size. Extracts and infusions from certain species are used as narcotics, folk medicines, insecticides, pigments, and tanning agents. Certain South American Indian tribes, primarily in the Amazon and Orinoco river basins, prepare hallucinogenic infusions variously known as caapi, ayahuasca, yajé, natema, and pinde from species of Banisteriop- sis C. B. Robinson ex Small. The most commonly used species is B. Caapi (Spruce ex Griseb.) C. V. Morton, which contains the 8-carboline alkaloid bases harmine, harmaline, and d-tetrahydroharmine. Harmine has also been isolated from B. inebrians C. V. Morton and from Cabi paraensis Ducke, a species used as a folk medicine but evidently not as a hallu- cinogen. Such harmala alkaloids do not occur in B. Rusbyana (Niedz.) C. V. Morton, a species used as an additive in hallucinogenic beverages, but relatively high concentrations of N, N-dimethyltryptamine and traces of other tryptamines do occur. A narcotic drink is also made from Tetrap- teris methystica R. E. Schultes, on which chemical studies have not yet been made. It seems likely that other genera and species of Malpighiaceae will be found to be used as narcotics, medicines, and poisons by aboriginal peoples. The fruits of Byrsonima, Bunchosia, and Malpighia species are edible, and the drupes of Malpighia glabra L.,? Barbados or West Indian Cherry, Acerola, Cereza, contain exceptionally high amounts of Vitamin C, from 15 to 100 times the concentration found in oranges. This species is grown to some extent in Florida, Hawaii, and the West Indies, and vitamin tablets and frozen or canned juice made from the fruits are Ecuiiiipevially available. The fibrous stems of Heteropteris umbellata and Banisteria longifolia are used for roping and basketry binding respectively. Strong and attractive wood used in furniture and construction is obtained from the arborescent species of Byrsonima. Many species are quite ornamental ° Malpighia glabra L. and M. punicifolia L. have been considered to be either con- specific or distinct, with the cultivated plants placed in either species or thought to be hybrids beeen them 1972] ROBERTSON, MALPIGHIACEAE 105 and are in cultivation in the tropics and in conservatories, although not to the extent that they deserve. Species of Bunchosia, Byrsonima, Galphi- mia, Heteropteris, Malpighia, and Stigmaphylion are cultivated to some extent in Florida and southern California. REFERENCES: ArENEs, J. Trois genres de Malpighiacées nouveaux pour la flore malgache et pour la science. Not. Syst. Paris 12: 126-136. 1946. [Revised subfamilial classification proposed in key, 136; cf. MACBRYDE. ] . Répartition géographique des Malpighiacées vivantes et fossiles. Compt. Rend. Soc. Biogeogr. 290: 81-108. 16 maps on 3 pp. 1959. [Proposes hypo- thetical scheme based on fossil evidence for the origin, evolution, and dispersal of the family. ] oe a AsENJo, C. F. Aspectos quimicos y nutritives de la acerola (Malpighia punici- folia L.). Ciencia México 19: 109-118. 1959. [Literature review. ] BAILLON H. Malpighiacées. Hist. Pl. 5: 429-469. 1874. [English transl. M. M. Hartoc, The natural history of plants. 5: 429-469. London. 1878. ] BENTHAM, G., & J. D. Hooker. Malpighiaceae. Gen. Pl. 1: 247-262. 1862. [49 genera in 4 tribes. Berry, E. W. The affinities and distribution of the lower Eocene flora of south- eastern North America. Proc. Am. Philos. Soc. 53: 129-250. 1914. [Malpi- ghiaceae, 187—190.] BOLKHOovsKIKH, Z., V , V. Grir, T. Matveyeva, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. Ed. by A. A. Feporov. (Russian and English prefaces.) 926 pp. Leningrad. 1969. [Malpighiaceae, 419.] Bristot, M. L. The psychotropic Banisteriopsis among the Sibundoy of Co- lombia. Harvard Univ. Bot. Mus. Leafl. 21: 113-140. 1966. CANDOLLE, A. P. pe. Malpighiaceae. Prodr. 1: 577-592. 1824. ae Cxopar, R., & W. Viscuer. La végétation du Paraguay. V. Malpighiacees. Bull. . Bot. Genéve II. 9: 55-107. 1917. [Includes anatomical and mor- phological details. Cuawson, B. J. aie of Janusia gracilis. Kansas Univ. Sci. Bull. 7: 189- 197. 1913. Cortint, C. Sviluppo del gametofito femminile in Malpighia fucata Ker-Gawl. (Malpighiaceae). (English summary.) Caryologia 11: 42-56. 1958. [Pe- naea, Scilla, and Malpighia types of embryo-sac development occur in amily. Cozzo, D. Anatomia del lefio secundario de Tricomaria Usillo Gill. ex H. et A. Lilloa 13: 17-21. 1 pl. 1947 an Cuarrecasas, J. Malpighiaceae. Jn: Prima Flora Colombiana. 2. Webbia 13: 343-664. pls. 28-57. 1958; Apéndice 1. Ibid. 15: 393-398. 1960. [21 gen- era, 168 spp.; comprehensive taxonomic treatment. | Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. 1966. [Malpighiaceae, 168, 169. ae ise DEULoFEU, V. Chemical compounds isolated from Banisteriopsis and re te species. Jn: D. H. Erron, ed., Ethnopharmacologic search for psychoactive : ; ' te- Duxe, J. A. On tropical tree seedlings. I. Seeds, seedlings, systems, and syste: matics. Ann. Missouri Bot. Gard. 56: 125-161. 1969. [ Bunchosia corni- folia, 151, fig. 38.] 106 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ENDLICHER, S. Malpighiaceae. Gen. Pl. 1057-1065. 1840. [Published in same month (April) as part of A. de Jussieu’s preliminary treatment of Mal- pighiaceae; see discussion in MAcBrybE, 1970. ErpTMAN, G. Pollen morphology and plant taxonomy. Angiosperms. Corrected reprint and new addendum. xiv + 553 pp. Frontisp. New York. 1966. [Mal- pighiaceae, 259-261. ] FovET, M. Contribution a l'étude cyto-taxinomique des Malpighiacées. Adan- sonia IT. 6: 457-505. 1966. [Includes list of all known counts in family and hypothetical evolution of chromosome numbers. | FRIEDBERG, C. Des Banisteriopsis utilisés comme drogue en Amérique du Sud. Jour. Agr. Trop. Bot. Appl. 12: 403-437. pls. 1-6; 550-594; 729-780. pls. 1-6, table, map. 1965. [Excellent ethnobotanical summary; also issued as a monograph with continuous pagination. ] GrisepacH, A. H. R. Malpighiaceae. Jn: C. F. P. Martius. Fl. Brasil. 12(1): 1-123. pls. 1-22. 1858. HauMAN-Merck, L. Observations sur la pollination d'une Malpighiacée du . Bo HEGNAUER, R. Chemotaxonomie der Pflanzen. Band. 5. Dicotyledoneae: Mag- noliaceae-Quiinaceae. 506 pp. Basel & Stuttgart. 1969. [Malpighiaceae, 23- 29. Hermscu, C., Jr. Comparative anatomy of the secondary xylem of the “Grui- nales ss and “Terebinthales,” of Wettstein with reference to taxonomic grouping. Lilloa 8: 83-198. pls. 1-17. 1942. [Malpighiaceae, 102, 103, 185, 186, figs. 11-13, 19.] Humpoxpt, F. H. A. von, A. J. BoNPLAND, & C. S. Kuntu. Malpighiaceae. Nov. Gen. Sp. Pl. 5: ed. fol. 112-134, ed. qu. 145-174. pls. 445-452. 1822. [Byrsonima, 113-118 (147-153). pls. 446-449.] Hutcuinson, J. The genera of flowering plants. Vol. 2. xi + 659 pp. London. 1967. [Malpighiaceae, 569-592. Jacogs, J. Malpighiaceae. In: C. G. G. J. vaN STEENIS, ed. Fl. Malesiana I. 5: 125-145. 1955. [7 genera. |] Jussteu, A. H. L. pe. Malpighiacearum synopsis, monographiae mox edendae prodromus. Ann. Sci. Nat. Bot. IT. 13: 247-291, 321-338. 1840. [Part pub- lished in same month (April) as Endlicher’s treatment. | . Monographie de la famille des Malpighiacées. Arch. Mus. Hist. Nat. Paris 3; 5-151, 255-616. pls. 1-23. 1843. [Issued separately with contin- uous pagination as “Monographie des Malpighiacées, ou exposition des caractéres de cette famille des plantes, des genres et espéces, qui la com- posent.”’| Knutu, P. Handbook of flower pollination. Vol. 2: (Transl. J. R. A. Davis.) viii + 703 pp. Oxford. 1908. [Malpighiaceae, 217, 218. In German ed. cf. 2(1): 215, 216.] Kostermans, A. J. G. H. Malpighiaceae. Jn: A. Putie, Fl. Surinam 2(1): 146-243. 1936a. . Studies in South American Malpighiaceae, Lauraceae and Hernandia- ceae, especially of Surinam. Meded. Bot. Mus. Utrecht 25: 1-70. 1936b. [Malpighiaceae, 4-12, 56, 57.] Launert, E. Malpighiaceae. Jn: E. Mitne-RepHeap & R. M. PoLuitt, eds., Fl. Trop. E. — 28 pp. 1968. [6 genera. ] Levin, R. B. The Barbados or West Indian cherry. Ceiba 6: 33-37. 1957. [Mal- pighia glabra, cultivated in Florida. ] 1972] ROBERTSON, MALPIGHIACEAE 107 LosrEav, D. Contribution a l’étude du pollen des Malpighiaceae d’Afrique. Pollen Spores 9: 241-277. 1967. [Four American genera included; - thetical scheme of pollen evolution in family; chart showing correlations between floral and palynological features. | . Le pollen des Malpighiacées d’Afrique et de Madagascar. Bull. Inst. Fondamental Afr. Noire A. 30: 59-83. 1968. [Includes palynograms of 19 genera, chart of types of symmetry. ] Lusgock, J. A contribution to our knowledge of seedlings. Vol. 1: vii + 608 London & New York. 1892. [Malpighiaceae, 291-293. ] Macsrie, J. F. Malpighiaceae. Jn: Fl. Peru. Publ. Field Mus. Bot. 13: 781- 871. 1950. [15 genera. ] MacBrype, B. A revision of the Galphimiinae Ndz., Malpighiaceae. 248 pp. Ph.D. Thesis, Washington University, St. Louis. 1970. [Diss. Abstr. Int. 31(4): 1763B. 1970.] . In: Chromosome numbers of phanerograms. 4. Ann. Missouri Bot. Gard. 57: 383. 1970 [1971]. [Galphimia angustifolia, G. gracilis, n = 12.] Metcatre, C. R., & L. CHALK. Malpighiaceae. Anat. Dicot. 1: 279-285. 1950. Morton, C. V. Enumeration of the Malpighiaceae of the Yucatan Peninsula. Carnegie Inst. Publ. 461: 125-140. 1936. [10 genera. ] _ A typification of some subfamily, sectional, and subsectional names in the family Malpighiaceae. Taxon 17: 314-324. 1968. Naranjo, C. Psychotropic properties of the harmala alkaloids. Jn: D. H. EFRON, ed., Ethnopharmacologic search for psychoactive drugs. U.S. Public Health Serv. Publ. 1645: 385-391. 1967. [Includes effects of harmine, harmaline, and tetrahydroharmine on human volunteers. | Naranjo, P. Etnobotanica de la ayahuasca (Banisterio psis sps.): religion y medicina. Ciencia Naturaleza 12(2): 3-92. 1969. [Extensive bibliography. ] NARASIMHACHAR, S. G. Megasporogenesis and female gametophyte in Mal- pighia punicifolia Linn, Curr. Sci. Bangalore 6: 507-509. 1938. NIEDENzU, F. Malpighiaceae. Nat. Pflanzenfam. III. 4: 40-70. 1897 . Malpighiaceae. Pflanzenr. IV. 141(Hefte 91, 93, 94): 1 [Worldwide revision of family.] O’DoneELt, C. A., & A. Lourterc. Malpighiaceae Argentinae. Lilloa 9: 221-316. pls. 1-18. 1943. [15 genera.] OstenporF, F. W. The West Indian Cherry. Trop. Abstr. 18: 145-150. 1963. [Malpighia punicifolia, review of nomenclature and literature. Peretra, E. Contribuicdo ao conhecimento da familia Malpighiaceae. Arq. Serv. Florestal Rio Janeiro 7: 11-70. 1953. Pryx, L. van vER, & C. H. Dopson. Orchid flowers, their pollination and evolu- tion. x + 214 pp. Coral Gables, Florida. 1966. [Similarity between Mal- pighiaceae and Oncidium flowers and photograph of Malpighia flower, 141, 142 -870, 1928. Recorp, S. J., & R. W. Hess. Timbers of the New World. xv + 640 pp. pls. 1-58. New Haven, Conn. 1943. [ Malpighiaceae, 346-350. ] Roy, R. P., & N. C. Misura. Cytological studies in Malpighiaceae. Proc. 49th Indian Sci. Congr. 3: 335. 1962. i ScHENCK, H. Beitrige zur Biologie und Anatomie der Lianen. II. Beitrage zur Anatomie der Lianen. xiv + 271 pp. pls. 1-12. Jena. 1893. [Malpighiaceae, 110-125. pls. 6, 7.] Scuoiz, H. Malpighiaceae. Jn: H. MELCHIOR, Engler’s Syllabus der Pflanzen- familien. ed. 12. 2: 272, 273. 1964. 108 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Scuuttes, R. E. The identity of the malpighiaceous narcotics of South America. Harvard Univ. Bot. Mus. Leafl. 18: 1-56. 1957. [Botanical determinations of plants used by Indians in preparation of caapi, yajé, and ayahuasca. | —. Hallucinogens of plant origin. Science 163: 245-254. 1969. [Malpighia- ceae, 250; good summary. | The plant kingdom and hallucinogens. Bull. Narcotics 21(3): 3-16; 21(4): 15-27. 1969; 22(1): 25-53. 1970. [Malpighiaceae, 22(1): 25-29; good summary. | . The botanical and chemical distribution of hallucinogens. Ann. Rev. Pl. Physiol. 21: 571-598. 1970. [Malpighiaceae, 576, 577, 587; references. ] SEMPLE, J. C. Jn: Chromosome numbers of phanerogams. 4. Ann. Missouri Bot. Gard. 57: 384, 1970 [1971]. [Galphimia angustifolia, 2n = 20; Heterop- teris angustifolia, 2n = 34; Lophanthera lactescens, 2n = 12. Smpprgur, S. A. The microsporangium and the male gametophyte in Malpighia coccigera Linn. Beitr, Biol. Pflanzen. 44: 361-364. 1968. StncH, B. Studies in the Malpighiaceae. I. Morphology of Thryallis glauca Kuntze. Hort. Advance 3: 1-19. 1959; II. Morphology of Malpighia glabra Linn. /bid. 5: 83-96. 1961 (also cf. Proc. Indian Sci. Congr. 3: 260. 1962); III. oe and structure of seed and fruit of Malpighia glabra Linn. Ibid. 5: ee ise . neo N. Am. Fl. 25(2): 117-171. 1910. [22 genera, 5 ec R. Notes of a botanist on the Amazon & Andes. Edited and condensed by A. R. WaALLace. Vol. 2. xii + 542 pp. 4 maps. London. 1908. [Banisteria Caapi, 414~425; there is a discrepancy between Spruce’s field notes, which are conserved at Kew, and the published account; see Schultes (1957), p. 15.] STANDLEY, P. C. Malpighiaceae. Jn: Trees and shrubs of Mexico. Contr. U.S. Natl. Herb. 23: 563-578. 1923. Stenar, H. Zur Embryosackentwicklung einiger Malpighiaceen. Bot. Not. 1937: 110-118. 1937. [Malpighia urens, Galphimia gracilis. Suppa Rao, A. M. A note on the development of the female gametophytes of some Malpighiaceae _ oe in Hiptage Madablota. Curr. Sci. Bangalore 6: 280-282. . The ovule and eel a development of some Malpighiaceae. /bid. 8: 79-81, 1939. . Studies in the Malpighiaceae. I. Embryo-sac development and embryo- geny in the genera Hiptage, Banisteria and Stigmatophyllum. Jour. Indian Bot. Soc. 18: 145-156. 1940. [Fertilization does not occur.| II. Structure and development of the ovules and embryo-sacs of Malpighia coccifera Linn. and Tristellateia australis Linn. Proc. Natl. Inst. Sci. India 7: 393- 404. 1941. Urnor, J. C. T. Dictionary of economic plants. ed. 2. 591 pp. Lehre. 1968. [Genera of Malpighiaceae, 66, 90, 92, 267, 269, 328, 516.] Vai, A. M. A preliminary list of the North American species of Malpighiaceae and Zygophyllaceae. Bull. Torrey Bot. Club 22: 228-231. 1895. VéELEz, L. Notes on the ecology of West Indian species of Malpighia. Science 124: 317. 1956. Witczek, R. Malpighiaceae. Fl. Congo 7: 214-234. 1958. [6 genera. ] Wituls, J. C. A dictionary of the flowering plants and ferns. ed. 7. (Revised 1972] ROBERTSON, MALPIGHIACEAE 109 by H. K. Amy Suaw.) xxii + 1214 pp. + liii pp. (Key to the families of flowering plants.) Cambridge, England. 1966. [Malpighiaceae, 689, 690.] Subfamily MALPIGHIOIDEAE Tribe MALPIGHIEAE 1. Byrsonima L. C. Richard ex Humboldt, Bonpland & Kunth, Nov. Gen. Sp. 5: ed. fol. 113, ed. qu. 147. 1822.4 Low spreading shrubs to small [or large] trees [rarely lianas], often much branched; bark warty [or smooth]; glabrescent, with the simple or T-shaped, often rust-colored trichomes largely confined to the younger parts [or the leaves, peduncles, pedicels and twigs densely pubescent]. Leaves opposite, glandless, simple, entire, usually coriaceous, sessile or shortly petiolate; petioles mostly stout; blades obovate-spatulate [or ovate to circular], the upper surfaces commonly glossy when alive but becoming glaucous upon drying; stipules axillary [or interpetiolar] and mostly connate, small and persistent. Flowers pedicellate, the pedicels jointed and 2-bracteolate, in terminal, many-flowered, simple racemes (oc- casionally compound at the base). Sepals 5, persistent, in anthesis the distal portion incurving between the petal claws and forming a con- striction around the ovary and base of styles, the tips recurving; abaxial glands 2 on each sepal, contiguous, somewhat fleshy [or absent]. Petals 5, glabrous, subequal [to distinctly unequal], white to pink or red [or yellow], turning a deeper color with age; claws long, recurved; blades strongly concave, reniform to subcircular with undulate and + entire [to slightly toothed] apices. Stamens 10, nearly equal in length, all bearing anthers; filaments thick, flattened and connate below, pubescent inside; anthers basifixed, 2-locular and introrse at anthesis, glabrous [or pubes- cent], connectives thick [and often forming prominent apical appendages}. Gynoecium syncarpous, 3-carpellate; ovary 3-lobed, conical [ovoid or globose], glabrous; styles 3, free, terminal, glabrous, slender, + straight, tipped by the subulate stigmatic surfaces; receptacle flat or slightly con- cave, densely pubescent. Fruit a green, yellow, orange, red or brown globose [ovoid or conical] drupe, crowned by the persistent styles; stone solitary; pericarp rather thick and fleshy; endocarp hard, woody, pitted [or smooth], angular, 3-seeded, rarely fewer seeded by abortion. Seeds subglobose; embryo circinate, the cotyledons subequal, terete. Lecto- TYPE spEcIES: Malpighia spicata Cav. = Byrsonima spicata (Cav.) DC.; see J. K. Small, N. Am. Fl. 25: 166. 1910.5 (Name from Greek, byrsa, “The generic name is usually attributed to L. C. Richard ex A. L. de Jussieu (1811). However, De Jussieu merely suggests that it might be possible to follow Richard’s idea of separating Malpighia L. into three genera, Malpighia, Byrsonima, and Bun- chosia; he does not indicate that he adopts this division. see discussion in MacBryde. : * Humboldt, Bonpland & Kunth referred M. spicata Cav. and seven other species of Malpighia to Byrsonima but did not make the formal transfers. It remains for a 110 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ally S \ Si \\\\ Lge US A = FicureE 1. Byrsonima. a-k, B. lucida: a, flowering branch, X 1; b, flower, X 5: C, Calys, sho wing ‘lands on bi sepal (glands on adjacent sepals removed), of flow vu semidiagram n h locule, X 4; j, k, opposite sides of the same embryo to show position of rar oriented as in seed, X 6. hide, and Latin nimia, very much (used), in reference to the use of the bark in tanning leather.) A tropical American genus of more than 100 species ranging from southern Florida, the West Indies, and Mexico to Bolivia and southern monographer to determine whether this species chosen by Small can be considered to be part of the protologue a thus eligible to be the lectotype or whether one of the ten species described under Byrsonima would be a more appropriate choice. 1972] ROBERTSON, MALPIGHIACEAE 111 Brazil. Many species grow in dry or harsh habitats such as savannas, rocky hills, coastal thickets, serpentine and palm barrens, and pinelands; other species occur in selvas (tropical rain forests) and montane or cloud forests. Byrsonima, Alcoceratothrix Niedz. (usually included in Byrsoni- ma), Diacidia Griseb., Burdachia A. Juss., and Glandonia Griseb. con- stitute subtribe Byrsoniminae Niedz. The genus is represented in our area by Byrsonima lucida (Miller) DC. (including B. cuneata (Turcz.) P. Wilson; cf. Sargent, 1921), locust- berry, a plastic species varying in habit from low spreading shrubs less than a meter high to trees exceeding eight meters with trunks 25 cm. in diameter. Distinguished by small spatulate-obovate leaves, elongate in- florescences with small bracts, pinkish flowers, glabrous anthers, pubescent filaments, and connectives much shorter than the anther sacs, B. lucida occurs in pinelands, low woods, and hammocks in the Everglades and Florida Keys and throughout most of the West Indies from the Bahamas to Barbuda and Trinidad. Following his penchant for subdividing taxa, Niedenzu (1928) divided Byrsonima into two subgenera, four sections, eight subsections, twelve series, and four subseries. In his system, B. lucida, along with three West Indian and six South American species, was placed in subgen. BRAcHyY- ZEUGMA, sect. SERICOLEPIS, subsect. Psilotheca, series Dictyoneura, sub- series Glossolepis. The correct sectional name is EPIPHYLLARION Griseb.; C. V. Morton (1968) chose B. lucida as the lectotype of this section. The genus needs a thoroughly documented modern taxonomic revision conducted on a broad geographic basis to delimit and classify the species and subgeneric taxa accurately and to apply the current rules of nomen- clature uniformly. Only Byrsonima verbascifolia (L.) DC. has been studied palynologically. The grains are 3-colporate with lalongate ora, prolate, rounded-triangular to subcircular in polar view, and with a reticulate exine. Similar pollen is found in tribe Galphimieae Niedz. Other genera of tribe Malpighieae which have thus far been examined (Malpighia and Bunchosia) have porate pollen. Although quite ornamental, Byrsonima lucida is infrequently cultivated. Byrsonima crassifolia (L.) HBK., 2n = 20, 24 (including B. cotinifolia HBK.; leaves large, elliptic or rarely somewhat obovate, flowers yellow to reddish), B. coccolobifolia HBK. (leaves large, subcircular to broadly ovate or obovate, flowers pink to red, connectives much longer than the anther sacs), and B. spicata (Cav.) DC. (leaves large, elliptic with acute apices, flowers yellow) are also grown to a limited extent in Florida. he fruits of all Byrsonima species reputedly are edible, and some are sold in local markets in Latin America. The wood of arborescent species is strong and heavy, has a reddish color, and is used for furniture and construction or is converted into charcoal. Certain species are used in tanning leather, as a source of inks, dyes or paints, and in folk medicine as an astringent and for the treatment of wounds, fevers, colds, diarrhea, snake bites, and leg tumors! 112 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 REFERENCES: Under family references see ARENES, 1959; BENTHAM & HOOKER; CUATRE- CASAS; FouET, GRISEBACH; Humpoipr et al. KosTERMANS; LOBREAU; Mor- TON, 1968: NIEDENZU, 1928: SMALL; UPHoF; aaa VAIL. Duke, J. A. Keys for the identification of seedlings of some prominent woody species in eight forest types in ae Rico. Ann. Missouri Bot. Gard. 52: 314-350. 1965. [B. coriacea, fig. 6 Jussieu, A. L. bE. Dixiéme rice sur les caractéres des familles tirés des graines, et confirmés ou rectifiés par les observations de Gaertner. Renon- culacées—Malpighiacées. Ann. Mus. Hist. Nat. Paris 18: 472-487. 1811. [Byrsonima and Bunchosia first published, but only provisionally, cf. 480, Se E. A. Flowering trees of the world for tropics and warm climates. 336 pp. 422 col. pls. on 65. New York. 1962. [Byrsonima, 148, pls. 181, 182.] Nanpa, P. C. Chromosome numbers of some trees ra shrubs. Jour. Indian Bot. Soc. 41: 271-277. 1962. [B. crassifolia, 273, 0. NIEDENZU, F. De genere Panes Index pee Lyco Hosiano Brauns- berg. Aestatem. 1897. . De genere Dt (Pars posterior.) Arbeit. Bot. Inst. Lyc. Hosianum 1901. Malpighiaceae Americanae III. Jbid. 5: 1-61. 1914. [Pp. 48-58 an addendum to preceding publication. | Bemerkungen iiber Byrsonima coriacea (Sw.) Kunth und die iibrigen westindischen Byrsonima-Arten. Repert. Sp. Nov. 24: 249-252. 1928. [B. Spicata Sp DC. placed in synonymy; key to West Indian spp. ] REcorp, S. J., . D. MELL. Timbers of Tropical America. xviii pp. pls. 1-50. ie Haven, Connecticut. 1924. [B. crassifolia, B. spicata, 364, 65. SANDWITH, N. Y. Contributions to the flora of Tropical America: XXIV. Re- view of the species of Byrsonima occurring in British Guiana. Bull. Misc. Inf. Kew 1935: 311-316. 1935. . New and noteworthy Polypetalae from British Guiana. Jour. Arnold Arb. 24: 218-226. 1943. [Malpighiaceae, 221-223; emended key for Byrsonima spp. ] SARGENT, C. S. Notes on North American trees. VIII. Jour. Arnold Arb. 2: 164— 174. 1921. [Nomenclature of B. lucida, 166.] . Manual of the trees of North America (exclusive of Mexico). ed. 2. xxvi + 910 pp. map. Boston & New York. 1922. [B. lucida, 632, 633.] Sims, J. Malpighia lucida. Wedge-leaved Barbadoes Cherry. Bot. Mag. 51: $l. 2462. 1824. [B. lucida.] Witson, P. The vegetation of Vieques Island. Bull. N.Y. Bot. Gard. 8: 379- 410. 1917. [B. cumeata (Turcz.) P. Wilson, 394, proposed to replace B. lucida; for discussion cf. SARGENT, 1921.] ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 1972] FISHER & TOMLINSON, CORDYLINE, II 113 MORPHOLOGICAL STUDIES IN CORDYLINE (AGAVACEAE) II. VEGETATIVE MORPHOLOGY OF CORDYLINE TERMINALIS J. B. FIisHER AND P. B. TOMLINSON IN AN INTRODUCTORY PAPER in this series we described in a general way the growth habit of many species of Cordyline and the way it is re- lated to a distinctive seedling morphology (Tomlinson & Fisher, 1971). We now restrict our inquiry to C. terminalis, the most widely cultivated species, describing its vegetative morphology in detail. This is the first comprehensive study of branching pattern in any member of the Agava- ceae (sensu Hutchinson, 1959), and serves as a basis for future experi- mental work on the controlling factors of apical dominance, rhizome ini- tiation, and secondary growth in this species. The continuing work will be reported in future articles. Cordyline terminalis grows rapidly and is easily propagated clonally so that it is an ideal experimental subject. It is therefore more suitable for experimental analysis than Dracaena fra- grans, which was used initially as a subject for a reappraisal of the structure and growth of monocotyledons with secondary thickening (Tom- linson & Zimmermann, 1969; Zimmermann & Tomlinson, 1969; 1970). Cordyline terminalis is probably native to the western Pacific and was at one time widely used as a food plant throughout Oceania (Barrau, 1961). The starchy underground rhizomes, erroneously called “root tubers,” were cooked before being eaten. The leaves were used to wrap food, as a source of fiber, and even as clothing (largely ornamental). These uses have been supplanted by other food crops, but C. terminalis is still a widely cultivated ornamental. It forms, for example, the chief constituent of the living fences around the “gardens” of villagers in the highlands of New Guinea. ‘There are numerous varieties referred to in horticultural literature (e.g., Graf, 1963). These differ chiefly in the color, shape, and size of the foliage leaves. The usefulness of the shoot system of this species for experimental morphology was appreciated by Sachs (1882) and Goebel (1905) but their studies were preliminary and inconclusive. The underground rhi- zomes were described briefly by Troll (1937; p. 722). MATERIAL AND METHODS Most observations were made on large, field-grown plants up to 20 feet tall growing at Shaw’s Nursery, 57th Ave., South Miami, Florida. These specimens included several varieties of unknown origin. Seeds were collected from mature fruits and sown in pots in a greenhouse to provide plants in early stages of development. 114 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Ficures 1-6. Cordyline reemiget# eedling. 1, early stage, plumule spe cotyledon retained within seed,. <2 2/3: 2, later stage with scale leaves 0 plumule, < 11/3; 3, older seedlings with elo ngated internodes sd pe eines roots, X 11/3; 4, still older seedling with distal foliage leaves X 1 1/3, X—inter- 1972] FISHER & TOMLINSON, CORDYLINE, I 115 For general anatomical observation, free-hand sections of fresh or FAA- fixed material were stained with aqueous solutions of either toluidine blue or basic fuchsin. SEEDLING DEVELOPMENT Germination in Cordyline terminalis is hypogeous (FIGURE 1), the greater part of the cotyledon remaining within the seed, and both coty- ledon and seed remaining below soil level. This contrasts with the epigeous germination of C. australis and C. banksii which we have described and illustrated elsewhere (Tomlinson & Fisher, 1971; Figs. 1, 2). In these species only the tip of the cotyledon remains within the seed, the whole organ elongating and becoming erect; usually bringing the seed with it. Despite these early differences, late stages of seedling growth seem very similar in all species of Cordyline. The early differences are much more pronounced than those which Arnott (1962; p. 47) described in Yucca when he distinguished “hypogeous” from “epigeous” species. In both his types, the cotyledon does rise above the soil surface, although its tip is still within the seed coat, which may or may not remain buried. Later stages of C. terminalis are illustrated in Ficures 2-6. The first plumular leaves are scale-like and there is a gradual transition to progressively larger foliage leaves. This should be contrasted with C. australis, where plumular leaves, though reduced, are green and never scale-like (cf. Tom- linson & Fisher, 1971; Figs. 1-4). The lowest, often underground, inter- nodes are sometimes elongated (Ficures 2-4), as in C. australis. Ad- ventitious roots are developed on all parts of the seedling axis, but pre- dominantly at the nodes. Successive internodes are progressively wider (FicurE 5) in the manner normal for establishment growth in mono- cotyledons (Tomlinson & Zimmermann, 1966). A major departure from normal monocotyledonous seedling develop- ment is then seen with the production of a descending (positively geo- tropic) rhizome from a bud in the axil of one of the lowest leaves (Fic- URE 5, b). This repeats the peculiar pattern of seedling growth we de- scribed for other Cordyline species, in which the original seedling axis, displaced by this lateral rhizome, persists only as a pseudolateral ap- pendage on older seedlings (FIGURE 6). POST-SEEDLING DEVELOPMENT The sapling subsequently develops as an apparently single axis which consists of the ascending, foliage-bearing aérial shoot continuous below node in which secondary vascular cambium first develops; 5, base of mature seedling, rhizome bud at b; 6, later stage with seedling rhizome well developed X 22/3, the original seedling axis as an apparent appendage. Arrows indicate nodes on original seedling axis. Larger roots cut off in FicuREs 5 and 6. A dia- gram representing Ficure 6 is given in FIGURE A. Gm, | aca | U . Ficures 7 and 8. Cordyline terminalis, shoot morphology represented diagram- matically. Leafy shoots are represented in outline with each node (double line) bearing a bud. Rhizomes shown solid black with each node (white line) bearing A [ 1972] FISHER & TOMLINSON, CORDYLINE, II 117 Fig.8 a bud. Inflorescence represented by branched structures with circles; sites of former inflorescences are represented by dotted outline. Ficures 7 A-K. Branching on vegetative parts. A, sapling axis with original seedling axis displaced (arrow); B , forking of sympodium by production of two axillary branches below a terminal inflorescence; C, forking of stem by produc- ium ; , bra auceie proximally behind intact apex; ~ rhizome on lower side of a leaning "branch, branching dis- tally behind injured or decapitated apex; K, i tepics al poe Paes of aérial rhizomes nm = me ban J Se ° 3 oO ° =] os ° = 19°) + 2. Qu tas) ° mh — Oo Ba Be =] aq tp oy “3 QO ca on leafy shoots, on lower surface of a leaning branch, and at base of an erect shoot. Specimens which illustrate some of ir these diagrams are: A = FIGURE > B = Fic 14; C = Ficure 9; F = Ficures 18-20 (various ages FIGURES 3A_D. Vegetative proliferation at base of inflorescence. A, terminal inflorescence evicted by an axillary branch, the lowest bract subtending a flow- er-bearing branch (the. twist of the base of the inflorescence is shown b e arrow); B, the same, with owest flowering branch replaced by a leaf shoot ; C, e e, wi lowest flowering branch replaced by an aérial sam the thizome; D, the same, with the lowest flowering branch replaced by a leafy — plus a rhizome. Specimens which cotta these diagrams are: A = Fic- ; B = Ficures 21, 22; C = Fic 118 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 with the descending scale-bearing rhizome. This is represented diagram- matically in FicurE 7 A, with the displaced seedling axis shown as an apparently lateral appendage (arrow). Shoot construction in the adult plant eventually becomes seemingly diverse. However, this diversity is based on a few very simple rules of branching. The various ways in which these rules can be expressed are illustrated diagrammatically in FIGURES 7 and 8. These are all based on observed examples some of which are shown in Figures 9-26. The examples which correspond to a particular diagram are cross-referenced in the legends to these illustrations. The morphological rules which govern this apparent elaborateness are simple. All leaves subtend a single bud. Buds are normally strongly in- hibited and remain inconspicuous (e.g., FIGURE 14; b. They are not drawn to scale in Ficures 7 and 8). Buds may grow out as either leafy shoots or as scale-bearing rhizomes. Rhizomes may become transformed directly into leafy shoots, but never vice-versa. Three main categories of branch expression can be recognized, and these are described separately. Leafy shoots. The commonest type of branching in undisturbed shoots involves normal sympodial development by substitution of a terminal inflorescence (FIcuRE 7 H). Usually the bud in the axil of the foliage leaf immediately preceding the first bract of the inflorescence (i.e. at the base of the long peduncle (Ficure 13), grows out as the substitution shoot. The first internode of this branch elongates (FicurE 11, int.) to about the same extent as internodes on the parent axis so that there is no major articulation in the axis of the sympodium at the level of branching. Where buds from two successive nodes grow out, the sympodium itself branches and the axis is forked (Ficures 7 B; 14). In this type of substitution growth, in which there is little delay in the development of the bud, its prophyll usually has a distinct lamina (FIGURE 12, prophyll) and the transition to foliage leaves along the branch is abrupt. Thus the second or third leaf of the continuing shoot has a more or less normal shape and size (FicureE 12). Essentially the same development occurs in the outgrowth of one or more buds below a damaged or decapitated apex as shown in Ficures 7 C and 9. However, in these examples the bud has previously been subjected to a long period of inhibition. This is reflected in the outgrown shoot which has a short, scale-like prophyll without any indication of a lamina (FicurE 10, prophyll) together with a longer series of transitional leaves before the normal foliage leaves are developed (FIGURE 10). Another difference is that the first few internodes do not elongate and there is con- sequently a more pronounced articulation at the insertion of the branch. Axillary buds on the upper side of leaning branches may grow out as leafy shoots (F1curE 7 D). Buds at the base of tall erect stems may also * These differences between early (“precocious”) and late-developing branches rrespond to those seen in woody dicotyledons, where they have been distinguished as “sylleptic” and “proleptic” shoots (Tomlinson & Gill, unpublished). Ficures 9-14. Cordyline terminalis, outgrowth of buds as leafy shoots. 9, two / ~ e renewal branches below a decapitated apex, X 1/2; 10, six successive leaves a 0 branc i 2 int. ase of the branch shown in FicurE 11, X 1/6, prophyll ? 13, development of shoots from two buds below a terminal inflorescence, X “e the right-hand (lower) shoot (S) ata much ae stage 120 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 grow out in the manner of epicormic shoots (FIicuRE 7 E), presumably because they are distantly removed from an inhibiting source. This ten- dency to branch by basal suckers is responsible for the “shrubby” habit of Cordyline terminalis in contrast to the tree habit of other species, as we described in our introductory paper All these methods of shoot development are precisely as reported for other arborescent Agavaceae (Tomlinson & Zimmermann, 1969). Rhizomatous shoots. Cordyline terminalis does differ from almost all other Agavaceae because the rhizomatous shoot which is characteristic of the seedling also occurs on aérial branches. For convenience we refer to these as “aérial rhizomes,” although we appreciate that there may be semantic difficulty in referring to above-ground shoots as rhizomes. Similar aérial rhizomes are developed in Cord yline stricta. This type of shoot can, of course, also develop on the underground parts of mature plants, for example, as a branch of an existing rhizome. The seedling rhizome is therefore distinctive only because it occupies a particu- lar place in the life cycle. Aérial rhizomes are strictly positively geotropic, fleshy shoots which always originate from axillary buds, i.e. they occupy the same position as leafy shoots. However, they most usually originate from axillary buds either on the lower side of leaning branches or at the base of erect aérial shoots (FicuRE 7 K). They typically bear scale leaves which are spiro- distichous, like the foliage leaves on erect shoots (Tomlinson & Fisher, 1971, p. 467). Like the erect shoots each scale leaf of the rhizome subtends a single axillary bud which is also strongly inhibited (e.g. FicuRE 16, b). Numerous root primordia are produced (FicurEs 16-19, r) but these remain as short initials which only develop further in a moist environment, as when the rhizome is buried. The range of leaf form on aérial rhizomes is considerable and varies from scales which are otherwise typical of underground rhizomes (Ficure 15) to leaves which have a distinct lamina (Ficures 16 and 17). The larger of these resemble the transitional leaves at the base of a branch which has developed from a dormant bud (cf. Ficure 10 and Ficures 16 and 17). The apex of the rhizome itself may indeed turn upwards and develop into an erect leafy shoot, as shown in FicurEs 7 F and 18. Under these circumstances a new a€rial rhizome usually develops from an axillary bud, either on the original rhizome (Ficure 7 F) or on the leafy shoot into which the rhizome has been transformed (FicuRE 7 G). Specific examples of this are shown in FicureE 19 (early stage) and Ficure 20 (late stage). The leafy shoot into which the rhizome has become transformed may grow on to produce a typical flowering sympodium (Ficure 7 G). Underground rhizomes can also give rise directly to leafy shoots. This occurs when the axillary (or even apical) bud of an underground rhizome grows directly into a leafy shoot (FIGURE 7H). Aérial rhizomes may also “mimic” the pattern of branching of leafy shoots. Injured or decapitated aérial rhizomes may be substituted by new _Ficures 15-20. Cordyline terminalis, aérial igo s, all X 1. 15, typical ith but distal end turning erect as a leafy shoot; 19, simila I, with, a "basal “bad tb) initiating a leafy shoot, and a distal bud (b2) ini- 122 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 lateral rhizomes which arise from distal buds (FicurE 7 J). In uninjured, but long aérial rhizomes, proximal axillary buds often grow out as further rhizomes (FicurE 7 I). The “mirror-image” relationship between leafy shoots and rhizomes is suggested in comparing FicurE 7 C with 7 J and 7 E with 7 I. Unlike aérial shoots, however, in rhizomes there is no substi- tution growth associated with flowering since aérial rhizomes never become transformed directly into inflorescences (they may become so indirectly, as in FicgurE 7 G). Peduncular branching. Both leafy and rhizomatous shoots can be developed at the base of the inflorescence, because the basal bract on the inflorescence subtends a bud which can grow out as a vegetative (FIGURE 25), rather than a reproductive (flowering) branch (FicurE 24). Pro- liferation at this node may be complex, because of the development of buds subtended by the basal prophyll of the branch, whether it is vegeta- tive or reproductive (FicuRE 26). This lower bract is always brought to the upper side by a characteristic twist of the base of the inflorescence axis. In this way the spirodistichous leaf arrangement of the shoot appears to be interrupted (Ficure 8 A) although in fact it is continuous into the inflorescence. This “inflorescence” bud may grow out in a variety of ways, either as a leafy shoot (Ficures 8 B, 21 and 22), or as an aérial rhizome (Ficures 8 C; 23). Both a leafy shoot and an aérial rhizome may occur together (FicurE 8 D) although one precedes the other. It is significant that the base of the old inflorescence stalk retains only primary tissue, regardless of the amount of secondary tissue developed by the new vegetative shoot at its lowest node (FicurE 23). Thus there is no secondary connection between the vascular tissue of the main axis (below the peduncle) and the new leafy shoot (above the peduncle). This is an observation with considerable developmental significance. ANATOMY OF THE AXIS Seedling. The underground seedling axis, both at early (Ficures 3 and 4) and late (Ficures 5 and 6) stages of development, never produces a cambium. There are usually 3 to 4 primary amphivasal bundles in the first internode and up to 8 to 9 bundles in the internode at soil level. These primary bundles are arranged in a ring and often are enclosed in an endodermis-like layer of sclerenchyma. The cambium is initiated first in the region of the axis at, or just above, soil level (as at internode X in FicurE 4), and is active distally throughout the axis. The seedling rhizome is produced from an axillary bud in this region where cambial activity is initiated (FicurE 5, b). Vascular connection is tiating a new aérial rhizome; 20, older stage with massive aérial branch rhi- zome (below) developed where a parent rhizome (upper left) has turned erect as a leafy shoot (upper right); b = dormant or devel ing buds; r = root pri- mordia. A diagram corresponding to Ficures 18-20 is given in Figure 7 F. 1972] FISHER & TOMLINSON, CORDYLINE, I 123 Ficures 21-26. Cordyline terminalis, proliferating inflorescences, all ce 1. 2h, vegetative branch developing as : leafy shoot, the simple example shown di- pice seen in FicurE 25; 22, older stage of leafy shoot with abundant secondary tissue, the rc a axis below this still percind only primary tissue; 23, vegetative branch developing as a branching rhizome; 24, diagram of first node of inflorescence, I = main axis; F = flowering ceanch: a bud (B) may or may not be nents by the first leaf on the branch (prophyll) as shown; 25, diagram of first node of inflorescence, proliferating vegetatively by replacement of flowering branch by vegetative bra inch (V), either a leafy shoot ra rhizome; 26, diagram showing development of a flowering branc (F) and roi vegetative branches (V) at the first node of a proliferating inflorescence, the two latter from buds both subtended by the prophyll; this probably ac- teas for ik proliferation of several branches in Ficures 22 and 23, all seem- ingly from the same node. Habit ne ay corresponding to these specimens are? 1: 22 = Ficure 8 B; 23 = Ficure 8 124 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 established mainly by secondary tissues derived from the cambium which is continuous between both organs. At the soil surface (FicurE 6) the anatomy of the axis is often intermediate between that of the rhizome and stem (as described below). If the junction of rhizome and stem is above the soil surface, the proximal region of the rhizome is stem-like. In older seedlings the junction of leafy shoot and rhizome is neither morphologically nor anatomically obvious since the tissue of the axis is secondary except for the few primary bundles of the original seedling axis. Adult axes. The anatomy of the three kinds of axis (leafy shoot; rhizome; inflorescence) is dissimilar and each may be described separately. The most distinctive feature of the inflorescence axis is that unlike the vegetative axes it never develops secondary vascular tissues. (a) Leary sHoor (Ficures 27 and 28). In a mature leafy shoot with well-developed secondary tissues there is usually a sharp boundary between the primary zone, with thin-walled, unlignified and irregularly-arranged ground parenchyma cells, and the secondary zone, in which the ground tissues are radially arranged (reflecting their origin from the cambium) and with thick, lignified walls. There is little distinction between primary and secondary bundles because both are amphivasal. The sharpness of the boundary depends also on the amount of secondary establishment growth which the shoot may have undergone. The situation in Cordyline seems quite comparable to that in Dracaena which has been described in detail elsewhere (Zimmermann & Tomlinson, 1970). The outer cortex is chlorenchymatous and the amount of secondary cortical tissue (if any) derived from the vascular cambium is difficult to establish, because cambial derivatives in this direction soon lose their radial alignment. Periderm (not developed in the stem from which FIGURE 27 was taken) is well developed in older parts of the axis. (b) Rxtzome (Ficures 29 and 30). Both underground and aérial rhizomes are similar in their internal structure, differing only in that the aérial rhizomes produce more periderm tissues and have a chlorenchy- matous cortex. Secondary tissues are well developed in older rhizomes, but the vascular bundles are very diffuse (FicurE 29). The boundary between primary and secondary tissue is indistinct because ground paren- chyma in both regions remains unlignified. The secondary tissue, however, shows clear radial seriation of this parenchyma. Vascular bundles are again amphivasal, the secondary bundles being narrower than correspon- ding bundles in the leafy shoots (cf. Ficures 28 and 30). The vascular cambium can be recognized much closer to the shoot apex in the rhizome compared with the leafy shoot, and this probably accounts for some of the differences between the two types of organ. (c) INFLORESCENCE AXIS. This is strikingly different from the axis of the leafy shoot, of which it is a morphological extension, in that a vascular cambium is never developed (Ficure 31). Most of the vascular bundles are collateral (FicurE 32); they are narrower and more compact 1972] FISHER & TOMLINSON, CORDYLINE, II 125 °° ¢. FicurREs 27-32. Cordyline terminalis, camera lucida drawings illustrating i i leafy shoot, 9, mature aérial rhizome, boundary between primary and secondary tissue indistinct; 30, detail of an amphivasal secondary vascular bundle; 31, mature inflorescence axis, no secondary vascular tissue; 32, detail of a collateral primary vascular bundle. 126 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 towards the periphery. Ground parenchyma cells have thickened and lignified walls. When a vegetative axis arises on the inflorescence, as shown in FIGURES 22 and 23, the peduncle persists and produces a periderm, but never a vascular cambium. Inflorescences of comparable age, but without a vegeta- tive branch, shrivel and die. DISCUSSION These observations on naturally grown specimens show that in Cordy- line terminalis there is strict apical dominance of axillary buds. These buds, however, do have the potential for further extensive development. They may grow out as leafy shoots under circumstances which are quite comparable for other woody monocotyledons (Tomlinson & Zimmermann, 1969) and indeed seemingly for woody dicotyledons as well (Phillips, 1969; Smith & Wareing, 1964). A unique property of these buds, however, compared with those in other genera of flowering plants is that they can develop as positively geotropic rhizomes in ways which “mimic” their development as leafy shoots. The rhizomes have the additional peculiarity of undergoing metamorphosis so that they can revert to negatively geotropic leafy shoots. It is therefore of interest to seek experimental evidence for possible hormonal mechanisms which control this divergent behavior. Future articles will present these experimental results which greatly clarify the situation and add to our knowledge of factors controlling shoot morphology in monocotyledons. Additional information on initiation of secondary vascular tissues will attempt to explain the inability of the inflorescence axis to develop a vascular cambium even when it has proliferated a shoot with this ability. SUMMARY Cordyline terminalis has hypogeous seedling germination, unlike the epigeous germination of closely related species. Late seedling stages, however, show features diagnostic for the genus as a whole, with the de- velopment of a lateral rhizome which displaces the original seedling axis. Axillary buds are subjected to strong apical dominance, but have the potential to grow out as either negatively geotropic leafy shoots or posi- tively geotropic, scale-bearing rhizomes. The latter can be transformed directly into the former. These different kinds of shoot expression seen on naturally grown individuals are described in detail. The value of this species as an experimental subject is emphasized. In outlining the dis- tribution of secondary tissues in the various parts of the plant, the inability of the inflorescence axis to develop a vascular cambium is pointed out, a peculiarity which is particularly striking in parts which are below a vegetative proliferation. These observations provide a background for future experimental work. 1972] FISHER & TOMLINSON, CORDYLINE, II 127 ACKNOWLEDGMENTS This study is part of a larger survey of the systematic and develop- mental anatomy of monocotyledons supported by N.S.F. grant GB- 5762X, P. B. Tomlinson, Principal Investigator. We are indebted to the proprietor of Shaw’s Nursery for access to specimens. Habit illustrations are by Miss Priscilla Fawcett, Botanical Illustrator at Fairchild Tropical Garden. Support for J. B. Fisher was provided by a fellowship from the Cabot Foundation, Harvard University. LITERATURE CITED Arnott, H. J. ae The seed, germination and seedling of Yucca. Univ. Calif. Pu bl. Bot. 35: 1-164. BARRAU, J. 1961. sibdhcindus agriculture in Polynesia and Micronesia. Bishop Museum Bull. 223. pp. 94. Gorse, K. 1905. Morphologische und biologische Bemerkungen 16. Die Knol- len der Dioscoreen und die Wurzeltrager der Oe oe Organe, welche zwischen Wurzeln und Sprossen stehen. Flora (Jena) 95 2 Grar, A. B. 1963. Exotica. ed. 3. Roehrs Co. Rutherford, pi Jers HuTcHINsoN, J. 1959. The families of flowering plants. Vol. II. oe dons. ed. 2. Clarendon Press. Oxfor Puiiuips, I. D. J. 1969. Apical dominance. In: M. B. Wrrxrys, ed. The physiology of plant growth and development. pp. 163-202. McGraw-Hill, London. SACHS, J. 1882. — und Form der Pflanzenorgane. Arbeit. Bot. Inst. Wirz- urg 2: 452-4 pga H., & P. . Hie 1964. Gravimorphism in trees. 2. The effect of vity on bud-break in osier willow. Ann. Bot. (Lond.) n.s. 28: 283-295. Toa ae P. B., & J. B. FisHer. 1971. Morphological studies in Cordyline (Agavaceae). I. Habit and general morphology. Jour. Arnold Arb. 52: 59-478. & M. H. Zimmermann, 1966. Anatomy of the palm Rhapis excelsa. II. Juvenile ge eas Arnold Arb. 47: 301-312. aes te 9. Vascular anatomy of monocotyledons with secondary wth — an bc Ma Ibid, 50: 159-179. Rien W. 1937. Vergleichende Morphologie der gree Pflanzen. Bd. 1. Vegetationsorgane. Teil 1. Gebriider Borntraeger. Ber. ZIMMERMANN, M. H. & P. B. ToMLINSON. 1969. The i system in the axis of Dracaena fragrans (Agavaceae). 1. a" and development of primary strands. Jour. Arnold Arb. 50: 370- 9 The vascular system in | axis of Dracaena fragrans. 2. Distribution and development of secondary vascular tissue. Ibid. 51: 478-491. DEPARTMENT wh Pov gen reas herr OxnI0 UNIVE PETE ATHENS, ia 45701 iaprraaretnee 01366 128 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 NOTES ON ASIATIC-POLYNESIAN-AUSTRALIAN SPECIES OF ERYTHRINA, II B. A. Kruxorr ! My FIRST PAPER on this subject was published in 1939 (Preliminary Notes on Asiatic-Polynesian Species of Erythrina. Jour. Arnold Arb. 20: 225-233. 1939) without the benefit of having studied specimens in Euro- pean and Asiatic herbaria because of the war. Previously there was no compact treatment of the Asiatic-Polynesian-Australian species. Since that time I have examined specimens in several major European herbaria (BM, BR, E, G, K, L, P, S, U, Z) and as a result have made considerable progress in the understanding of several species. Six species (namely E. glauca Willdenow, E. boninensis Tuyama, E. phiebocarpa F. M. Bailey, E. ros- trata Ridley, E. sandwicensis Degener, and E. stipitata Merrill) are here reduced to synonymy for the first time. A conspectus of the species of Erythrina worldwide is in the course of preparation; in it I plan to give complete synonymy and type localities of valid species and synonyms, cite the type collections, give carefully compiled distribution, and discuss habit and habitat. In the present paper I am giving largely the information which I consider important or inter- esting and which it is not possible to include in the conspectus. Complete synonymy is given only for E. variegata L., but I am citing, of course, under other species those which are reduced to synonymy for the first time. Specimens are cited only for E. microcarpa Koorders & Valeton, E. euodiphylla Hasskarl, E. merrilliana Krukoff, and EF. insularis F. M. Bailey which are very poorly represented in herbaria. Of 13 species reviewed in this paper the chromosome numbers were published for seven (Phytologia 19: 162, 163. 1969; Taxon 18(3): 314. 1969; and Ann, Missouri Bot. Gard. 56: 474. 1969). No counts have been published for Erythrina microcarpa, E. resupinata, E. euodiphylla, E. merrilliana, and E. insularis. A 2n number of 42 chromosomes was found in all Asiatic-Polynesian-Australian species sampled. Three of the arboreous species under review (namely E. arborescens, E. suberosa, and E. stricta) and the herbaceous E. resupinata probably will prove to be frost tolerant in the Los Angeles area. It is kept in mind that by a formal proclamation on March 7, 1966, Erythrina was selected as the “Los Angeles official tree.” E. resupinata will doubtless prove to be an excellent addition to rock gardens. I wish to express obligation to Prof. Dr. C. G. G. J. van Steenis for his suggestions and help in obtaining certain critical type material. I am es- * Consulting Botanist of Merck Sharp & Dohme Research Laboratories, Rahway, New Jersey. 1972] KRUKOFF, NOTES ON ERYTHRINA, II 129 pecially indebted to Mr. R. Barneby for his assistance during the progress of this work, and for his cooperation in reading the manuscript. KEY TO THE SPECIES ? 1. Keel petals connate; seeds not red; rachises, ner oees and leaflets (at least on petiolules and costa beneath) not stellate pubesce . Standard long—stipitate, subrotund-rhombic; pods prides seeds — umber to blackish with black markings. .................. 1. E. bd 2. Standard and seeds not as above; pods follicular REN HAES ay in E. arborescens). 3. Keel petals subequal to wings; pods seedless and et in the lower AOU |. @h0. ges, eOUISoT: SA, Buea: 2. E. subumbrans. 3. Keel petals conspicuously longer than wings; pods bearing seeds throughout Pe Ne TN Pave can aaaitt 5. E. microcarpa, 4. Plants not native to Jav. 5. Perennial herbs ae thickened roots endemic to northern OG a ee eee a 7. E. resupinata. 5. Large trees. 6. Pods ligneous; seeds shiny black; leaflets cago ceriferous beneath; keel petals 1 1/3—2 times creat than wings. .... Peta SP Our A BARNS SE 8 2 x bee sakie. Pods follicular, chartaceous; seeds solos to dark brown or sooty; leaflets ceriferous on both margins of veinlets be- neath or intricately ena “aiateie keel petals 2 1/2 or more times longer than w re Calyx campanulate; Rahicts without numerous white a s si SUES WI TEC 6s ee ee os suberosa. if rene subspathaceous; branchlets heavily armed with é spines when fresh. ............... 6. E. stricta, 1. Keel petals separate es subequal to wings (flowers of E. imsularis not known); seeds scarlet-red or brownish red; rachises, pedicels, and leaflets (at least on petiolules and costa beneath) stellate pubescent. 8. Seeds large (+ 19 mm. long and + 11 mm. broad), brownish sc pods with prominent veining; standard red. ............-... 8. ariegata. . Seeds and pods not as above. Species keyed by geographic areas 9a, 9b, et co etc. 9a. Endemic to New Guinea; seeds scarlet with black hilum and with a black line extending from the hilum for approximately 2 mm. toward the chalazal end, 9-10 mm. long and 6-7 mm. nae standard bright OFAN ro cee. ie Pee aie ee . E. merrilliana. . Endemic to Australia. 10. Seeds scarlet with white hilum, uniformly ee + 13 mm. long and + 7 som, Drobtiag: ces oo snc ss te 3. E. vespertilio. 10. “pain agi? with black hilum and with a ae line extending um for approximately 2 mm. toward the chalazal why 9. ant mm. long and 7-8 mm. broad; known only from the type collection from Turtle Island. ...... 12.'R, 3 N=) ao ® Characters indicated by italics do not occur in other species. 130 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 9c. Endemic to Tahiti and Hawaii; standard orange; + 17 m lone-end-8:mam; brenda: cc. wage ertinss eked sak. bs i ‘alanis 9d. Eocene to Java; all petals green; leaves fetid when fading. cies sao peg pee ie 5 dys ia nae ee ean . E. euodiphylia. 1. Erythrina fusca Loureiro, Fl. Cochinchinensis 427. 1790. y Erythrina glauca Willdenow, Ges. Nat. Freunde Berlin Neue Schr. 3: 428. 1801. This species is a very large tree that thrives in a variety of conditions, but seems to prefer lowlands (seashores, swamps without outlets, low overflow lands, river banks, shores of lakes, etc.). In low swamp grounds where Erythrina fusca reaches especially huge dimensions, it often occurs in pure stands. This is the most widespread species in the genus and the only one which occurs on three continents. It is undoubtedly dispersed by ocean currents. In the Old World Erythrina fusca is found in a very extensive area, from Madagascar and the Mascarene Islands in the west; India, Burma, the Philippines and Carolines in the north; Samoa and Tonga Islands in the east; and New Caledonia, Papua, and Timor in the south. In the New World it is found in the West Indies and on the Continent from Belize and Guatemala in the north to and including Peruvian, Brazilian, and Bolivian Amazonia in the south. I did not place Erythrina glauca in synonymy in my monograph of the American species (Brittonia 3: 205-337. 1939) since before doing this I wanted to see E. fusca, as it appears in the Old World, in the field. There are no differences between E. fusca as it occurs in Asia, Polynesia, and Africa and as it is represented in the New World, in fact even its habitat is the same. According to a private communication from Prof. E. A. Bell and John Romeo of the University of Texas at Austin the alkaloids and amino acid patterns of seeds of this species, as it is rep- resented in America and Asia, are very similar. 2, Erythrina subumbrans (Hasskarl) Merrill in Philip. Jour. Sci. Bot. 5S: 433. 1910. Plants of this species are medium size trees confined to low and middle elevations in light forests, open places, and secondary forests. The species is found from southern India, Burma, Thailand, and Viet- nam to and including Java, Borneo, and the Philippines. 3. Erythrina arborescens Roxburgh, Fl. Ind. 3: 256. 1832. Erythrina moori Todaro, Hort. Bot. Panorm. 2: 7. pl. 26. 1879. This species is composed of small to medium size trees, found at ele- vations between 1200 and 2100 meters. Leaves usually appear in May and fall off as soon as the tree has flowered in October. 1972] KRUKOFF, NOTES ON ERYTHRINA, II 131 Erythrina arborescens is found in India, Nepal, Burma, and China. In China it is the only species which is largely found to the north of the Tropic of Cancer (Sikang, Szechwan, Kweichow, and Yunnan). In my previous paper (l.c.) I reported: “A doubtful species Erythrina moori was described on the basis of a cultivated plant of unknown origin and said probably to be from India. Examination of the original de- scription and the plate establishes definitely the fact that it is an Asiatic rather than an American or African species. Examination of the type, if it exists, is necessary to check my conclusion that Erythrina moori is probably conspecific with E. arborescens.” Since that time I have become familiar with all known Asiatic and African species and there is no doubt in my mind that E. moori is a syn- onym of EF. arborescens. The connate keel petals about 21/2 times longer than the wings, the ligneous pods with black seeds as shown on the excellent plate, are sufficient to establish its identity. In fact the black seeds are otherwise found only in three unrelated species of Ameri- can Leptorhizae. 4. Erythrina suberosa Roxburgh, Fl. Ind. 3: 253. 1832. This is a large deciduous tree confined to India (northern, central and southern), Nepal, Bhutan, Burma, Thailand, and Vietnam, ascending to 900 meters. Without field studies and without seeing specimens in herbaria in India, it is not possible to decide whether Erythrina glabrescens (Prain) R. N Parker (Indian Forester 46: 647. 1920) should be placed in synonymy here or retained as a variety. 5. Erythrina microcarpa Koorders & Valeton, Booms. van Java 2: 9 61. 1895, Erythrina stipitata Merrill, Philip. Jour. Sci. Bot. 5: 112. 1910. Erythrina microcarpa is a large tree, to 18 meters high, found in east Java along and near beaches on low elevations up to 900 meters. Java: Koorders Herbarium 46 (lvs) (L), 49 aoe ae ©, 50 (lvs) (L), 51 (Ivs) (t), 65 (lvs, firs) (K, L), 66 (lvs, firs, frts, seeds) (K, L-lectotype), 71 (firs) (1), 72 (lvs, firs, frts, seeds) (x, L), 12785 (lvs, imm. flrs) @, 14598 (firs, imm. di: (K, L), 14599 (lvs, firs, imm. map &: Herb. Hort. Bot. Bogor 9620 (mixed coll., lvs-E. subumbrans, firs-E. rocarpa (K, L); coll. undesign. II/0-16 (1/VIII) (Ivs, firs, frts, ‘seeds) (a): Prvidinon Java 837 (firs) (x), 879 (firs, frts) (L). Philippines: Lubang Island, near the town of Lubang, in open lan at sea level, E. D. Merrill 958 (Ny-isotype of E. oie s.n. (lvs, firs, frts) (kK) with a note on the label “found without a ticket”; on the same sheet is mounted A. D. Elmer 7132 (Jan. eo) (Ivs, an) ia Talend Leyte, prov. Leyte, loc. Palo — which is plainly E In the protologue of the species, Koorders & Valeton did not designate the type. It is proposed here to designate Koorders Herb, 66 (t) as lecto- type. 132 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Examination of the two collections from the Philippines shows clearly that E. stipitata is conspecific with Erythrina microcarpa It will be noted that abundant material of the species from Java was distributed by S. H. Koorders in Europe only to Leiden and Kew; also that apparently only two sheets of E. microcarpa from the Philippines are available, as far as European and American herbaria are concerned. In the preparation of this paper I did not consult Asiatic herbaria. The species is related to Erythrina suberosa and its relatives. 6. Erythrina stricta Roxburgh, Fl. Ind. 3: 251. 1832. This species is a large deciduous tree usually aphyllous at anthesis. It is very spiny even in the context of its genus. The spines are white, a unique feature in the entire genus Erythrina stricta is found in India (northern, central and southern), Nepal, Burma, Thailand (where it is common in deciduous forests), Viet- nam and China (eastern Tibet, southern Yunnan, and Kwangsi tH. 7. Erythrina resupinata Roxburgh, FI. Ind. 3: 257. 1832. This is an herb with a perennial thickened root which is in flower in April, at which time no part of the plant is visible but the raceme; the stem is formed later and is from 10 to 13 centimeters high. This species is endemic to northern India. Fruits and seeds of Erythrina resupinata were not seen by me in 1939. As is the case with the related species (namely E. suberosa, E. microcarpa, and also E. stricta), pods of it are follicular, chartaceous, not at all or slightly constricted between seeds, which are isabelline to dark brown or sooty. Seeds of the four species just mentioned are very similar. Those of E. resupinata are probably the smallest (8 mm. long and 6 mm. broad), followed by E. microcarpa; those of E. suberosa and E. stricta are larger, often 11 mm. long and 7 mm. broad. Erythrina resupinata is very poorly represented in American and Euro- pean herbaria. Four American species of Erythrina (namely E. leptorhiza DC., E. hor- rida DC., and E. montana Rose & Standley confined to the higher ele- vations in Mexico, and £. herbacea L., as it occurs in the northern part of its range in the U.S.A.), as well as three species from Southwest Africa (E. baumii Harms, E. mendesii Torre and E. pygmaea Torre), although they are not related, are also herbaceous plants with thickened roots, Up to now no one appears to have studied the thickened roots of these species chemically. Such studies probably will prove to be of considerable interest. This plant will likely be an excellent addition to rock gardens. 8. Erythrina variegata L. Herb. Amboin. 10. 1754; Amoen. Acad. 4: 122. 1759, based on Gelala alba Rumphius, Herb. Amboin. 2: 234, $. 77, 1780. 1972] KRUKOFF, NOTES ON ERYTHRINA, II 133 Erythrina corallodendrum L. var. orientalis L. Sp. Pl. 706. 1753, typified by Mouricou (Rheede, Hort. Malab. 6: 13. ¢. 7. 1686). In the second edition of the Species Plantarum Linnaeus identified Rumphius’s Gelala litorea Rumph. 2: 230. ¢. 76. with Rheede’s plant and cited it : synonymy. Tetradapa javanorum Osbeck, Dagbock Ostind. Resa 93. a Erythrina picta L. Sp. Pl. ed. 2. 993. 1763, based on ites alba Rumphius Herb. Amboin. 2: 234. ¢. 77. 1750. Erythrina orientalis (L.) Murr. Comm. il ad 8: 35. &, 7. 1787. Erythrina indica Lamarck, Encycl. 2: 391. 1788 Erythrina corallodendrum Loureiro, FI. Cochinchinensis 427. 1790. Erythrina spathacea DeCandolle, Prodr. 2: 412. 1825 Erythrina divaricata DeCandolle, Ibid. 414. Erythrina lithosperma Blume, Cat. Gew. Buitenzorg 92. 1823 (nomen nudum) ; Hasskarl, Flora 25. Beibl. II. 77. 1825, descr. ~ te Jav. Rar. 381. 1848 (not E. lithosperma Miquel, Fl. Ind. Bat. 1: 209. Erythrina loureiri G. Don, Gen. Syst. 2: 372. 1832 (used on E. coralloden- drum Loureiro). ebainkate carnea Blanco, FI. Filip. 564. 1837; ve 2. 393; 1845; ed. 3, 2.359, 217. 1879; not E. carnea Ait. Hort. Kew Jae. Cabicdys indicus (Lamarck) Walpers, Flora 36: 148, 1853. Chirocalyx pictus (L.) Walpers, Ibid. Chirocalyx candolleanus Walpers, Ibid. Chirocalyx divaricatus (DeCandolle) Walpers, Jbid. Erythrina parcellii Bull, Gard. Chron. n.s, 1874 (II): 392, fig. 82. 1874. Erythrina alba Cogniaux & Marchal, Plantes 4 feuillage ornamental 2: pi. 56. 1874. Erythrina marmorata Veitch ex Planchon, Fl. des Serres 23: 21 (pl. s.m.). 1880. Corallodendron divaricatum (DeCandolle) Kuntze, Rev. Gen. PI. 1: 172. 1891. Corallodendron spathaceum (DeCandolle) Kuntze, /bid. 173. 1891. Corallodendron orientale (L.) Kuntze, Ibid. 172. 1891. Erythrina phlebocarpa F. M. Bailey, Queensl. Agr. Jour. 1: 368, 1897. Erythrina variegata L. var. orientalis (L.) Merrill, Interp. Herb. Amboin. 276. 1917, and in Trans. Am. Phil. Soc. 24: 208. 1935, based on Erythrina coral- lodendru m L, var. orientalis L. Sp. Pl. 706. 1753. Erythrina mysorensis Gamble, Kew Bull. 1919: 222. sd Erythrina rostrata Ridley, Flora Malay Penins. 1: 580. 1922. Erythrina so Pra var. alba Blatt. & Mill. in om Bombay Nat. Hist. Soc, 33: Erythrina pie Tuyama, Bot. Mag. Tokyo 49: 373. 193 Erythrina variegata L. var. orientalis (L.) Merrill, forma ae (L) Mahesh- wari, Bull. Bot. Surv. India 3: 47. 1961. Erythrina variegata L. var. orientalis (L.) Merrill, forma parcelli (Hort. ex Bull) Maheshwari, bid. Erythrina variegata L. var. orientalis (L.) Merrill, forma marmorata (Hort. ex Veitch) Maheshwari, bid. Erythrina variegata L. var. orientalis (L.) Merrill, forma mysorensis (Gamble) Maheshwari, /bid. Erythrina variegata ae var. orientalis (L.) Merrill, forma orientalis (L.) Maheshwari, /bid. Erythrina variegata hy var, orientalis (L.) forma alba (Blatt. & Mill.) Ma- heshwari, Ibid. 134 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Erythrina variegata L., E. corallodendrum L. var. orientalis L., and E. picta L. were based on plants from Herb. Amboin.; the type of Tetradapa javanorum is from Java; that of E. carnea Blanco from the Philippines; that of E. corallodendrum Loureiro from Vietnam; that of Z. indica from India; that of E. spathacea from Santo Domingo. E. divaricata was based on a Sessé & Mocifio plate and was said to be a Mexican plant, but was doubtless drawn from a cultivated plant, probably collected in the West Indies. The type of E. lithosperma is from the Botanic Gardens, Bogor, Java; that of E. parcellii and of E. alba a cultivated (?) plant “from South Sea Island’’; that of E. marmorata a plant of unknown origin “des jles de la mer du Sud”; that of E. phlebocarpa a plant cultivated in New South Wales, Australia; that of E. mysorensis from Mysore, South India; that of E. rostrata from Kedah, Malaya; and that of E. boninensis from the Island of Bonin. DisTRIBUTION. Specimens were seen from the following countries: Tanzania (Tanganyika: Dar es Salaam & Mafia Island, Zanzibar), Mada- gascar, Maldive Islands, Ceylon, India (west Bengal, Mysore, Madras), Burma, Andamans, Thailand, Cambodia, Laos, Vietnam, China (Yunnan, Kwangtung, Hainan), Hongkong, Formosa, Ryukyu Islands, Malaya, Su- matra, Java, Timor, Borneo, Philippines, Celebes, Moluccas, New Guinea, New Britain, Solomon Islands, New Caledonia, New Hebrides, Fiji, Samoa, Tonga, Marianas, Carolines, Society Islands, Tahiti, and the Marquesas. Hasit and Hasirat. This species is a large tree of the lowlands, being especially common along the sea coasts. It is now widely cultivated in the tropics; it was an early introduction to the New World and was even described twice as a new species, native to America, by DeCandolle. n my previous paper (l.c.) I accepted the name E. variegata L. var. orientalis (L.) Merrill for this species, as suggested by E. D. Merrill. However, the arguments of A. A. Bullock (Kew Bull, 20: 294. 1966) are convincing and I now accept Erythrina variegata L. as the valid name for the species. Erythrina parcellii and E. alba are obviously a form of the species with variegated leaves; I examined the specimen at Kew on which they were based. The basis of E. marmorata is a sterile specimen of uncertain origin col- lected “des iles de la mer du Sud.” It is obviously a horticultural form of the species with variegated (white-spotted): leaves. Reexamination of the type-material of E. phlebocarpa, now mounted on three sheets (BRI) carefully annotated by S. T. Blake, shows that the pro- tologue of this species was derived from elements of two species. The fruiting element (BRI 007963) represents E. variegata L., but the leaves associated with the pod by Bailey represent an African species, E. lysts- temon Hutchinson. No flowers survive with the type-series as preserved at Brisbane, but the ostensible cotype at Kew has flowers of E. variegata mounted with foliage of E. lysistemon, whereas that at the British Mu- 1972] KRUKOFF, NOTES ON ERYTHRINA, II 135 seum (Nat. Hist.) has flowers and one pod of £. variegata mounted with leaves of E. lysistemon. In order to settle the synonymy, it is proposed here to designate the fruiting element, that part of the mixture which suggested the epithet phlebocarpa (Bri 007963), as lectotype. The name then falls into the synonymy of E. variegata. It should be noticed that the flowers preserved at the British Museum and Kew cannot in any circumstances be considered part of the type- series for they disagree seriously with the original description. Bailey described the calyx of Erythrina phlebocarpa as ‘“campanulate, oblique at the top,” whereas the calyx of E. variegata is spathaceous, deeply split ventrally. It seems likely, therefore, that Bailey actually saw and de- scribed flowers of E. lysistemon, the species to which belong the leaves preserved at Brisbane, and that the flowers associated with leaves of E. Lysistemon in the English herbaria were added as an afterthought, none of them, interestingly enough, being retained in Australia. My remarks on E. phlebocarpa (1939, p. 233) were based on study of the flowerless plants with associated pod and seeds now at Brisbane and flowers of an Erythrina, now recognized as the African E. lysistemon, which is ex- tensively cultivated in Australian gardens. The following material of the type collection of EZ. phlebocarpa was ex- amined: tralia. New SoutH WatEs: Queensland Herb. 007961, 007962, 007963 (3 sheets deposited at Queensland Herbarium) (type coll. of E. phlebocarpa, lvs of E, lysistemon Hutchinson, and frts and seeds of E. variegata L.), the Kew sheet - of the same collection consists of leaves of E. lysistemon and flowers of E. varie- gata L., whereas the British Museum sheet consists of leaves of E. lysistemon and one fruit and flowers of E. variegata L. The following material of E. lysistemon from Australia was examined: New Souru Wates: Sabine Helms 1447 (firs & mature lvs) (cult.) (us), F. A. Rodway 813 (firs) (Nowra, extensively planted in district, does not form fruit in this district — flowers fall as soon as they begin to grow) (K), 813A (mature lvs) (x), Kaspiev 396 (firs) (G), 1308, 1318 (firs) (Z). QUEENSLAND: C. T. White 9073 (firs & mature lvs) (cult. Brisbane Bot. Gard.), Brass 33486 (firs & mature lvs) (Bromfield Crater, Atherton Tableland) (x), Kaspiev 883 (firs) (z). Lorp Howe Istanp: M. M. J. v. Balgooy 1109 (t) (firs) (treelet c. 2 m., land- slide of sand dune). New Zealand. AUCKLAND: W. R. Sykes 703/64 (firs & young lvs, 30/9-1964) (Parnell Rose Garden) (cas), 787/64 (firs & young lvs, 2/10-1964) (Mt. Albert, exposed bank) (cas), 697/64 (firs & mature lvs, 29/9-1964) (Auckland Domain) (cas), 700/64 (fl 30/9-1964) (Parnell Rose Garden) (cas), 812/64 (frts, seeds & mature Ivs, 3/10-1964) (Garden) (cas); Schweinfurth 1075 (firs & lvs) (North Island, North Auckland) (™). According to W. R. Sykes this species is known in cultivation - — Zealand as “Erythrina indica Lam.” Seeds on his specimen are typical of E. lysistemon with black hilum. 136 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 On the basis of the original description of Erythrina mysorensis (“calyx spathaceous. . . . carinae petalae liberae’”’) I suggested in my previous paper (l.c.) that E. mysorensis might be merely a related variety or a form of the common and widespread E. variegata. In my previous paper (l.c.), I suggested that the doubtful E. rostrata may prove to be a synonym of the species. With the help of B. Verdcourt, I recently located the isotype of this species at Kew (Ridley 15782, April, 1890, from Kedah, Pulau Adang Seashore, Malaya). The specimen con- sists of young and mature leaves, 3 pods, and seeds. It manifestly belongs with the species. I have examined the holotype of E. boninensis (Takasi Tuyama s.n. (tr) April 14, 1934, from the Island of Bonin, Chichijima, Omura). It consists of the upper portions of 4 rachises with flowers and a considerable number of loose flowers in poor condition. The flowers are rather small for the species but obviously belong here. I also examined 2 other sterile collections of Erythrina from Bonin H. Hattori s.n. (2/8-1905) (T1) and Takasi Tuyama s.n. (Dec. 24, 1968). They also are obviously of this species There seems to be no need to distinguish as species different forms of Erythrina variegata, and there is no justification for separation at the varietal level of Gelala alba Rumphius and Gelala litorea Rumphius. As early as 1917 Merrill stated: “Strictly, the specific name variegata should be adopted to include not only the form with the variegated leaves, but also the much commoner and widely distributed form with uniformly green leaves.” It should be kept in mind, too, that forms with white flowers and/or variegated leaves occur also in other species (E. berteroana, E. falcata, and others) Inasmuch as various species of Erythrina (especially E. crista-galli L., E. lysistemon Hutchinson, and E. variegata L.) were cultivated in green- houses or in the open in Europe and the U.S., as well as on other continents for many years, there are many horticultural forms, some of which are referred to in the literature. I mentioned some of the horticultural forms of E. crista-galli in my monograph of the American species (Brittonia 3: 231. 1939). C. A. Backer in 1911 (Schoolflora voor Java 1: 360) and then C. A. Backer & R. C. Bakhuizen van den Brink, Jr., in 1963 (Flora of Java, 1 628) suggested that Erythrina crassifolia is probably a hybrid between what is now known as E. variegata and E. subumbrans. In the second publication, they give a rather complete description of E. crassifolia. It is cultivated locally in Java, its pods are usually poorly developed, and seeds are deformed or quite abortive, which is usually the case with Erythrina hybrids. I accept the suggestion that E. crassifolia is a hybrid without reservation, but I suggest that it is probably a hybrid between E. variegata and E, fusca (see for example S. M. Popta 00527/96 (1) which has leaves closely resembling those of the latter species. Among other specimens of E. crassifolia | have examined Koord. Herb. 76 (L) and Herb. Hort. Bot. Bog. Kds. 32440 (1) marked as “n. sp.” by Koorders. 1972] KRUKOFF, NOTES ON ERYTHRINA, I 137 9. Erythrina tahitensis Nadeau, Enum. Pl. Tahiti 80. 1873. Erythrina sandwicensis Degener, Fl. Hawaii. 2: Fam. 169c. 1932. This is a large tree, often found on arid rocky grounds including lava WS. The species is found in Tahiti and Hawaii. In a previous paper (l.c.) I stated: “I have seen Nadeau 499 (type of E. tahitensis) deposited at Geneva. It consists of a single inflorescence with small flower buds, two flowers, one pod, one seed and no leaflets. From this available material it is impossible to decide whether or not this plant is specifically distinct from the plant native to Hawaii.” Recently I examined in Paris two additional sheets of this species, as it is represented in Tahiti. The holotype (Nadeau 499) consists of a fruit- ing inflorescence with 3 pods and seed, also one flowering inflorescence with immature flowers. Another collection (M. Pancher 801 (1860), “sur montagne élevée des Tahiti) consists of a branchlet with two mature leaves and one immature, one flowering inflorescence, also one pod, an one flower in good condition. It is the basis of the name E. montana Forst. (Cuzent, G., Iles Tahiti 230. 1860). I concluded that the Hawaiian plant is conspecific with the plant from Tahiti. It was suggested that the plant is probably extinct in Tahiti. How- ever, taking into account its habitat, this seems to be most unlikely. The type collection is from “‘precipicie, de Tautana, 800 m”; for the habitat of the Pancher collection see above. Recently Brother Alain told me of his unsuccessful attempt to collect the very rare E. buchii Urban in the Dominican Republic where he saw it in flower on a precipice inaccessible without special climbing equipment. E. tahitensis is in flower in Tahiti in August, when collection is easier, as its striking orange flowers are visible at considerable distances. The native name for this species in Tahiti is ‘‘atae oviri.” 10. Erythrina euodiphylla Hasskarl, Hort. Bogor. 178. 1858. This is a large tree of the lowlands, including savannas with high grasses and scattered trees and shrubs. It is endemic to eastern Java and neighboring Madura and Bali. This species is represented in herbaria of Europe and U.S.A. only by the two specimens cited below: Eastern Java: M. Jacobs 4886 (L), H. Zollinger 1440 (G). This species is unique in the genus, having all petals green and leaves fetid when fading. The species is immediately distinguished from Ery- thrina fusca, E. macrocarpa, and E. subumbrans which also occur in Java, as its rachises, pedicels, and leaflets (at least on petiolules and costa be- neath when young) are stellate pubescent. The only other species which is found in Java, E. variegata, is distinguished by the prominent veining on the pod and standards. The species is related to E. stricta. 138 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 53 I have not seen fruits and seeds of Erythrina euodiphylla as they are not available in European and America herbaria. However, immature fruit and seeds were described by Hasskarl. 11. Erythrina merrilliana Krukoff, Jour. Arnold Arb. 20: 227. 1939. This is a striking tree with bright orange flowers which reaches the canopy in the rain forest on low elevations up to + 1200 meters. Trees up to 24 meters high are recorded. The species is endemic to New Guinea Since this species was described in 1939, seven additional collections were examined by me. w Guinea. CENTRAL District: L. J. Brass 5265 (Mafubu, elev. + 640 m.) (pm, NY-holotype), Derbyshire 843 (Kairuku subdistrict, + 51/2 km. north of Ikikina village) (L), R. Schodde 2491 (near Karema, Brown River, elev. 30 m.) (G, K, L), 2642 (near southwest base of Mt. Lawes, elev. 30 m.) (c, L); MoroBe District: McVeigh & N. G. F. Ridgwell 7336 (Bulolo Valley, elev. 900 m.) (Lt), B. Blackwood 270 (elev. 1200 m.) (K); Mutne Bay District: L. J. Brass 24251 (elev. + 64 m.) (G, L), Papua: C. E. Carr 12835 (Rouna) (BM). Mature seeds were not seen at the time the species was described. Seeds are scarlet with black hilum 9 to 10 mm. long and 6 to 7 mm. broad and have a black line extending from the hilum for approximately 2 mm. toward the chalazal end. They are unlike those of any other Asiatic- Polynesian species but are indistinguishable from those of the Australian Erythrina insularis. Seeds of E. merrilliana resemble those of many African species (including E. abyssinica Lamarck) in being scarlet with a black hilum, and those of many American ones (including E. berteroana Urban) in having a black line extending from hilum about 2 mm. toward the chalazal end, although in none of the latter is the hilum itself black. Beside Erythrina merrilliana two species have been recorded from New Guinea: FE. fusca Loureiro and E. variegata L. The former has connate keel petals, a long-stipitate standard, and campanulate calyx; from the latter E. merrilliana is distinguished by its bright orange flowers and el- liptic-oblong standard without prominent veining. This species is related to E. tahitensis. It is immediately distinguished, however, by the small seeds and by the calyx which in the part opposite the cleft has five long (up to 5 mm.) spurlike teeth, resembling in this character again certain African species. 12. Erythrina insularis F. M. Bailey, Queensl. Agr. Jour. 1: 228. 1897. This is said to be a spreading tree about 4.5 to 6.5 meters high. It is known only from the type collection. Australia. QUEENSLAND. Turtle Island: F. M. Bailey 29 (June 1897), BRI- holotype — detached leaflets, petioles, fragments of branchlets, and a single 1972] KRUKOFF, NOTES ON ERYTHRINA, II 139 rachis with pods and seeds, sm — four leaflets and pods with seeds, k — three leaflets, rachis with pods and seeds (plus a pod of E. variegata L. mounted on ight). This species is related to Erythrina merrilliana. It is stellate pubescent and its fruits and seeds are indistinguishable from those of the latter species (see under E. merrilliana). 13. Erythrina vespertilio Bentham in Mitch. Jour. Trop. Austr. 218. This species is a short-boled tree, often from 0.3 to 0.9 meters in diameter and is rather common along the coastal areas of southern Queens- land. It is confined to tropical and subtropical Australia (Western Australia, Northern Territory, Queensland, and New South Wales). The species shows extraordinary variations in the form of leaflets, to which Bentham (FI. Austr. 2: 253. 1864) has already called attention. Two other species of Erythrina native to Africa, namely E. livingstoniana Baker and E. humeana Sprengel also show extraordinary variations in the form of the leaflets. They are not related to E. vespertilio. P.O, Sieh SMIT NEw Mewes 11787 POSTAL SERVICE SEE INSTRUCTION: STAT —— IT OF OWNERSHIP, M MANAGEMENT Nt AND hn a det ON PAGE 2 (REVERSE) es e@ Arnold Arboretum dcto r 971 Quarterly (January, April, July, October] 4 } 22 Divinity Avenue, Cambridge, » 02138 & y 22 Divinity Avenue, Cambridge, » 021338 6. NAMES AND ADDRESSES OF PUBLISHER, EDITOR, AND MANAGING EDIT | PUBLISHER [Name and address] The Arnold Arboretum of Harvard University, 22 Divinity Ave., Cambridge 3) . Bernice G. ens sd 22 Divinity Avenue, Cambridge 02138 Bernice G. “Schubert, 22 Divinity Avenue, Cambridge 02138 ce Samed = Saperion Nant oid een WB Had nd a neatly arr Ti aa od TT individual if 5 tir manse aac dddvena, ae Sic Gs inet of ack individual must be given.) NAME ADDRESS e Arnotd e Harvard University D sue, pridge; Mass; | TOTAL NAME ADD 3s he den and e * 5 ‘ary “C University mOoridzge, Mi i o 0 ° 39 U. S. C, 3626 provides in pertinent part: ih it shal mail mich matter ot the rats pr nder iY ‘aq permission to mal BY NONPROFIT (Check one) The purpose, function, and Nonprofit status of this Have not changed Have changed during (If changed, publisher must honcee Mest see gy exempt status for Federal s} roth tog es Preceding 12 months submit Mite soe with this sta 11, EXTE a cr SSUE DURING. | sincLt issue rus 4 NEAR. s AND N. ACH ISSUE DURING SINGL UE PUBL tad ATURE OF CIRCULATION EDING 12 MONTHS EST TO FILING DATE TOTA oa f ¥ 748 750 B. PAI 1 "Les Tico DEALERS AND CARRIERS, STREET VENDORS AND COUNTER SALES none none 2. MAIL SUBSCRIPTIONS 622 615 C. TOTAL PAID CIRCULATION 622 615 Db. Fnge fC, DISTRIBUTION BY MAIL, CARRIER OR OTHER MEANS 18 18 SAMPLES, COMPLIMENTARY, AND OTHER FREE COPIES” 2. COPIES DISTRIBUTED TO NEWS AGENTS, BUT NOT SOLD none none Pee Dy 640 633 F, OFFICE USE, LEFT-OVER, UNACCOUNTED, SPOILED AFTER L | PRINTING od _ 108 117 DERE itt isha tech oa G. TOTAL (Sum of F 4 F. A) 7 > of editor, publisher, business manager, or owner) Se a a an me sem oe eres wet conte. GLE see fae AS & 1971 ARNOLD ARBORETUM. 1872-1072 CENTENNIAL Boks OF THE ARNOLD ARBORETUM Volume 53 Number 2 : HARVARD UNIVERSHY tug, : = eR Oi ir eee eae ON adc nana oo Journal of the Arnold Arboretum Published quarterly in January, April, July, and October by the Arnold Arboretum, Harvard University. Subscription price $16.00 per year, beginning January, 1972. Subscriptions and remittances should be sent to Ms. Ellen Bernstein, Arnold Arboretum, 22 Divinity Avenue, Cambridge, Massachusetts 02138, A. Volumes I-XLV, reprinted, and some back numbers of volumes 46-50 are available from the Kraus Reprint Corporation, 16 East 46th Street, New York, New York 10017, U.S.A. EDITORIAL STAFF B. G. Schubert, Editor C. E. Wood, Jr. S.A. Spongberg CIRCULATION E. Bernstein Printed at the Harvard University Printing Office, Boston, Massachusetts COVER: A CeNTuRY OF TREES. The cover design was planned and drawn by Karen gcc Vel- mure especially for the Journal of the Arnold Arboretum. It will be used during 1972 to celebrate the Centennial Year of the Arnold Arboretum _ of Harvard University. The design consists of one hundred trees ree a the range and rey of the collestions of of the eens GEORGE B. HINTON 1882-1943 JOURNAL OF THE ARNOLD ARBORETUM VoL. 53 APRIL 1972 NUMBER 2 GEORGE B. HINTON, COLLECTOR OF PLANTS IN SOUTHWESTERN MEXICO JAMES HINTON AND J. RZEDOWSKI * AMONG THE PLANT COLLECTORS of Mexico during the Twentieth Cen- tury, George B. Hinton stands preéminent. Hinton was a metallurgist who decided at the age of fifty that his real and final interest lay in botany. During ten years of intense activity, from 1931 to 1941, he collected 16,300 numbers, concentrating his efforts in 26,000 square miles of the most inaccessible parts of the States of Guerrero, Michoacan, and Mexico. His botanical explorations, made on muleback in areas far beyond the railroads and the highways, in areas previously unvisited by other botanists, con- stitute a contribution to the knowledge of Mexican flora without recent parallel, not only because of the great numbers of new species and new genera he discovered, but also because of the geographical extension his collections provided for the ranges of thousands of species known from other localities or other states. It is estimated that at this writing, his collections include approximately three hundred and fifty new species and four new genera, while future studies of his material will undoubtedly produce even greater numbers of new species. Hinton was born in London in 1882, one of four sons of Charles Howard Hinton, author and professor of mathematics, as well as a novelist; and of Mary Everest Boole, also a mathematician and author. Hinton was the grandson of George Boole, now known as the inventor of computer mathe- matics. He was a nephew of George Everest, Surveyor General of India, whose name was given to the world’s highest mountain. He was also a nephew of Mrs. Voynich, author, and of James Hinton, surgeon and author. In 1889, at the age of seven, Hinton left England, never to return; his family went to live in Japan, where his father was a professor of mathe- matics at the University of Tokyo. After seven years in Japan, the Hintons emigrated to the United States, where Charles Howard Hinton taught mathematics at Princeton, and later worked as a mathematician in the Patent Office in Washington, *The elaboration of this paper by the junior author was subsidized by the C.O.F.A.A. of the Instituto Politécnico Nacional. 142 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Until the age of fifteen, because of an extreme weakness of the eyes, Hinton’s education was conducted by his mother and father at home. Thereafter Hinton attended various American universities, among which were the Minnesota School of Mines, Columbia University, the Arizona School of Mines, and the University of California at Berkeley. After the death of his father, Hinton paid not only his own way through college, but also that of his youngest brother Sebastian, by working during sum- mers as an assayer in the mines of Mexico, which he probably first visited circa 1901. In 1910, he married Emily Percival Wattley, who died in 1921, leaving him three sons, Howard, George, and James. When his children were grown he remarried twice, both marriages ending in divorce. Once Hinton came to live permanently in Mexico, in 1911, he never left the country, not even temporarily during the worst of the revolutions. He was contemptuous of all those Mexicans and foreigners alike, who hastened out of the country at the first rebellion. He had come to live in Mexico for better or for worse, and there he stayed. Hinton loved Mexico and had a great affection for the people, especially the Indians and the peasants, in whose simplicity he found an echo of his own. In the days when so many people despaired of the country’s future, Hinton was one of the very few who kept the faith he had in the future of Mexico and its people. Hinton worked in Mexico as an assayer, metallurgist, civil engineer, architect, industrialist, and finally as a botanical collector. During his apogee, he was considered the finest metallurgist ever to practice in Mexico, having solved problems which had defied the best metallurgical brains of the world for decades. He invented several important and widely used mining processes, as well as ‘‘Floating Cement,” which was patented in Mexico, the U.S.A., England, Germany, and Sweden. A million dollar company had been organized to exploit his patent, when the collapse of Wall Street in 1929 obliged him to return to mining. Hinton began his study of botany in 1931, as a hobby, first teaching himself the requisite Latin. In 1936, at the age of fifty-four, he retired from mining to dedicate himself full-time to his botanical explorations. In the early years of his botanical work, while still occupied with mining, he used as assistants Luis Martinez and Gerénimo Garcia. Upon retiring from mining, however, he enlisted the full-time codperation of his youngest son, James, whom he sent to explore the remotest sierras of Guerrero and Michoacan. While he could ride horseback all day long, day after day, dismounting only to collect, an impaired heart prevented him from climb- ing mountains and cliffs on foot. Furthermore, a certain bronchial weakness prevented him from undertaking exploration of the highest sierras, where working and sleeping both cold and wet for days at a time was almost a condition of collecting at altitudes above three thousand meters. The twenty-six thousand square miles where Hinton concentrated his explorations had remained botanically unknown for two main reasons, first, because of the banditry which had always, even to the present day, 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 143 characterized the area, and second because of the precipitous terrain, not only unpopulated, but virtually inaccessible. The banditry obliged the collector to expose himself not only to great physical hardships but also to the possibility of loss of his life by assassi- nation. While law and order had never been distinguished by their efficiency in the localities Hinton collected, a substantial portion of his work was done beyond their last outposts, where bandits and outlaws took refuge precisely because they were beyond the utmost reaches of the law. This hinterland of the great Guerrero and Michoacan sierras lay anywhere from one to seven days’ journey by horse from the nearest agent of the law. The distance covered in a day’s ride in this country, one of the most broken on earth, might be only five or ten miles, especially when barrancas over two thousand meters deep were encountered. But it was still a hard day’s ride, or a hard day’s walk. Botanical exploration took him as far as seven days on horseback from the nearest outpost of the Mexican Govern- ment, military garrison, post or telegraph office, and three days from the nearest human abode. The only way of being allowed to work in such country, where many of the few inhabitants were outlaws, was to establish personal friend- ships with the people, and to live right in their huts, where the protec- tion of their hospitality was of perhaps greater importance than the protection of their wattle walls or their leaky grass roofs. The first trip was the test, and Hinton had to arrive not alone but with a friend of his future host. Once under his roof, he was not only safe from his host but from the friends and companions of his host. Hinton then had to persuade his host to take him to the next hamlet or hut, and there present him with his following host, so that initially he was handed from one mountaineer to another, until he returned to civilization. Once the first shock of looking at a blonde and blue-eyed outlander was over, once the shock of fear and suspicion had been assuaged precisely by another of their kind, these mountain people were good friends to Hinton. He took presents to them and their children, and developed strong bonds of mutual respect and affection. He was so often implored to prescribe cures for their sicknesses, that little by little their plight caused him to overcome his extreme reluc- tance to prescribe medicine. He took with him such medicines as were most commonly needed, and gave them to his hosts, who received them with gratitude. When the nearest doctor, held in suspicious awe by the mountaineers, might be found only after a ride of several days, an effective remedy given by Hinton on the spot was never forgotten. After the first trip Hinton came and went with impunity. Merchants, cattle buyers, missionaries, prospectors, and even soldiers often dared not enter the mountain fastnesses where Hinton came to feel more at home than in the towns and villages where his operations were based. Because they were not in any way participants of the feuds so numerous in the sierras, both Hinton and his son moved freely across areas infinitely wider than those permissible to the mountaineers themselves, who were locked into small localities by their fear of feuding enemies. 144 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The second reason why those twenty-six thousand square miles had remained botanically unexplored was the nature of the country itself, broken, remote, without access other than by foot or by mule, and at the same time affording extremes of heat and cold without comforts or con- veniences of any kind. Since he could not finance great caravans, Hinton had to live off the land, which often meant nothing but tortillas and salt, and sometimes plain hunger. For Hinton the privations of hunger and thirst, of barely passable living conditions, were of small account. Like the mountaineers, who chose freedom in the lawless mountains rather than the security of the village jails, Hinton chose freedom to explore the wild mountains rather than the comfort and the security of the cities where he might have lived well and long. Of his hosts he could ask for no more than what they themselves ate, and since they were so poor, their best food was often only corncakes and salt. While Hinton might enjoy good food and good liquor, he would permit neither to govern or influence his way of life. Although his explorations forced him to live for long periods on tortillas, with beans and cheese as an occasional luxury, his passion for collecting sustained him. These privations of diet were, of course, responsible for the acute anemia which, in part at least, was to cause his early death. Since it was a financial and tactical impossibility to carry into the wilderness tents, furniture, or food, Hinton planned his expeditions to take advantage of every available hut or hamlet, where he might obtain shelter in which to prepare specimens, a place where he might obtain food not only for himself and his muleteer, but also corn and fodder for his animals. A typical shelter would be a one-room hut, with palm-thatch or grass roof, wattle walls, and too little space to accommodate the mountaineer, his wife and many children, let alone Hinton with his cot, his drying stove, and his gigantic tarpaulin specimen bags. The muleteer-assistant would sleep in the corn trough. During the dry season, of course, Hinton pre- ferred to have his cot and his drying stove under a tree, thus avoiding the crowd in the hut. But in the tempestuous rains, he must crowd into the hut. Since there was usually no table available, Hinton an his tiny cot as a table for changing cardboards and packing specimen By using the lonely huts and the hamlets, Hinton prion himself of the great difficulty of providing his own shelter and of preparing his own food. He would therefore stay in one hut or hamlet until he had thoroughly collected the vicinity, and then he would move to the next hut or hamlet, contriving of necessity to cover the distance between hamlets in a single day, however great that distance might be. It was only the remotest mountains of the Sierra Madre del Sur which required camping out, and to these areas Hinton sent his son James, who made do using a pup tent now and then but who most often relied upon securing shelter in caves or under overhanging cliffs and often lay out the tremendous rainstorms under a riding cape, with the specimens well pro- tected in their multi-layered tarpaulin bags. Riding through the rainstorms by day, James and the muleteer would strip naked, keeping their clothes 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 145 dry under the tarpaulin, and dressing again in these dry clothes once the storm had passed. For food they took “gordas,” thick corncakes made with lard and cone sugar, which might last ten or fifteen days. Heated over glowing embers, these gordas were nutritious and not unpalatable. Monotony dissuaded gluttony, and, held to real hunger, a man could live two weeks at a time on a few kilos of gordas. During his stay in the lonely huts of the sierras, Hinton could usually dispose of no more than a strip of seven or eight feet long by thirty inches in width, for both his cot and his stove, and this he must often dispute with dogs, sheep, chickens, cats, and pigs, all just as anxious as he to elude the icy driving rain. Quite often he must struggle with pigs or dogs scratching themselves or fighting ferociously under and over his cot, and from chickens defecating upon his cot from the rafters. Privacy was the one commodity that he sometimes really missed. To the problem of exploring with an impaired heart the pathless soli- tudes of one of the most broken and mountainous areas in the world, Hinton replied with a superb saddle mule, the only item of expensive equipment that he permitted himself the luxury of using. His saddle mule, and his son’s, were worth more than all the rest of their equipment com- bined. These saddle mules could not be bought with money alone. It was a question of money, luck, and skill. When a good saddle horse could be bought for fifty pesos, a fat cow for sixty pesos, and a pack mule for a hundred and fifty or two hundred pesos, Hinton had been offered as much as two thousand pesos for either one of the saddle mules. Pack mules they bought and sold every year, but the saddle mules they kept until they died. The first was Lenina, whom he bought in 1932, and who died in 1943. He gave Lenina to his son in 1937, when he bought Isabel, who died in 1956. His son James wrote to Dr. Carl Epling, asking that one of the twenty-six new Salvias she had helped discover be named after his saddle mule. Dr. Epling obliged by describing Salvia leninae Epl. Since a saddle mule was the pivot of Hinton’s expeditions, she was lavished with every care. In planning an expedition, each day’s journey must terminate where water and corn were available for the mules. If corn was not avail- able, then it must be carried along, and whatever happened, the mules must have their corn at dawn and at dusk. When it came to finding food for himself and his muleteer, Hinton might give up late on a thunder- storming night. But until his mules had been fed, he would neither eat nor rest. An animal could not be denied, because, not knowing the reason, it felt only its hunger, but a man could bear his hunger because he knew the reason for it. ‘ On an expedition, many stops must be made by the collector during the day, but the pack mules must go steadily on, else the journey would not be completed. The saddle mule must therefore be able to overtake the pack mules after long stops on the way, and to do this she must be a superb animal full of sinew and spirit, because the pack mules them- selves had been carefully selected for speed and endurance. After each day’s journey, Hinton’s first preoccupation was for the mules, that after 146 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 they were fed and brushed, their backs must be carefully examined for the first sign of a crippling laceration. The eye could not be trusted to do the job, the fingers must feel the fur behind the withers, below the tail where the cruppers caught on the long downhills, and behind the forelegs where the cingle chafed on the upward steepnesses. Until the mules were fed and groomed, the men must abide their thirst and their hunger. Out in the wilderness, after eating their corn, the mules were let loose to forage for grass all night long. At dawn they would be back at the hut or the camp, squealing for their corn, which they loudly munched in nosebags while being carefully brushed and saddled and loaded. Each morning, in the lowlands, the mules must be scrupulously examined for the bite of a vampire bat, which, if neglected, would turn into a suppurating and sometimes crippling ulcer. Such care did Hinton take of his animals that during ten years of using them to cover five or six thousand miles a year, no mount was ever lost. To this end a blacksmith was not entrusted to shoe his saddle mule, but he would shoe her himself, and during long expeditions, enough sets of new shoes were carried to reshoe his mule once every six weeks. Too often a blacksmith, if not watched closely as he worked, would save himself the hard work of paring down the hooves with the butteris and then rasping them down with a file, by chopping them down with a bolo, dumping the hooves as it is called, which results in destroying the animal’s balance, often a question of life or death. Just as a nail driven into the quick may cripple a mule so can a dropped shoe. Aside from extra sets of shoes and nails, extra saddlepads and halters must be taken on an expedition for each mule. Hinton had experimented with both horses and donkeys, but found that horses were too delicate and unreliable, while donkeys were not only too small for the deep river crossings and trails, but were cowed into utter helplessness by the incredibly fierce thunder and lightning storms which smote the mountains during the rainy season. Horses had too thin a skin, and no care was enough to prevent saddle sores caused by the excessive rocking of saddles on the broken terrain. Driven to utter exhaustion as might be necessary, they had not the mule’s power of recuperation. Once burros or horses bogged down in a river or a swamp, they ceased to fight and resigned themselves to death. A mule, however, might be trusted to keep on struggling until she got out or killed herself trying. While Hinton and his son were famed for their superb saddle mules, and while as a consequence they might ride farther and faster than the most knowledgeable of the mountaineers, this was the only good equipment they ever used. All other equipment must of necessity be cheap, because there was a strict limit to the amounts Hinton could subsidize with his small personal fortune. He was not able to afford good waterproof boots or leakproof riding capes, waterproof jackets, woolen socks, or adequate tents and sleeping bags. Neither he nor his son owned a pair of gloves. But if his life held many physical privations he was not even aware of them. 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 147 On various occasions prior to 1936, Hinton suspended his collecting for several months at a time because of the banditry. But this factor, which had been the real limiting factor to botanical exploration in the area he collected, was not a major problem for him. He had had so much ex- perience with bandits during and after the revolution that he was confident of his ability to deal with any given situation. It was only when he had certain intuitive doubts about a group of robberies or assassinations that he stopped collecting the localities where these took place. That he had survived literally scores of bandit raids and scores of ransomings during the revolution did not seem to Hinton a matter of luck at all. He could always point out the glaring mistakes made by some one else who had been seriously wounded or had lost his life at the hands of bandits or rebels. During his first years of exploring the remote sierras, Hinton reaffirmed the conclusion he had reached during the revolutions with respect to carrying firearms: that it was safer to go unarmed than to carry a gun. Not only did an unarmed person elicit less stranger-hostility from the mountaineers, but in a dangerous situation, he was forced to rely on his tact, on his skill, on his intelligence, and therefore he was safer than an armed man. A man’s brain, used to the best advantage, was a better weapon than any firearm. A good weapon encouraged the folly of arro- gance and pride. Moreover, in the lawless mountains, a pistol was one of the possessions most coveted by every man. If there were two things paramount in every man’s mind, things for which he hungered, for which he sweated and strove, which completed him as a man and for which he might easily be provoked to assault or to kill, the first was precisely a good pistol or rifle. A good firearm put him on an equal footing with any man that lived. The second was a good saddle mule. Without these two possessions he was just any man. With them, he was a Don. With a good mule he was not only more able to defend himself against his enemies, known and unknown, but since anywhere he went, he must climb some of the steepest and roughest hills in the world, there was no greater convenience for him than a fine mule. With their good mounts, Hinton felt that he and his son were already to a sufficient degree the objects of envy. It was unwise to compound the danger by carrying a good pistol. He finally compromised by carrying in his saddlebag a snub-nosed thirty-eight pistol, which he kept under his pillow at night. Upon one occasion, at a hamlet called Soyate, having been treated there with insulting harshness by the people, he strapped on his pistol openly, and required them to treat him civilly, which in future they did. However, when he moved to Aguililla, where every man carried a pistol and a cartridge belt, after the people had come to know and like him, they said to him, “In this country, it is the custom that every man should carry his good pistol. And if he does not carry one, then we take it amiss, and feel that he is a hypocrite, a man up to no good, a man to be mistrusted.” Therefore Hinton said to his son, “When in Rome . . .’ and in western Michoacan they carried pistols and cartridge belts openly. 148 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 In his house at Aguililla, Hinton kept his loaded pistol and cartridge belt hung from his bedpost, and he slept behind a locked door. The rise of banditry in 1940 led him to suspect that he might be taken for ransom, and he had as a young man been ransomed so often during the revolution that he had no patience for it now in his old age. If ever there should come a ransoming party for him, he said, he would shoot it out with them through the locked door, rather than submit to the nuisance of being ransomed again. To a born naturalist like Hinton, the fears of the local inhabitants concerning wild animals brought only perennial surprise and amusement. The only creature to be feared, he often said, was another man. He could listen with fascinated pleasure at night to the roar of a jaguar nearby, while his muleteer cowered by the fire in horrified awe. Hinton knew that in his work as a collector, it was simply impossible that a jaguar would attack either him or his muleteer. It was even extremely unlikely that a jaguar or a mountain lion would attack his mules. And he lacked what was the native’s worst fear of all— the fear of unknown, of imaginary beasts infinitely more terrifying than the jaguar, the puma, or the coral snake. The dangers of being stung by scorpions or poisonous insects were ever present, but he had been stung so many times by scorpions that the stings were only a nuisance. What was to be really feared was an infected cut or scratch, especially in the lowlands, where scratches must be scrupulously treated and watched to avoid serious consequences. Another danger much greater than that of poisonous insects was the niguas, which, if given a chance, bored under the toe nails, spread over the sole of the foot, and in a month might cripple a man completely so that he would require massive surgery. This danger Hinton avoided by never touching the ground in the huts with his bare feet. In the lowlands, bedbugs, and in the highlands, fleas, required a man to sleep on a cot as long as he slept in a hut, else they would eat him alive. The bite of a rattlesnake, or to a much lesser extent the bite of a coral snake, was a real and frequent danger, but to Hinton it was obvious that a man was much less likely to be killed by a rattlesnake in Guerrero than by an automobile upon a rare trip to Mexico City. Diseases, from leprosy to typhus, from dysentery to spinal meningitis, were constantly encountered, but Hinton had a sublime confidence in his natural resistance to them. Moreover, there was often no choice. Once he had been in a hut for a few days and had come to the conclusion that his host had smallpox or typhoid fever, what could he do but refuse to be worried? He thought that as he had lived and worked in the Balsas River Valley, and had slept in the huts of families having pinto, from whose hands he received his tortillas, that sooner or later he would contract pinto himself. But did he not intend to live the rest of his life with the pintos anyway? And, after all, it was not a disease which incapacitated a man, it gave him only a startling complexion. Hinton’s own frequent disease, malaria, he combated with prophylactic doses of 25 milligrams of quinine 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 149 sulphate a day. And the swarms of gnats and mosquitoes which pestered him in the lowlands, he combated felicitously with clouds of smoke from the cigars and cigarettes which he chain-smoked throughout his waking hours, both day and night. It is, of course, one thing for a strong, healthy young man to bury himself in those wildernesses, and quite another thing for a man of later years, with an impaired heart and declining health. Every time he leaves the last railroad station, the last village, to ride into those solitudes, he is laying his life on the line. If he has put himself four or five days’ ride from the nearest doctor, and six or seven days’ ride from the nearest hospital, then any emergency of health may prove fatal, for he is far beyond medical help. When Hinton left the comforts of home and her- barium at Aguililla to explore the remote sierras, it was because he loved collecting as much as he loved his life. He was not only willing but quite literally eager to give his life to its pursuit. In 1937, on the way to his herbarium at Los Placeres del Oro, he had an internal hemorrhage one night in Coyuca de Catalan, a hemorrhage he attributed to malaria. Passing and vomiting great quantities of blood all night, no doctor was to be found in all the village. He considered himself lucky in finding ice, which he ate until the bleeding stopped. He waited in Coyuca for a few days to regain his strength, and then rode on to Los Placeres del Oro, where he continued his work as if nothing had happened. While Hinton was only too quick to point out that a man did not make his ancestors, yet certainly the accomplishments of his ancestors pressed him to think in terms not of immediate temporary gain but of lasting accomplishments. When he spoke of fame, as he rarely did, it was clear that his interest lay in being remembered only for having been of some lasting use to the world. There crowded upon him the work of so many ancestors who were remembered for their ideas, their books, their inven- tions, that he was driven to do things which he hoped would also be of permanent use to the world. : Having educated his family, and having acquired enough to live on for the rest of his life, Hinton could now devote himself full-time to the task which he loved above all others — plant collecting. Frequently, when he despaired of getting determinations from the specialists, when he saw this delay as a mounting obstacle to his work, he was tempted to get into taxonomy himself. But he would always realize that his greatest usefulness to Mexican botany lay not in taxonomy but in exploration. It was all very well to say that exploration could wait, that sooner or later, if he did not undertake it himself, some one else would. But no one knew better than he how swiftly the great forests were being destroyed by nomadic predatory farmers, who felled the great woods on the steep slopes, burned the pyre, planted a few handfuls of corn one or two years, and then, when the torrential tropical rains had denuded the slopes of their soil, moved deeper into the virgin wilderness. It was a job that had to be done now. By the time the area was crisscrossed with the roads that would bring in any and 150 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 all botanical collectors, the entire flora would have changed substantially and many of the species he was discovering would be lost forever. At the end of ten years of botanical exploration, Hinton had inadver- tently practiced so much medicine that he was contemplating a return to the University to become a titled doctor. There was such a crying need for medical knowledge everywhere he went, and he was so horrified by the witchcraft he saw practiced, that he had begun to think that if he was to continue to collect plants, if he was to continue to live at all, he must surely combine botany with medicine. That was the problem, to continue to live. He was under no illusions concerning his longevity. He came from a short-lived family, and he had lived a life full of hardships and privation. When he retired from mining in 1936, he often told those close to him that all he asked for was another five years of collecting plants. By 1943, he was indeed thankful for the eight years, and he did not know how he could live long enough to learn medicine. He could see himself breaking down physically, and when he died, he was satisfied that he had done as much as he could with his life. Feeling that in the future his work would be severely hampered by physical ills, he was not only ready but eager to die. Impatiently asking both doctor and nurse how much longer he must linger, he insisted that he could do no more, and his last words were that his work was done. Hinton had the utmost confidence in the sure-footedness and intuition of his saddle mule, so that he rode down gorges and up precipices where only a mescal-crazed mountaineer would dare ride. When night found him still on the trail, he allowed his mule to find her way back to the hut, unworried and lost in his own thoughts. During a railroad accident, in which the train from Mexico City to Uruapan jumped a curve and rolled down an embankment, as Hinton lay recovering consciousness on the over-turned car, he heard something roll, and received a stunning blow on the hip. It was a castiron spittoon which had caught in the hat rack, and now rolled off and fell upon him. It broke his hip. But within six weeks he was not only out of plaster and out of the hospital; he was in the saddle collecting plants again, and anxious to forego travel by train. During his last days, coming from Aguililla to Apatzingan, the truck he was riding in turned over, and he received a severe bruising which perhaps contributed to his death a few days later, from coronary thrombosis. Hinton was unorthodox in much of his thinking. When he was collecting along the coast of Michoacan, below Coalcoman, in 1939, he was tracked down by a workman’s commission from the Lane Rincén Mines. They delivered to him a petition signed by the 1200 workmen of the mine, who had received the mine from the company as payment of severance pay, back wages, and other legal benefits. They suspected that the mine had been sabotaged by the company, with the idea in mind that after a few months they would have to beg the company to come back and run the mine after the workmen had given up millions of pesos in benefits. There- fore Hinton left his beloved task of collecting plants in the twenty-six- thousand square miles of his garden and went back to the mine to help 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 151 his men. He found that the flotation mill had indeed been carefully and thoroughly sabotaged. However since he had himself built it, he had it running efficiently within a few weeks. And a few months later, he had the mine running at a profit. The grateful workmen not only rewarded him well, given their limited means, but made him a member of the union, Sindicato de Mineros, which at that time was so hated by both foreigners and upper class Mexicans alike. He used to show his union card to his friends at the British Club, remarking with delight that now he belonged to a “real” club. To the physical problems of collecting in the mountain fastnesses on a limited budget, Hinton brought an inventiveness developed in engineering. What was needed was a system which would enable a collector to travel independently of a base for as long as three months, during which no new cardboards could be obtained, and during which all the specimens collected had to travel on the backs of two or three mules. Further, the specimens must be bright and crisp, and as impervious to moulding as quick thorough drying could make them. Hinton developed a drying furnace which could dry average plant specimens not with special blotters but with ordinary corrugated cardboards within twenty-four hours. First he designed several two-pack furnaces which he cached at strategic points in the sierras. Then he designed especially for his son a telescoping furnace, with a charcoal burner, and three compartments. The whole furnace, telescoped, formed only half a muleload. Set up, this model could handle two or three hundred specimens at a time. Quick drying not only retained the color of most leaves and flowers, but also permitted the preparation of excellent and beautiful specimens under the most difficult conditions. In addition it released cardboards for the next lot of plants. As paper folders became filled with specimens, these were tied into bundles, and the bundles forced into multilayered tarpaulin bags, which when properly tied resisted more than tempestuous rain storms of the Sierra Madre del Sur. They occa- sionally resisted the crossing of rivers when mules fell into unsuspected holes and had to flounder and swim their way out. These furnaces, had they required fuel from the outside, would have presented impossible requirements of weight and volume. But charcoal could be made anywhere and in any amount. To Hinton, these were the happiest days of his life. He delighted in nature, and the idea that he could do something of small but lasting value for his adopted country and for botany filled him with satisfaction. He often commented that collecting plants satisfied the base instinct of acqui- sition and enriched science without exploiting anyone. Hinton counted as a blessing each day that he could spend in the moun- tains, in sunlight, in cool shaded barrancas, on the sweltering plains, and under starry skies at night. Even the routine work of the herbarium, indoors, was not unpleasant to him. He said that as an amusement, handling the labels compared very favourably to playing cards, but was infinitely more useful. When Hinton left mining to devote his full time to plant collecting, he 152 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 had to make a considerable adjustment in his relationship to those who worked for him. In mining he had been accustomed to nothing less than strict and immediate obedience to his orders. He had often had twelve or fifteen hundred men working for him. It usually did not take long for them to learn that here was a man who, while he might be exacting in command, was also uniquely and genuinely interested in their welfare. While still at the Lane Rincén Mines, he had made an arrangement with the union for the company to provide an hour a day, and the men an hour of their own time, in order that the illiterate workmen should all learn how to read and write. For those who could read and write, he had set up classes so that electricians gave them lessons in electricity, mechanics in maintenance, carpenters in woodwork, etc., etc. He himself gave classes in metallurgy. It was understood that those who did not learn how to read and write within six months would be fired. Likewise, the literate must pass examinations in crafts and other skills. The result was that with the cooperation of the union, he was able to raise the level of skills and also of pay, because fewer and fewer men were able to do the same or more work. But out in the lawless hinterland, he realized at once a very different problem: to use his own words, the unwillingness of the people to subject themselves to the indignity of being commanded by another for pay. Sympathetic with their pride and their freedom, he tried nevertheless to make use of the mountaineers but finally realized that it was simply impossible. They had not come upon a free life in order to work for a stranger, however unexacting the work. Therefore he used muleteers from the outside, and when they were not driving his mules from place to place, they helped him collect. This meant that he must often attend to all the routines himself, from collecting to drying and packing the specimens. The first base that he used was El Rincon itself, from which he collected the District of Temascaltepec in the State of Mexico, and gradually extended his collecting southward to the Balsas River. During these years, from 1931 to the end of 1935, when he was becoming more and more intensely interested in the exploration, he collected on weekends and holidays, and had Martinez and Garcia collecting intermittently toward the Balsas, first from El Rincén itself, and then from substations at Tejupilco, Cutzamala, and Pungarabato. As long as plant collecting was a part-time occupation for him, it was only natural that his main base should be at his house. El Rincén was very well situated for collecting the State of Mexico to which he devoted his collecting at the outset. Little by little, as he felt that the job neared completion, he extended his collections southward toward Guerrero. Upon retiring from mining, he immediately moved his base to Los Placeres del Oro, Guerrero, which he chose, as he would choose subsequent bases, for the following reasons: First, because it was cheap, a factor of prime importance. Second, because Los Placeres was at the far edge of the area he had already collected, on the south side of the Balsas River and at the foot of the Sierra Madre del Sur, which he was to begin to collect. 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 153 The village had a postoffice, which was needed for mailing specimens, but beyond Los Placeres there was no outpost of law and order until the other side of the Sierra Madre, say, Petatlan, six days’ ride away, on the Pacific Coast. From Los Placeres he collected for eighteen months, making one trip all the way to Atoyac de Alvarez, ninety kilometres west of Acapulco, an area which he afterwards entrusted to his son James. From Los Placeres del Oro, the base was changed to Zitacuaro, Micho- acan, a fifteen day trek with the entire herbarium packed into zinc cans and loaded on forty-eight burros. The trip was hard on the zinc cans, but for Hinton it was a leisurely trip with his son, more pleasant and cheaper than had he employed trucks, which he considered infinitely less reliable than burros. He had investigated prices in Zitacuaro, and found that they were attractive. What he wanted now was a base from which his son James could continue to make trips into the State of Guerrero, and from which he himself could make trips into western Michoacan. Shortly after establishing the base at Zitacuaro, he established a sub- station at Coalcoman, Michoacan, whence he proceeded to collect westward to the coast, southward to Aguililla, while his son spent half a year col- lecting the area of Zitacuaro before resuming his trips to the Sierra Madre del Sur, with a substation at Atoyac de Alvarez, Guerrero. However, once the area of Zitacuaro had revealed excessive botanical similarity with Temascaltepec, Hinton found the base too far both for himself to collect western Michoacan and for his son to collect the Sierra Madre of Guerrero, Also, he felt the pressure of inflation in Zitacuaro. The last and final base was selected at Aguililla, Michoacan, which lay on no crossroads, and which, it might be said, lay at the end of the trail to nowhere. Like Placeres, it was the last outpost of law and order. (Once an outlaw crossed the Aguililla River, he would be pursued by no soldiers.) The village could be trusted to hold prices for some time, being at that time five days’ ride beyond the nearest railroad or highway at Uruapan. At Aguililla Hinton bought a house and store which he converted into a pleasant home and herbarium. He liked it so well that he planned to live there to the day he died. It was convenient for his exploration of the ranges between the hotland and the coast, as well as the coast of Michoacan itself. From there his son collected Mount Tancitaro, Michoacan. Hinton had a respect for law which had no relationship to his personal opinion of a particular law. All his life he ran his businesses not as close to the margin of the law as he might, but rather as well within the law as he could. If a law was to a point senseless, Hinton did not chafe at it. He abided by it. He no more thought of violating the law than of falsifying scientific data. During his time of collecting, and now also, it was the law that for each collection sent abroad one specimen must be deposited with an agency of the government. To comply with the law, Hinton began by taking dupli- cates of his early collections to the National Herbarium of the Instituto de Biologia, a dependency of the Universidad Nacional Auténoma de México. Dr. Isaac Ochoterena, Director of the Instituto de Biologia at 154 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 that time, refused Hinton’s collections, first on the grounds that he had no room for them; and then on the grounds that he was not interested in Hinton’s work, since there were already too many foreign names in Mexi- can botany. Actually, the Institute of Biology was not in any sense of the word a government agency, nor was there at the time any direct govern- ment agency to which Hinton might have presented his specimens. At Hinton’s insistence, Dr. Ochoterena finally accepted a small number of Hinton specimens and gave both Hinton and his son letters designating them Honorary Collectors — letters which were certainly of considerable help with local authorities. Thereafter, as his collections drew more and more attention abroad, and as Hinton realized their potential importance, he insisted more and more upon fulfilling his legal obligation by depositing one specimen of each of his collections in any government agency which might be induced to accept them, but to no avail. In spite of his wishes, therefore, Hinton found himself in a position of being unable to comply with the law. This was a perennial source of worry, since it was obvious that his explorations might at any minute be stopped on the grounds that he was infringing the law. He tried therefore to obtain, for years, a letter from the Institute, in which they refused his offer of specimens, in order to be able to prove at any given time that he had tried without success to comply with the law. With the same insistence it refused his specimens, the Institute, with strict adherence to the law, denied him such a letter of corroboration. Toward the end of 1939, several lots of specimens addressed to En- gland, Switzerland and the U.S.A., were detained at the post office, and Hinton was required to produce a letter proving that he had deposited duplicate specimens with the government. Subsequently, he was informed that such a letter could now be obtained at the Instituto Biotécnico, a new department of the Ministry of Agriculture, (S.A.G.), and therefore a direct dependency of the federal government. To his great relief, Hin- ton found the Instituto agreeable to receiving a complete set of all his collections in return for a blanket permit to send his specimens abroad. Accordingly, the Instituto Biotécnico was presented with a set of nearly ten thousand specimens, and the lots at the post office were released. On the other hand, he was not given the blanket permit that he had been promised, nor were his specimens officially acknowledged. When his son went to the Instituto to recuperate the multiwalled tarpaulin bags in which the collection had been sent to the Instituto Biotécnico, he was appalled to find the bags half submerged in the water of a flooded cellar However, regardless of being denied his permit, he had no further trouble in shipping specimens abroad for another two years. During this period he continued to send a duplicate of everything he collected to the Instituto Biotécnico, which just as consistently refused to acknowledge them. In 1942, botanical parcels at the post office were once again seized, and Hinton was unable to obtain a permit in exchange for specimens 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 155 which he was equally ee to resort to smuggling or any other sub- terfuge to make the shipme His relationships with — government officials in the provinces, say, on the level of village postmasters, garrison commanders, or municipal presidents were often marred by attempted impositions which he refused to tolerate. His work was frequently viewed with suspicious hostility. Actually, it could be stated in all truth that in Guerrero and Michoacan, a bandit-infested area, Hinton was more hindered by authorities than he was troubled by outlaws. It was a relief for him to ride beyond the last outpost of law and order, into the lawless and unexplored mountains, where the people went out of their way to help him because they admired him for his courage, respected him for his knowledge, and loved him for his generosity. As far as could be done, Hinton used numeration chronologically. How- ever, when two collectors worked simultaneously, especially during the four years his son collected for him, there would be one series of numbers available for his son collecting, say, in Teotepec, while he had another series of numbers for collecting, say, in Coalcoman. For this reason, numbers cannot be taken chronologically save in certain instances, when only one person was collecting at the time. Hinton claimed to have acquired from his own father a strong collector’s instinct, and had at various times in his life made serious and valuable collections, such as of metallic ores mounted on slides, orchids, and beetles. For many years he had collected orchids with his good friend Eric Ostlund, but he sold his orchid collection in 1929, when he went to live in Temas- caltepec, and when his interest in plants became general. He collected beetles mainly for his son Howard, during the early 1930’s. During other times he also collected sea shells and did some excavating for archeological objects. This latter field held great fascination for him, but he could never come to terms with its fundamentally illicit aspects, and therefore he abandoned it. A major point of his satisfaction in living far beyond the last railroad or highway, of living life as it had been lived in the sixteenth an seventeenth century, was that he could enjoy its spiritual values, without being chafed by its lack of mechanical and material facilities. Hinton loved village life, with its definite but varied patterns, to the extent that he at last became a Catholic in order to partake more fully of Mexican village life. He had estimated that in Aguililla, the entire populace attended church once a day. When he was in the village, he loved to rise before dawn, bathe in a tub filled with cool water drawn by hand from a well, and then attend mass and come out from the church with the people in order to see the sun rise, to walk home slowly as the villagers went about their business, some on horseback to distant fields, some to the stalls to sell milk or to buy meat and fruit for the day, the blacksmith heating his forge, the merchant opening his store, the cowherd driving his cattle to pasture, and so on. He ate a simple breakfast of fresh milk, cheese, and tortillas topped off with hot black coffee from his own bushes. 156 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 53 Compared to the breakfasts he had out in the mountains, such a breakfast was luxurious to Hinton. Then he went into his herbarium and worked hard from seven to twelve, when he had a simple but to him luxurious lunch of broth, chicken stew, tortillas, Mexican rice, perhaps preceded by a cup of mescal and followed by black coffee. Then he slept until three o’clock and rose to spend another three or four hours in his herbarium. At night he supped on sweet bread with chocolate or coffee, and after taking a turn about the village square, retired early. In the village there were no professional people of any kind, except now and then a medical student doing his social service. His friends were the village craftsmen and the merchants, the graziers, and the mountaineers who came to visit him when he was at home and whose hospitality he was delighted to repay whenever he could. He had felt the chill of patternless living in the cities, and now he basked in the warmth and charm of patterned living in the villages. That there was no electric light, no sewers, no refrigeration, and no commu- nications bothered him not the slightest. He felt that the full daylight hours were ample for his work, and indeed they were, since there was no waste of time. He could visit the post office, the town hall, the church, the market, and be back in his herbarium within a few minutes. Not that he worked in haste. His pace was intense but unhurried. He enjoyed the dawns, the long slow days, the contented evenings, and the long quiet nights. He slept only a few hours, but spent long hours lying awake, rested, thinking and smoking while he awaited the coming of dawn. He enjoyed his friendship with the villagers, above all with the children, for whom he had swings built in his garden and often made little parties. He enjoyed having the time to consult his botanical works, of which he had a small but very good collection. When the corn harvest failed in Aguililla in 1940, Hinton imported corn from the coast, and had it given to the children who came to ask for it. This he was able to do without incurring the hostility of the merchants who were speculating with high prices while he was giving it to those who could not pay such prices. During most of his time in the village, he gave English classes to those who wanted them and usually had a class of five or six adults to whom he gave an hour a day without charge. At the same time, while he hated to do it, he could not avoid prescribing medicine, principally to those who were too poor to buy the medicine or pay the doctor. His cures were so good that against his will he found himself obliged increasingly to prescribe for the most desperate cases. In a financial sense, Hinton’s big problem was not that his combined income from the theoretical sale of plants together with rent from houses could not cover the cost of exploring what he called his garden. During the four years that his son James assisted him full time, the sale of twelve specimens of each number, plus the income he had from his houses and other sources, would easily have paid for the cost of collecting. The prob- lem was that he could not get determinations nearly as fast as he and his son could collect. The more determinations lagged, the more he had to 1972 | HINTON & RZEDOWSKI, GEORGE B. HINTON 157 dig into his savings to finance the collecting, while his herbarium grew bigger and bigger. The additional specimens were a problem to collect and a problem to store. But the specimens in his herbarium could not pay the cost of collecting because they were increasingly undetermined. No more eloquent explanation of this point can be found than his letter to one of the herbaria responsible for determining his specimens, when he wrote in 1939: “You ask me to be patient about the determinations on the lots of plants sent you in 1934. My mules have been eating corn for five years!” Once he had the collecting started, properly organized, he could hardly stop it. For one thing, he had the judgement to realize that his own lease on life would not wait. For another, he knew that his son could not continue to help him indefinitely. Therefore he must push the collecting NOW, regardless of determinations, regardless of the sale of specimens, trying to measure out his own personal fortune in proportion to his own life expectance, but preferring to gamble rather than to play it safe. Having started with the herbarium at Kew as the principal institution in Classifying his specimens, the advent of the war obliged him to change to the Smithsonian Institution and the Gray Herbarium. Finally, at the time of his death, he had in many cases provided four or five specimens, and he still did not have the determination which would enable him to sell the remaining specimens, if any. In 1941, the war brought botanical exploration to a close. Hinton left his herbarium to solve metallurgical problems at Taxco, with the intention of serving the cause in this manner for the duration of the war and not unmindful of repairing his financial position. His son James went to work for the Rubber Development Corporation. After a year at Taxco, when Hinton returned to Aguililla, it was to find that during his absence, the cyanide poisoning he had left in the herbarium had vanished, and the insects had attacked his collection with an astonishing rapacity. After going through the entire collection, a work of months, he calculated that something like thirty or forty thousand specimens had been lost. At a later date, after his death, the collection was again ravaged by insects and sustained another loss almost as great, after which his son removed it to Mexico City where the collection would be more frequently watched. It might be said that the difference in what Hinton spent collecting and what he received from the sale of specimens, was determined rather pre- cisely by the loss of specimens destroyed at one time or another by insects. Had all his specimens been determined and sold, the collection would have paid for itself through the sale of specimens. During his expeditions, Hinton constantly shrugged off the lure of gold. Most of his collections were made in gold country, and he seldom visited a hut where his host did not show him a piece of gold-bearing ore, or perhaps a nugget, and try to inveigle him into developing a placer or a vein. Being a mining man, it was perhaps easier for him to refuse to touch a host of bonanzas, not only in gold, but also in silver, copper, lead, mercury, guano, precious woods, and cattle. He had not the time to become 158 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 rich, he had barely enough time to explore his wild and lonely garden. He was also often lured to visit other areas in other states by natives who knew fabled wildernesses. At first enthralled by their tales of the vegeta- tion, he would finally shake his head and laugh and tell them that he had time to collect only one place well. He did not want to skim new species here and there, he wanted to establish a permanent record of everything that had once grown in his garden. During the four years of most intense collecting, between 1936 and 1940, Hinton distributed his specimens as follows: one specimen of everything he collected to the Royal Botanic Gardens, Kew, the British Museum (Nat. Hist.), the herbaria at Zurich and Geneva, Harvard Uni- versity, and the Smithsonian. Frequent lots were sent to the Field Mu- seum and Missouri Botanical Garden, as well as to the New York Botanical Garden. Special groups, such as lupines and salvias, were sent also directly to Dr. Charles Piper Smith and Dr. Carl Epling. Infrequent sets of a thousand specimens were sent to Naturhistoriska Riksmuseum, Stockholm; Catholic University, Washington, D.C.; Mrs. Douglas of Mexico City, etc. Hinton had originally made only two or three specimens of each collec- tion. When the possibility of financing some of his expenses out of the sale of duplicates came to his attention, he began to take six specimens of each collection, wherever possible. When his son James returned from studies in Canada to devote himself full-time to assisting him, the number of specimens in each collection was increased to twelve, with the idea in mind of being able to finance future collecting. By this time, also, it was apparent that in order to have anything left for distribution, it was necessary to provide as many as three or even four specimens of each collection in order to obtain a determination without which the collection had little or no commercial value. At the ratio of three specimens for a determination, if only six specimens were collected at the outset, then the price of fifteen U.S. cents per speci- men for three specimens gave him 7 1/2 U.S. cents per specimen collected. However, if twelve specimens were collected at the outset, and three given for a determination, then in theory the gross return rose to 11 1/4 U.S cents per specimen. After Hinton’s death in 1943, his son transferred some fifty-six thousand specimens to the New York Botanical Garden, which distributed these to forty-four herbaria listed on page 160 There now remain in Hinton’s collection approximately twenty thousand specimens. It is estimated that the total number of specimens collected by Hinton, including those lost to insects and at government agencies, cannot exceed 150,000. Total remuneration received from the distribution of specimens, including subsidies from the Royal Botanic Gardens, Kew, is estimated at twelve thousand dollars. His records show that Hinton subsidized his collections to the amount of roughly another twelve thousand dollars — and, of course, worked without personal remuneration, other than the 1972 | HINTON & RZEDOWSKI, GEORGE B. HINTON 159 enormous personal satisfaction, which he would not have traded for any money, full-time for five years, and part time for another five. Because of his drying system, Hinton was able to produce very ample specimens, to the degree that Kew and others were often able to divide one specimen into several, and even this, sent to some other herbarium was again divided. As a consequence, Hinton’s specimens found their way to a great many herbaria impossible to record in this wor As his collecting progressed and determinations lagged more and more, he began to perceive that the bottleneck to increased knowledge of the Mexican flora was not so much botanical collecting as it was botanical determinations. To this end, he began to work on a mechanical device to reduce the task of determinations. This was a punching device, later per- fected elsewhere, which he did not live long enough to build himself. Hinton’s interest in botanical collecting was principally scientific. If he could find only one or two specimens, he collected and sent them to the herbarium or to the specialist. This, naturally, deprived him of a specimen for his own collection, and it deprived him of specimens which could be sold. But this was for science, which was his main purpose. Furthermore, he had the means to indulge his pleasure in collecting. Had there been no scientific interest in his work, it would have destroyed his sense of mission, that sense of accomplishment which was so important to him. But he would still have enjoyed the work, the use of the faculties it demanded. He also enjoyed the work for its own sake, the exploration of barrancas and mountains which had perhaps never been looked upon by a trained eye backed by an academic education. He enjoyed riding horseback, the awaking long before dawn, the brushing, feeding, and saddling of the mules, and the early start so that the sunrise might be appreciated out in the woods or the top of the hills. Hinton loved the wilderness, and had it not been botany, some other mission would surely have taken him to those solitudes. It was only when he could collect ten or twelve sheets that he had eight or ten for distribution to the herbaria. These, then, were to defray ex- penses. While there were rare times when expenses were fully met, his system of collecting meant, of course, not only that he and his son worked for many years without any personal income, but also that he used both the interest and the principle of a good part of his personal fortune. Had he brought to collecting a business attitude, and had he collected only the species which could have provided at the moment large, remunerative numbers of specimens, his work would have lost much of its botanical interest and value. On the other hand, Hinton collected also the most common of species, in order to establish a record of distribution of known plants in botanically unknown areas. Without these numerous collections of common plants, the proportion of new species he collected would have increased spec tacularly, but he would have lost the opportunity of extending the ain geographical range of common as well as of new and rare species. The more valuable he thought a specimen, the less he wanted it in his 160 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 own herbarium, where it would receive the least protection. In his arrange- ment with the different herbaria which worked upon his collections, Hinton supplied large numbers of additional specimens without cost. In the final outcome, a great many of the determinations he received cost him as many as four specimens, and even for this number, a very substantial portion of his collections had not been classified before his deat Basically, the purpose of his work as Hinton saw it was to place before the specialist ample or at least adequate botanical material in the best possible condition. His attitude toward collecting was one of uncompro- mising honesty. When, as inevitably happens, there arose any confusion between a plant and its label, he destroyed both without hesitation, for if the main purpose of his collections was knowledge, then he could not possibly be instrumental in the fabrication of misbeliefs and errors. The purpose of his collections was indeed knowledge, and he was richly satisfied that in his own small way, he had served this purpose well. Institutions to which duplicates of G. B. Hinton’s collections were distributed by the New York Botanical Garden after Hinton’s death. (See comments on page 158.) New York BoTANICAL GARDEN Museum NATIONAL D’HISTOIRE Bronx, New York. NATURELLE Paris, France. SOUTHERN METHODIST UNIVERSITY Dallas, Texas. TEXAS RESEARCH FOUNDATION Renner, Texas. ACADEMY OF NATURAL SCIENCES HILADELPHIA UNIVERSITY OF TEXAS Philadelphia, Pennsylvania. Austin, Texas. UNIVERSITY OF WASHINGTON CONSERVATOIRE ET JARDIN BOTANIQUES Seattle, Washington. Geneva, Switzerland. U.S. DEPT. or AGRICULTURE ScIENCE MUSEUM, THE INSTITUTE OF Beltsville, Maryland. AMAICA Kingston, Jamaica. ARNOLD ARBORETUM, HARVARD UNIVERSITY RuTGERS UNIVERSIT Cambridge, Massachusetts. New Brunswick, New Jersey. UNIVERSITY OF CALIFORNIA DupLEY HERBARIUM Berkeley, California. Stanford, California. UNIVERSITY OF MICHIGAN INSTITUTO MIGUEL LILLO Ann Arbor, Michigan. Tucuman, Argentina UNIVERSITY OF ARIZONA FIELD MUSEUM Tucson, Arizona. Chicago, Illinois. 1972] PoMONA COLLEGE Claremont, California. NATURHISTORISKA RIKSMUSEUM Stockholm, Sweden. BOTANICAL MUSEUM AND HERBARIUM Utrecht, Netherlands. HINTON & RZEDOWSKI, GEORGE B. HINTON JARDIN BOTANIQUE Leningrad, U.S.S.R. JARDIN BOTANIQUE NATIONAL DE BELGIQUE Brussels, Belgium. BOTANICAL INSTITUTE OF THE 161 ACADEMY OF SCIENCES OF THE U.S.S.R. Leningrad, U.S.S.R. NATURHISTORISCHES MusEUM Vienna, Austria. Roya Botanic GARDENS Kew, Richmond, Surrey, England. UnIversITY oF ILLINOIS Urbana, Illinois. BOTANISCHES MusEUM Berlin-Dahlem, Germany. UNIVERSITY OF CALIFORNIA Goleta, California. Missouri BOTANICAL GARDEN : St. Louis, Missouri. WASHINGTON STATE UNIVERSITY BoranicaL Museum, Tue Untversrry Pullman, Washington. Lund, Sweden. UNIVERSITY OF GEORGIA BriTisH Museum (Nat. Hist.) Athens, Georgia. d London, England. UNIVERSITY OF MINNESOTA BOTANICAL MusEUM AND HERBARIUM Minneapolis, Minnesota. Copenhagen, Denmark. U Towa STATE UNIVERSITY abiraepee pe ee Ames, Iowa. Laramie, Wyoming. UNIVERSITY OF OKLAHOMA BOTANISCHE STAATSSAMMLUNG Norman, Oklahoma. Munich, Germany. NATIONAL MusEUM OF CANADA SMITHSONIAN INSTITUTION D Ottawa, Canada. Washington, D.C. DUKE UNIVERSITY CORNELL UNIVERSITY Durham, North Carolina. Ithaca, New York. Note: Dr. Rzedowski has prepared with meticulous care as complete a list as possible of determinations of the Hinton collections. Completion of this task was made possible through the cooperation of Mr. C. V. Morton, of the Smith- sonian Institution, who made available records compiled from his own and other institutions, and also by use of reports found in Hinton’s files. The list of collection numbers (with or without names) is almost 500 pages long making it too costly to print, as well as unsuitable for the pages of a journal. Therefore, until the end of 1972, the editors of the Journal of the Arnold Arboretum will undertake to have prepared, at cost plus postage charges, xerox copies of the list for those individuals or institutions finding a copy important for proper use of their collections. Charges will be about $18.00 per complete list. — Eds. 162 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 PARTIAL LIST OF NEW TAXA BASED ON THE COLLECTIONS OF G. B. HINTON GENERA Hintonella Ames (Orchidaceae) Hintonia Bullock (Rubiaceae) Onocleopsis Ballard (Aspidiaceae) Symplococarpon Airy-Shaw (Theaceae) SPECIES AND VARIETIES oTE: The number following the name of the authority is the number of the N type collection. ACANTHACEAE Aphelandra lineariloba Leonard, 3156 Beloperone longibracteata Leonard, Berginia hintonii Leonard, 5956 Buceragenia hirsuta Leonard, 5318 Carlowrightia coyucana Leonard, 6129 Carlowrightia lanceolata Leonard, 2672 Carlowrightia mucronata Leonard, 5533 Dicliptera aquatica Leonard, 3337 Jacobinia capitata Leonard, 1354 Pseuderanthemum axillare Leonard, 3200 Stenandrium mexicanum Leonard, 4292 Tetramerium hilliit Happ, 5724 Tetramerium hintonii Happ, 3375 ANNONACEAE Desmopsis mexicana Fries, 15836 APOCYNACEAE Echites woodsoniana Monachino, 15325 ARISTOLOCHIACEAE Aristolochia bullata Pfeifer, 14569 Aristolochia cardiantha Pfeifer, 9096 Aristolochia glossa Pfeifer, 15151 ASCLEPIADACEAE Gonolobus megalocarpus P.G. Wilson, 7141 Matelea hintoniana Woodson, 1450 Pseudomarsdenia lanata P.G. Wilson, 7602 ASPIDIACEAE Onocleopsis hintonii Ballard, 3297 BEGONIACEAE Begonia asteroides Smith & Schub., 2206 Begonia cylindrata Smith & Schub., 5042 Begonia extranea Smith & Schub., 993 Begonia fernaldiana Smith & Schub., 9425 Begonia hintoniana Smith & Schub., 2050 Begonia michoacana Smith & Schub., 15186 Begonia nemoralis Smith & Schub., 15063 BIGNONIACEAE Adenocalymma hintonii Sandw. [fl.] & 7259 [fr.] , 3784 BORAGINACEAE Bourreria hintonii 1.M. Johnst., 10866 Bourreria longifora I.M. Johnst., 15908 Bourreria superba 1.M. Johnst., 13834 Cordia coyucana I.M. Johnst., 8156 Cordia hintonii I.M. Johnst., 7740 Cordia subvelutina I.M. Johnst., 15066 mp ie acres cremnogenum I.M. Johnst Neheticsion pee var. hintonii 1.M. Johnst., 3631 1972] HINTON & RZEDOWSKI, GEORGE B. HINTON 163 Heliotropium michoacanum 1.M. _ Bidens minensis Sheriff, 9453 Johnst., 15069 Chaptalia hintonii Bullock, 3098 Coreocarpus hintonit Sherif, 9379 BROMELIACEAE Cosmos hintonii Sherif, 10647 Catopsis mexicana L.B. Smith, 4692 Cosmos purpureus var. flavidiscus Sherff, 1700 Pitcairnia flexuosa L.B. Smith, 3948 Pitcairnia hintoniana L.B. Smith, 2040 BURSERACEAE Bursera citronella McV. & Rzed., 12299 Bursera coyucensis Bullock, 6895 Bursera crenata P.G. Wilson, 12025 Bursera denticulata McV. & Rzed., 1202 Bursera dubia Bullock, 9349 Bursera fragrantissima Bullock, 9731 Bursera grandifolia f. robusta Bullock, 9 Bursera heteresthes Bullock, 7812 Bursera hintonii Bullock, 6991 Bursera occulta McV. & Rzed., 12555 Bursera sarcopoda P.G. Wilson, 12293 Bursera sessiliflora var. pubivalvis Bul- lock, 9052 Bursera staphyleoides McV. & Rzed., 1207 Bursera trifoliolata Bullock, 6340 Bursera trimera Bullock, 6262 Bursera velutina Bullock, 6261 Bursera velutina var. parvifolia Bul- lock, 9193 CAMPANULACEAE Diastatea expansa McV., 2644 CAPPARACEAE Forchhammeria hintonii P.G. Wilson, 7617 Physostemon aureum R.C. Foster, 10537 CARYOPHYLLACEAE Drymaria grandis Bullock, 5427 CELASTRACEAE Euonymus corymbosus Sprague & Bul- lock, 325 COMPOSITAE Bidens gracillima Sherff, 4546 Dahlia hintoni Sorensen, 14499 Eryngiophyllum pinnatisectum P.G. Wilson, 659 Florestina lobata Turner, 1261 Guardiola pappifera P.G. Wilson, 9333 Heliopsis purpurea P.G. Wilson, 8096 Melampodium pilosum Stuessy, 11428 Microspermum flaccidum P.G, Wilson, 761 Microspermum tenue P.G. Wilson, 53 Perezia scaposa Blake, 15838 Perezia simulata Blake, ae Piqueria hintonii King, 3517 Porophyllum warnockii cee 8469 Psacalium nanum Pippen, 14464 Rumfordia pinnatisecta pa Wilson, 11289 Sabazia liebmannii var. hintonii Long- pre, 148 Sabazia microspermoides Longpre, 14788 Schkuhria greenmanii Heiser, 4507 Senecio bombycopholis Bullock, 7790 e, Zinnia jugistyla P.G. Wilson, 12038 CONVOLVULACEAE Breweria elliptica Smith & Schub., 2176 Ipomoea cordata Smith & Schub., 6984 Ipomoea decemcornuta O’Donell, 4991 CRASSULACEAE Sedum hintoni Clausen, 15926 CRUCIFERAE Cardamine obliqua var. stylosa Rollins, 6080 CUCURBITACEAE Microsechium hintonii P.G. Wilson, 8596 164 DIOSCOREACEAE Dioscorea guerrerensis R. Knuth, 6659 Dioscorea hintonit R. Knuth, 5215 Dioscorea pantojensis R. Knuth, 6226 Dioscorea temascaltepecensis R. Knuth, 2716 ERICACEAE Befaria hintoni Camp, 14237 EUPHORBIACEAE Acalypha erecta P.G. Wilson, 4596 Acalypha trilaciniata P.G. Wilson, 12026 Argythamnia micrandra Croizat, 6307 Euphorbia hintonii Wheeler, 6257 Euphorbia succedanea Wheeler, 6580 Pedilanthus coalcomanensis Croizat, Pedilanthus gracilis Dressle Phyllanthus 15857 r, 10973 coalcomanensis Croizat, FAGACEAE Quercus apiophylla Warb., 6568 Quercus sageta Warb., 6577 GENTIANACEAE Halenia hintonit Bullock, 8273 GERANIACEAE Geranium cruceroénse R. Knuth, 4167, Geranium hintonit H.E. Moore, 14787 Geranium hystricinum H.E. Moore, 14972 Geranium pata R. Knuth, 8945 Geranium pinetorum R. Knut h, 1887 Geranium a ee R. Knuth, 27 Geranium unguiculatum H.E. Moore, 14906 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 53 GESNERIACEAE Achimenes bella Morton, 4512 Achimenes brevifolia Morton, 10766 Achimenes dulcis Morton, 8207 Achimenes flava Morton, 4485 Achimenes flava var. puberula Morton, 105383 Achimenes obscura Morton, 4705 Niphaea mexicana Morton, 4544 Phinaea multiflora Morton, 10626 GRAMINEAE Aristida hintonii Hitchc., i Hilaria hintonit Sohns, 450 Muhlenbergia eriophylla Seal ie Muhlenbergia hintonit Swallen, 3 Panicum hintonii Swallen, 2527 Paspalum hintonii Chase, 1 Sporobolus hintonit Hartley, 2610 Zeugites hintoni Hartley, 4538 Zeugites sagittata Hartley, 4793 GUTTIFERAE Hypericum hintonii Bullock, 1796 HAMAMELIDACEAE Matudaea hirsuta Lundell, 7381 IRIDACEAE Cypella mexicana Morton, 10322 Sphenostigma hintonit R.C. Foster, 13996 Sphenostigma longispathum var. fili- forme R.C. Foster, 4614 Sphenostigma mexicanum R.C. Foster, 8010 Tigridia hintonii Molseed, 9748 Tigridia mortonii Molseed, 6527 JUGLANDACEAE Juglans major var. glabrata Manning, 3973 LABIATAE Hyptis iodantha Epling, 9844 Hyptis perpulchra Epling, 8574 Hyptis pseudolantana Epling, 11266 Salvia cyanantha Epling, 15350 1972] Salvia cyanicephala Epling, 12792 Salvia dichlamys Epling Salvia dryophila Epling, 5406 Salvia gravida Epling, 12354 Salvia hamulus Epling, 9884 Salvia hintonii Epling, 10150 Salvia jacobi Epling, 14110 Salvia languidula Epling, 10351 Salvia leninae Epling, 14897 Salvia nigriflora Epling, 13956 Salvia palealis Epling, 14040 Salvia perblanda Epling, 9946 Salvia praestans Epling, 11095 Salvia punicans Epling, 11224 Salvia sapinea Epling, 14798 Salvia scaposa Epling, 5332 Salvia subobscura Epling, 12570 Salvia subpatens Epling, 4376 Salvia synodonta Epling, 12576 Salvia tricuspis Epling, 11260 Salvia viscidifolia Epling, 10725 Scutellaria blepharophylla Epling, 10507 Scutellaria hintoniana Epling, 1136 LAURACEAE Cryptocarya hintonii Allen, 13737 Persea hintonii Allen, 3533 LEGUMINOSAE Aeschynomene hintonii Sandw., 6104 Astragalus hintonii Barneby, 13240 Brongniartia cuneata Smith & Schub., 6311 Brongniartia funiculata Smith & Schub., 5397 Brongniartia proteranthera Smith & chub., 775 Caesalpinia hintonii Sandw., 7746 — multifoliolata . Wilson, 12352 ra stenosepala Smith & Schub., 530 oe. a var. velutina Smith & Schub., 11876 Cassia ies Smith & Schub., 6144 Cassia trichocraspedon Sandw., 5582 Crotalaria pilosa var. robusta Senn, HINTON & RZEDOWSKI, GEORGE B. HINTON 165 eas gs tuerckheimii var. macrantha Senn, 506 Dae ace. caudata P.G. Wilson, 2333 Dalea hintonii Sandw., 5314 Erythrina mexicana ee 5333 Inga hintoni Sandw., 7617 Lonchocar pus peste Sandw., 5936 [fl.] & 6508 [fr. Lupinus aralloius C.P. Smith, 15592 was re aschenbornii var. cervantesii C.P. Smith, 8340 Lupinus dispar C.P. Smith, 15712 Lupinus elegans var. quercetorum C.P. Smith, 13535 Lupinus ferax C.P. Smith, 15724 t Lupinus metallactus C.P. Smith, sb Lupinus monteportae C.P. Sm 12720 Lupinus octablomus C.P. Smith, 14521 Lupinus paraplesius C.P. Smith, 15654 Lupinus perbonus C.P. Smith, 15702 Lupinus persistens var. praeclarus C.P. Smith, 831 Lupinus quercuum C.P. Smith, 8362 Lupinus quercuum var. perflexilobus C.P. Smith, 13285 Lupinus rhodanthus C.P. Smith, 13117 Lupinus semiaequus C.P. Smith, 15596 Lupinus simulans var. abietum C.P. Smith, Lapis squamecaulis C.P. Smith, Lapin tancitaricus C.P. Smith, 15599 inus taurimortuus C.P. Smith, Mimosa egregia Sandw., 6314 [fl.] & 6907 [fr. Nissolia hintonii Sandw., se Nissolia leiogyne Sandw,, 629 —— tripetalus P. G. ies 3351 pets grandifolia var. glabrescens Sandw., 5419 Platymiscium lasiocarpum Sandw., 3174 beasts mexicana C.E. Wood, 2698 Tephrosia pogonocalyx C.E. Wood, 101 Tephrosia vernicosa C.E. Wood, 9950 166 LENTIBULARIACEAE Pinguicula imitatrix Casper, 14219 LILIACEAE Calochortus hintonii Bullock, 1383 Milla delicata H.E. Moore, 9920 Milla mortoniana H.E. Moore, 9725 Schoenocaulon mortonii Brinker, 13465 Smilax occidentalis Morton, 14364 LYTHRACEAE Cuphea avigera var. bracteolata R.C. Foster, 3408 Cuphea hintonii Bullock, 2691 Cuphea michoacana R.C. Foster, 12268 Cuphea ornithoides R.C. Foster, 10678 Cuphea pertenuis R.C. Foster, 13460 Cuphea pulcherrima R.C. Foster, 14837 Cuphea trichochila R.C. Foster, 9583 Cuphea vesiculigera R.C. Foster, 10570 Cuphea wrightii var. compacta R.C. Foster, 15125 MALPIGHIACEAE Malpighia hintonii Bullock, 1192 MALVACEAE Malvaviscus hintonii Bullock, 3928 MELASTOMATACEAE Tibouchina latibracteata P.G. Wilson, 7165 MENISPERMACEAE Disciphania mexicana Bullock, 4314 MYRSINACEAE Parathesis hintonii Lundell, 10402 Parathesis tetramera Bullock, 10779 Parathesis villosa Lundell, 13721 MYRTACEAE Eugenia — McV., 16292 Eugenia crenularis Lundell, 7695 Eugenia waders McV., 8957 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Eugenia hintonii Lundell, Eugenia michoacanensis epee 13812 NYCTAGINACEAE Okenia parviflora P.G. Wilson, 12101 ONAGRACEAE Lopezia hintonii R.C. Foster, 14902 ORCHIDACEAE Bletia concolor Dressler, 10481 Brassia oestlundiana L.O. Williams, 180 Bulbophyllum cirrhosum L.O. Wil- liams, 14191 Epidendrum eximium L.O. Williams, Hid mexicana Ames, 8928 Lockhartia mexicana L.O. Williams, 640 Oncidium hintonii L.O. Williams, 13784 oe hintonit L.O. Williams, Prin exilis Dressler, 2346 gar et sie pachyphylla L.O. Wil- liams, PIPERACEAE Peperomia hintonii Yuncker, 4648 POLYGALACEAE Polygala hintonii Blake, 256 PTERIDACEAE Adiantum arcanum Maxon & Morton, 731 Adiantum hintonii Maxon & Morton, 6832 RANUNCULACEAE Delphinium subscandens Ewan, 6579 Thalictrum hintonii Boivin, 6743 RUBIACEAE Bouvardia capitata Bullock, 6319 Bouvardia cataphyllaris Bullock, 1131 Bouvardia hintonii Bullock, 8107 1972] Bouvardia standleyana Blackwell, 12677 Crusea grandiflora P.G. Wilson, 4848 Deppea hintonii Bullock, 4824 Hintonia latiflora var. leiantha Bullock, Sickingia mexicana Bullock, 4818 SCROPHULARIACEAE Buchnera retrorsa Philcox, 2127 Buchnera tenuissima Philcox, 5275 Penstemon hintonii Straw, 11290 Russelia hintoniit Lundell, 10921 Russelia leptopoda Lundell, 13845 Russelia tenuis Lundell, 15841 Russelia teres Lundell, 16127 SOLANACEAE Physalis hintonii Waterfall, 8457 Solanum brachycarpum Correll, 13995 Solanum hintonii Correll, 4416 STERCULIACEAE Ayenia pringlei Cristobal, 12082 Byttneria atrata Bullock, 4830 STYRACACEAE Styrax hintonii Bullock, 3798 THEACEAE aa ea airy-shawianum Ko- buski, 3081 Symplococarpon Tt Airy-Shaw (Bullock) TILIACEAE Triumfetta columnarioides Bullock, 4917 HINTON & RZEDOWSKI, GEORGE B. HINTON 167 eee heliocarpoides Bullock, 515 Triumfetta hintonii Sprague, 7018 Triumfetta quercetorum Bullock, 6746 Triumfetta stellata Lay, 4460 UMBELLIFERAE Donnellsmithia hintonii Math. & Const., 8463 VALERIANACEAE Valeriana tanacetifolia Meyer, 1663 VERBENACEAE Citharexylum affine var. glanduliferum Moldenke, 7344 Citharexylum hintonii Moldenke, 3399 Lantana frutilla Moldenke, 6455 Lantana hintonii Moldenke, 9307 Lantana scandens Moldenke, 12315 Phyla yucatana var. parvifolia Mol- denke, 6 Stachytar pheta 8236 hintonii Moldenke, Stachytarpheta mexicana Moldenke, Stachytarpheta velutina Moldenke, Verbena hintonii Moldenke, 11991 Verbena litoralis var. albiflora Mol- denke, 13965 VITACEAE Cissus tuberosa var. hintonii Bullock, 8172 ZYGOPHYLLACEAE Kallstroemia hintonii Porter, 12106 The partial lists are arranged by districts (distrito Correspondence to modern municipali taken from modern maps and the position of a given locality on Hinton’s original map ma Collecting Localities of G. B. Hinton s), ancient administrative units, which were used by Hinton on his labels. ties (municipios) is indicated. The municipalities, access roads and coordinates were not coincide exactly with the corresponding listed coordinates. Altitude and vegetation data are taken from an archive of duplicate labels which was the _-main source of information in the elaboration of the list. ESTADO DE MEXICO DistRIcTt OF TEMASCALTEPEC APPROXIMATE PREDOMINANT CoorDINATES ALTITUDE oO LocaLity MUNICIPALITY LAT.N LONG Access Roaps VEGETATION Acatitlan Tejupilco 19°01’ 100°19’ 1130 Awdr * to Valle de Bravo and Tejupilco Tropical Anonas (Las Anonas) Tejupilco 18°48’ 100°22’ 880 Awdr to Tejupilco and Bejucos Oak Bejucos Tejupilco 18°47’ 100°26’ 610 Awdr to Tejupilco Berros (Cerro de los Berros) Tejupilco 19°00’ 100°05’ 950 Awdr to Temascaltepec and Tejupilco Oak Cajones Texcalitlan 19°03’ 99°53’ 2460-2500 Awdr to Toluca and Sultepec Oak-pine (2900?) Tejupilco 18°48’ 100°24’ 770 Awdr to Tejupilco and Bejucos Cafitas Tejupilco 1a°st 100°18’ 1000-1350 Awdr to Tejupilco and Bejucos Carboneras Temascaltepec 19°03’ 100°00’ 1900-2030 Paved highway to Toluca and Temascaltepec Oak-pine Cerro de Los Berros Cerro Mufieca (see Ypericones) Chorrer: Tejupilco 18°55’ 100°14’ 1230 Awdr to Tejupilco and Bejucos ? Oak Comunidad Temascaltepec 19°08’ 99°56’ 2250-2750 Paved highway to Toluca and Temascaltepec Pine-oak Crucero (El Crucero) Temascaltepec? 19°08’ 99°50’ 2670-3400 ? Fir-pine WOLAXOTAV GIONUYV AHL AO TVNUNOL €S “I0A] Cruces (Las Cruces) Cucha Cuentla Cumbre (La aig Cumbre de Tejup El Crucero (see Pio) Guayaba Hornos (Los Hornos) sca (see Ypericones) Ixtapa La ire (see Cumbre) Labor (La Labor) La Sierrita (see Sierrita) gas Los oe (see cn) Limon Luvianos Mesas (Las Mesas) Meson Viejo Mina de Agua Nanchititla eres Palmar a Chico) Pan Paso ae Vigas (see Vigas) Pen * oan = All-weather dirt road. Temascaltepec San Simon de Guerrero ? Tejupilco Texcaltitlan Tejupilco Tejupilco Temascaltepec Tejupilco Temascaltepec Tejupilco Tejupilco Temascaltepec Temascaltepec Temascaltepec Tejupilco Tejupilco Valle de Bravo? 19°12’ 18°57 13°51 19°03’ 18°48’ 19°02’ 18°50’ 18°55" 19°01’ 19°10’ 18°52! 19°08’ 99°53’ 99°59’ 100°18’ 99°56’ 100°09’ 99°58’ 100° 19’ 2900-3400 1960-2240 790 2500-2700 1000-1100 1900-2100 (640?) 910-960 1080-1450 000 2830-3300 1800-2000 1100-1850 1680-1960 ? ? ? Awdr to Toluca and Sultepec ? Awdr to Tejupilco and Bejucos ? Awdr to Tejupilco and Amatepec ? Awdr to Tejupilco and Bejucos Awdr to Tejupilco and Valle de Bravo Paved secitWe to Toluca and Temascaltep: Awdr to Tejupilc Awdr to Tejupilco and Bejucos Awdr to Tejupilco Fir-pine Pine-oak Pine-oak-Alnus Oak Pine-oak Oak Oak Oak Pine-fir Pine-oak Pine-oak Oak-pine Oak ? [ZZ61 NOLNIH 4 ADYOUD TASMOGAZA ¥ NOLNIH 69T ESTADO DE MEXICO DIstRICcT OF TEMASCALTEPEC APPROXIMATE PREDOMINANT CoorDINATES ALTITUDE LocaLitry MUNICIPALITY LAT.N. LONG. W METERS Access Roaps VEGETATION tanal Tejupilco ? ? Oak laza de Gallos Tejupilco 18°48’ 100° 18’ 350-1200 ? Puerto Salitre (see Salitre) Pungarancho Tejupilco 16°53" 100°35’ ? ak cén Temascaltepec 19°01’ 99°58’ 1930-2400 ? Oak-pine Rincon del Carmen Tejupilco 18°53’ 100°08’ 1340-1460 ? ak Salitre (Puerto Salitre) Tejupilco 18°53’ 100°17’ Awdr to Tejupilco and Bejucos San Lucas del Maiz ? Pine-oak San Mi Temascaltepec 19°08’ 99°53’ 2730-2770 ? Pine-fi Sierrita (La Sierrita) Texcalitlan 18°58’ 99°55’ 2200-2480 += Awdr to Toluca and Sultepec Oak-pine-Alnus Socabén Temascaltepec 19°02’ 99°58’ 2320 Oak-pine Tejupilco Tejupilco 18°55’ 100°09’ 1290-2000 Awdr to ee and Valle de Oak aaa tana Valle de Bravo? 19°06’ 100°04’ 1800-1960 Oak — Temascaltepec 19°03’ 100°02’ 1700-2080 Paved highway to Toluca Oak-pine Tena Zacazonapan? 19°04" 100°13’ 1220-1720 Oak sivas Temascaltepec 19°04’ 99°57’ 2480-2800 Awdr to Comunidad Oak-fir Timbres Valle de Bravo? 1660-1900 ? Oak Tule Valle de Bravo? 1500-1800 ? Oak Vigas (Las Vigas, Paso de igas Tejupilco 18°55’ 100° 15’ 1080-1300 Awdr to Tejupilco and Valle de Bravo Villaneda (Villa Neda) Tejupilco Volcan (Volcan de Tejupilco) Tejupilco 18°51’ 100°09’ 1380-1530 Awdr to Tejupilco and Amatepec Oak Ypericones (Ipericones includes Cerro Mufieca) Tejupilco 18°48’ 100° 16’ 1500-2300 ? Oak OLT WOLAXOPTAVY GTIONYV AHL AO TVNUNOL ¢¢ “10A] ae ey tor} 30° i ant — aul hia alin e jot geo | ent, Mango s San FAlipe | | ( We pene a wee : VOLCAN @ URUAPAN A me, : COLIMA ‘ *Tancitaro % | ©, \ sense B. Juarez \ a / @ Los, Moja Arete . : Los Aiyhores “a te-Fipride, jandio pe 12 oe te ay > °°22eeeeee TACAIMBARO bere cages e TOLUCA 1 2 Ze * Meson ov, ipa TA TEPALcATEPEC | » Acahuato : 4 Fe e * Son Miguel NEVADO - @ Buenavista a -) G / o 5 ‘ 2 DE a \° Ss APATZINGAN@ i = la gee TOLUCA Pens \* 4 y 3 Crucero * qa i ¥ ‘d 4 bs ” | % Obregon ‘ \ + ¢ oN Tequisquicpen H | eco! a se a a Poe © Tingambato Vorise 4 1 e Tiquiluca \ } Pet a ee TTI 2 soba iy 4 % \" Ss z % | ‘ . st ‘s \ «Paredes aca 3 cS i NE cape d * Puerto de ity, s | oPizono \ oe 4 v a : ee Cumbre ‘ - 7a ¥ sereavae” > nchititla ‘ s | thane «/Salitre ~“ x: / mA . fp ae A J Luvionos . s 1 : * San Mateo 3 od ames Co 995 © Chorreras o L P % Fits SF Be AL L avy oe \ Sto. Domingo ( 4 L & ae eae Esto * Te} H arotas a = x | a | *ee, “Aquale Sate) iY | Anse” ‘ ogi: a ragratte sounygbol oy . Peta Bile ee H | » Pie dela Cueste %e, / Guadalupe , 7x id “y prpeaigy oe ye bea spss H _— oChecalepa Torrecilias ry \ ia > “tg & xtapane @ SULTEPEC r | aw a a Pto. Bersenor, @COALCOMAN © Rancho Viejo iy R10 OFC \ eck i Tuncato eo i <= j tarron lp Coahuayana Pto. del Aire « 2 Purucho > ee =e In \ ii = * Renae i > Aguililla Que a ae a ee, Bejucos § 8°. * | Villa Victoria | e Barroloso 5 ” * e El Valle : a a a ey Pochote 2 Se H “ a | 4 e algerama . oAmatepec Zacoaigan ~—t ) S ae — ints: tveenaide moa SULT i Cc a: MEX/CO 4] eEI Moreno Se » ane apa a Carrizo gf. Pr istoba ZINACANTEPEC 4 | | SKE \ A | W AP | ceutla 1 Repare | | El Corinto San Juan , | wae | MINA (COYUCA) ST) ee es ce ce a | age GALEA NA Se GUERRERO PLAN ag LOCALIT ae @ San Andres s ae ia: a Saree oe SF if zie oid ee mecca oe re reas pacar aarmam sp onr A, MONTES DE OCA sh | ant iags arriza | Tierra Colorada » sruleueine 2 COLLECTED BY G.B. HINTON ETAL. | HUE TANMO 9 4 semaines A COMPILATION OF THE BEST | -, // *?MICHOACAN | | URUAPA 12 | AVAILABLE DATA ADJUSTED | | | APATZ/INGAN 43 ARTEAG. 14 WITH DIFFICULTY TO THE | eg COALCOMAN S| Be SS |} = — ___Coyco de Benitez o TOPOGRAPHY OF THE FIELD — re | SCALE 1: 25 Drawn By G.B. Hinton ur. Date: 2/15/39 | | | | | | | | | | | | | onsen" q 10315’ "00" isin jo2* 30" ear i Bre : | _— — : a seat — — ses aoe Abe BI) 101° 00" (00*|45" 100"|30 Agua — Icaparros Carrizo (see Plan de Carrizo) El Plato (see Plato) El Reparo (see Reparo) El Verde (see Verde) orida La Soledad (see Soledad) escaltepec — (Plan del Molino) Io So a Pie de la Cuesta Piedra Ancha Plan de Carrizo (Carrizo) Plato) o de Rio de Sea (Santiago) San Andrés San Juan de las Flores San Vice Santiago (se ag de Santiago) Santo Dom Sierrita Atoyac Tecpan Tecpan Atoyac Atoyac Atoyac Tecpan GUERRERO DIstRIcT OF GALEANA 100°21’ 1700-2200 100°57’ 260-560 100° 27° 0-660 100°38’ 250-5 100°57’ 760-880 100°23’ 120-140 100° 23’ 750-850 100°36’ 175-250 100°40’ 600 100°21’ 1800-3000 100°19’ 2500-3150 100°23’ 550-1050 100°39’ 800-1400 100°16’ 1000-1160 100°36’ 75-100 100° 14’ 2820 100°14’ 100°21’ 680-700 100°19’ 100°22’ 380-620 100°51’ 0-140 100°18’ 100°21’ 850-1500 100°37’ 450-950 Paved highway from Acapulco Paved highway from Acapulco Pine Oak-pine Oak-pine Oak Oak-pine Oak- ~pine Oak-pine Oak Oak-pine Pine Pine Oak-pine NOLNIH “4 ADYOAD ‘IMSMOGAZY ¥ NOLNIH [7261 TLT GUERRERO DIstTRIcT OF GALEANA APPROXIMATE PREDOMINANT CoorDINATES ALTITUDE Forest LocaLity MUNICIPALITY LAT.N NG. W METERS Access Roaps VEGETATION Soledad (La Soledad) Atoyac 17°F 100°14’ 1080 Oak-pine Tec Tecpan +7719" 100°37’ 20 otepec Atoyac 17°28’ 100° 10’ 1800-3300 Awdr to Xochipala Oak-pine Toro Muerto Atoyac ae 100° 16’ Verde (El Verde) Tecpan 17°16’ 100°37’ DIstRICT OF BRAvos Pueblo Viejo Tlacotepec 17°30 Ora 2000 Oak Puerto Hondo Tlacotepec 17°31° 100°05’ 2100 Pine District oF Mina (Coyuca) Aguacate Coyuca 18°07’ 101° 17 1100 Oak Aguazarca Coyuca 18°07’ 101°00’ 1400-1760 Oak Alacranes Zirandaro 18°14’ 101°0 borejo Cutzamala 18°40’ 100°32’ Ancoén (El Ancén) Coyuca ie 2" 100°41’ Animas Coyuca 13°27" 100°41’ Anonas Coyuca 18°20’ 100°52’ 200-360 i Arcelia 18°18’ 100° 16’ Paved highway to Iguala rmenia Zirandaro 18°08’ 101°19’ 2000-2400 Oak-pine Arrastras (Las Arrastras) Totolapan 18°01’ 100°18’ 600 ne agg (see Valderrama) Bar Bejuco | sem Bejucos) Coyuca 18°12" 100°54’ 400 rasiles Log es) Coyuca 18°10’ 100° 52" 400 aa _ Totolapan 17°46’ 100° 16’ 1650-2275 Oak-pine ZLI WOLAYOTUV GIONUV AHL AO TVNUNOL €S$ “I0A] Calavera Camarén (Camerén) Campo Morado Carriceras (Carrizera) Casas Viejas Cayunche (see Cuayunche) Cerro Azul Chacamerito Chilacayote Chirapitiro (see Zirapitiro) Chiriagua Cigarrillo Ciudad Altamirano pea ‘cag de Catalan) Cruz P. Cuahuilote (Cusitore) Cuayunche (Cayu Baia Cuevas Cuirio (Quiri Cundan “Candin Chico) Cutzam El satay (see Ancén) Espinazo a Diablo Filo May ation Fresnos nes Guayameo Coyuca oyuca Tlacotepec oyuca Ajuchitlan Coyuca Ajuchitlan Ciudad Altamirano Ajuchitlan Coyuca Coyuca Ajuchitlan Coyuca Tla Zirindaro 18°01’ 101°15’ 17°37’ 100°39’ 18°18’ 100°34’ 17°37? 100°38’ 17°54’ 101°02’ 18°12’ 100°55’ 15°21" 100°40’ 17°52" 101°00’ 17°34’ 100°37’ °O1’ 18° 101 main crest of the Sierra Madre del Sur 18° 1 14’ 18°12’ 101°17’ 18°06’ 100°59’ 18°17’ 101°15’ 450-1000 900-2780 1000-1500 1300-1750 1500-1640 2250-2550 1450-1675 1580-1860 1350 900-1320 2320-2520 1500 1040-1400 1260-2180 560-900 Awdr to Ciudad Altamirano Awdr to Iguala Awdr to Iguala Dirt road to Coyuca Awdr to Ciudad Altamirano Awdr to Iguala and Coyuca Dirt road to Zirandaro Oak-tropical Oak-pine Oak Oak Oak Oak-pine Oak-pine Oak-pine Oak-pine Oak Oak-pine-fir Oak-pine Oak-pine Oak NOLNIH ‘4 ADYOAD ‘TASMOGAZA ¥ NOLNIH [Z261 A | LT GUERRERO District oF MINA (Coyuca) APPROXIMATE PREDOMINANT CoorpDINAt ALTITUDE Locality MUNICIPALITY LAT.N LONG METERS Access ROApS VEGETATION Idolo Coyuca 18°03’ —-101°16’ 1140 Jaripo Coyuca 18°18’ 100°44’ Dirt road to Coyuca Jiotes (Jiotis) 18°06’ 100°57’ 500 Juntas (Las Juntas) Tlacotepec 17°S4 100° 10 1500 La Soledad (see Soledad) Laguna Zirandaro 18°18’ 101273" 1500-2500 Oak-pine agunas Zirandaro 18°09’ 101°18’ 1900 Las Arrastras (see Arrastras) Las Juntas (see Juntas) as Rosas (see oa, Llanos de Zacatlan (see Zacatlan) Los Barrales (see Barrales) a on yuca 18°07’ 100°59’ 1100-1760 Oak Man Ajuchitlan 17°42! 100°38’ 1500-1750 Mesa “ axle aca Queisle) oyuca 18°03’ 101°04’ Mesa del Frij Coyuca 18°03’ 101°11’ 1000-1140 Oak no oyuca 18°02’ 101°07’ 800 Nancintla Totolapan 17°43’ 100° 13’ 1460 Oak Ojo de Agua Cutzamala 18°43’ 101°30’ Ojo de Agua Coyuca 18°00’ 101° 15" 1500-1640 Oak Otatlan Totolapan 17°40’ 100° 13’ 950-1550 Oak Palma 18°03’ 101°15’ 1050-1100 Paracho Zirandaro 18°10’ 101°23’ 2300-2400 Pine-oak Paraje Javalin Totolapan 17°47’ 100° 16’ 2000-2500 i Coyuca 18°17! 100° 52’ 320 Parotas Coyuca 18°08’ 100°58’ 800 Oak WOLAIYOdUV GIONUV AHL JO TVNUNOL ¢S “I0A] Paso de Arena Paso de San José Patambo Petlacala Pilas inzan Morado Placeres (Placeres del Oro) Pochote Poleo Poliutla (Policla) Puerto del Oro Puerto del Clarin Puerto Puerto Sus tata ta Ciudad tamirano Querendas Quirio (see Cui Rancho ola oe Bejuco) Rio ety Ros: fas (Las Rosas) San Migu San ot (San Nicolas del Oro) San Pedro an Rafael Santa Ana (see Santana) Santa Barbara Santa Rosa Coyuca Cutzamala Zirandaro Ajuchitlan Zirandaro iudad Altamirano Coyuca Totolapan Totolapan Zirandaro Altamirano yuca 18°07’ 18°20’ 17°54’ 17°40’ 18°12’ 18°18’ 18°18’ 101°08’ 100°38’ 100° 38’ 400-800 750 1600-1850 1340 1260-2000 1550-1600 230-320 1400-1600 800 Dirt road to Coyuca Dirt road to Cutzamala Dirt road to Coyuca Awdr to Iguala Dirt road to Ciudad Altamirano Oak-pine Pine Oak-pine Oak-pine Oak Pine NOLNIH ‘4 ADYOAD ‘IMSMOGAZY ® NOLNIH [ZL61 SLT Locality Santa Ter Santana ena Ana) Soyate (Zoyate) Suspiro (see Puerto Suspiro) Tanganhua Tario : Tehuehuetla (Tejuejuetla) Teote Ticuiches (Tequeches) Tierras Timbre Tina Toro acas Valderrama (Balderrama) etarrén Vinatita Yesqueros (Yesceros) Zacatlan (Llanos de Zacatlan) Zapo (see Sapo Zihuaquio (Zihuagio) jas Mue ecw Gan Miguel Totolapan) cher MUNICIPALITY Zirandaro Coyuca Cutzamala Cutzamala oyuca Totolapan Coyuca Coyuca GUERRERO District or Mina (Coyuca) APPROXIMATE PREDOMINANT CoorDINATES ALTITUDE s LAT, LON METERS Access RoaAps VEGETATION 18°16’ 101°05’ 450-1040 Oak 8°04’ 101°13’ 1100-1350 Oak 18°23 100°43’ 17°55" 101° 740-860 18°09’ 101°21’ 2140-2400 Oak-pine is"t7’ 100°35’ 18°25’ 100°45’ 17°48’ 100°25’ 1200 . 1500-3600 Oak-pine-fir 18°14’ 101°18’ 1650-1880 Oak-pine 8°09’ 101°17' 1400-1800 Oak 18°15’ 100752" 400 18°27 100°41’ 18°31’ 100°16’ —- 1800-2900 Oak-pine 18°01’ 100°18’ 600 8°05’ 101°16’ 1200-1300 18°44’ 100° 28’ Dirt road to Cutzamala 8°04’ 101°10’ 18°43’ 100°31’ Dirt road to Cutzamala 18°43’ 100°29’ Dirt road to Cutzamala 8°06’ 101°16’ 1250 17°40 100°29’ 1600-2250 Oak-pine 8°05’ 101°07’ 1500 18°02’ 101°14’ 1050-1100 Oak 9LT WOLAYOTUVY GIONUV AHL AO TYNANOL ¢€S “I0A] Zirandarangio Zirapitiro (Chirapitiro) Zoyate (see Soyate) Cc Carboneras Ciruelo (Ciruelos) Filo Guadalupe (Mineral de Guadalupe, Real de ntonio Soledad (Soledad Palma) Tuzal) Vallecitos Acahuato Aguaje ililla Apatzingan Buenavista Coyuca Cutzamala Petatlan 18°20’ 100°41’ ta 35" 100°33’ DIsTRICT OF MONTES DE OCA Figs & TOs 1d 560-720 560-800 17°47’ 1001’ 840-960 La Union main crest of the Sierra Madre del Sur La Union La Union Apatzingan Tepalcatepec Aguililla Apatzingan Buenavista 18°01’ 101°19’ 1150-1400 $7753 LOL" 21" : 17°59' 101°18’ 900-1200 17°35 100°59 700 17°39’ 101°17’ 100 17°38" 101°06’ 400-1000 17°3V' 101°16’ 100 17°49’ 101°07’ 2100 Aw ee 101°17’ 800-1040 17°50’ 101°08’ 720 0 17°55" 101°19’ 400-760 MICHOACAN District oF APATZINGAN 19°10’ 102°20’ 800-1400 19°00’ 102°43’ 200-350 800-1100 19°04’ 102°22" 300-400 19°12" 102°33' 400-800 Dirt road to Cutzamala Paved highway to Acapulco Paved highway to Acapulco Awdr to Apatzingan Awdr to Apatzingan Oak-pine Oak Oak Oak-pine Oak Oak-pine Oak-pine Pine Thorn Thorn Tropical-thorn NOLNIH ‘€ ADYOAD TMSMOGAZA ® NOLNIH [ZL61 LET MICHOACAN District oF APATZINGAN [cont. ] APPROXIMATE CoorDINATES ALTITUDE LocaLity MUNICIPALITY LAT.N. LONG, W. METERS Access Roaps Rancho Viejo Buenavista 19°12’ 102°40’ 700 Awdr to Apatzingan Tepalcatepec Tepalcatepec 19°11’ 102°50’ 300-400 Awdr to Apatzingan Tiquiluca Tepalcatepec 700 Awdr to Apatzingan DistrIcT OF URUAPAN Tancitaro Tancitaro 1600-3900 DIstRICT OF COALCOMAN quila Aquila 18°37’ 103°30’ 50-250 Awdr to Tecoman, Col. Barroloso 1300-2900 Chacalapa 150-250 Coalcoman Coalcoman 18°46’ 103°09’ 1000-1250 Awdr to Apatzingan Huizontla Villa Victoria 18°09’ 103°24’ 350-1050 rta Coalcoman 18°36’ 103°07’ 850-1700 tula quila 18°30’ 103°29’ 550 arotas Coalcoman 750-800 Puerto Las Cruces Coalcoman 1200-1450 Puerto Zarzamora 1200-1800 Salitre Coalcoman 1100-1800 San José Aquila 750-1600 San Pedro Aquila 18°15’ 103°06’ 450-600 Sierra Naranjillo Coalcoman 450-1550 Sierra Torrecillas Coalcoman 1400-2700 Tizuapan (Tizupa) Aqu 18°11" 103°02’ 0-50 Troj Coalcoman 18°57’ 103°21’ 400-500 Villa Victoria Villa Victoria 00-1600 PREDOMINANT VEGETATION Tropical-thorn Tropical-thorn Tropical Oak-pine-fir Tropical Oak-pine-fir Tropical ep ine Oak & — Oak-p' Tropical oe -pine Pin Pine & tropical Tropical Tropical Tropical & pine WOLIXOIUVY GIONUY AHL AO TYNANOL €$ “IOA] Huetamo Tiquicheo Aguila (El Aguila) Ahorcados (Los Ahorcados) Alumbres (Los Alumbres) Angangueo Aputzio Benito Juarez (see Laureles) u Campana (La oe) raged (Las Canoas) Capu Cerro de los Magueyes adie Pein (Pelon Hill) a Lee Coyota Dos uce) Enandio (Nandio, Salto de Enandio) Huetamo Tiquicheo Zitacuaro Zitacuaro Zitacuaro Zitacuaro Zitacuaro District oF HUETAMO 18°36’ 100°54’ 18°34’ 100°47’ 18°25 100°36’ 18°55" 100°45’ 380 450-600 DIsTRICT OF ZITACUARO 19°36’ 100°18’ 19°22" 100° 18’ 19°20’ 100° 28’ 2700-2800 2400 1 2100-2200 950-2300 2350-3350 2000-2300 2600-2650 2600-3600 1950 1900-2600 1700-1900 1800 1500-1850 Awdr to Morelia Awdr to Huetamo Awdr to Morelia Paved highway to Zitacuaro Awdr to Zitacuaro Awadr to Zitacuaro Tropical Tropical Pine-oak Pine-oak Pine 0. Oak Pine-oak Oak Oak-pine Oak Pine-oak Oak-pine NOLNIH ‘€ ADYOAD TYSMOGAZA ¥ NOLNIH [7261 6LT MICHOACAN District OF ZITACUARO APPROXIMATE PREDOMINANT a ALTITUDE Forest LocaLity MUNICIPALITY LAT.N, G.W. METERS Access ROAps VEGETATION Encarnac Zitacuaro Awdr to Zitacuaro Florida (La Florida) Zitacuaro 1300-1600 Tropical Zitacuaro 2400-2500 Guajes (Los Guajes) zantla 1000-1350 Tropical Guanoro Benito Juarez 1500-1900 Oak Ipazote (Ypazote hill) Zitacuaro 2150-2450 ak-fir Jungapeo Jungapeo 19°28’ 100°30’ 1500-1850 Paved highway to Zitacuaro Tropical Jurungueo (Jurungeo) Zitacuaro La Campana (see Campana) La Florida (see Florida) Las Canoas (see Canoas Las Lomas (see Lomas) as Pilas (see Pilas) Laureles (Benito Juarez) Benito Juarez 19°18’ 100°27’ 00-1700 Awdr to Zitacuaro Oak Zitacuaro ae t950 Oak Lomas (Las Lomas) Zitacuaro 1900-1950 Los Ahorcados (see Ahorcados) Los Alumbres (see Alumbr es) Los Guajes (see G Los Macheros (see Macheros) Los Pinzanes (see Pinzanes) Los es ae Tepehuaje: — (Las Macheros) Zitacuaro 2300-2400 Oak Macho Zitacuaro 2300-2400 Oak-pine ango Zitacuaro 2000-21 Manzanillos Zitacuaro 2100-2200 Paved highway to Zitacuaro Ost WOLAYOTAV GIONUV AHL AO TVNUNOL ¢¢$ “I0A] Nihuatlan Ocu Pelén 0 Cerro Pelén) Piedra Pilas (Las Pi i Pinzanes (Los Pinzanes) Puerto (El Puerto) Rincon Salitr San hades (San Andrez) San Feli San José del Rio San José Purua a Ana Sauce oe Sauce) Seis P Sicate yl Zicate ) a Tepehuajes (Los Tepehuajes) Tiamoro aba (Tiripitillo) hed ntle Ypasote hill A Tpazote) _ - te) Zir ~ie Donato Guerra, Estado México ? Zitacuaro Tuzantla Zitacuaro Tuzantla Zitacuaro Zitacuaro Tuzantla Benito Juarez Zitacuaro Zitacuaro [J.R.] ESCUELA NACIONAL DE , ae eee INSTITUTO POLITECNICO san Mexico 17, D. F., MExic 19°35’ 19°12’ 19°26’ 100° 28’ 100°33" 100°22’ 2500-2550 2000 950-1300 1450-1550 2050-2150 1900-2000 1900 1450-1950 1325 1550-1900 2400 2100-2400 00 1900-2900 1600-2200 Paved highway to Zitacuaro Pine-oak Tropical Tropical Tropical Oak Tropical Tropical Pine Tropical Tropical Oak Oak Paved highway to Mexico and Morelia | | [J. H.] Av. JALIsco 180 5° Piso TacusayYa, D. F., México [z61 NOLNIH ‘4 ADYOAD TYSMOCGAZA ¥ NOLNIH 182 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 THE GENERA OF GERANIACEAE IN THE SOUTHEASTERN UNITED STATES 1 KENNETH R. ROBERTSON GERANIACEAE A. L. de Jussieu, Gen. Pl. 268. 1789, “Gerania,” nom. cons. (GERANIUM FAMILy) Annual, biennial or perennial herbs [infrequently shrubs to trees; stems sometimes succulent] with simple and often glandular trichomes; under- ground stems, rhizomes [enlarged tubers and roots| frequently produced. Leaves opposite or alternate, palmately or pinnately lobed or compound [rarely simple and entire], palmately veined, stipulate, the petioles some- times persistent. Inflorescences terminal or axillary, variable but basically cymose and sometimes condensed into umbels [rarely solitary], bracteate. Flowers perfect [or imperfect] and regular to slightly [or distinctly | irregular, 5-merous [rarely some parts fewer by abortion]. Sepals mostly distinct, persistent, imbricate. Petals distinct, often emarginate, imbricate, or convolute in aestivation [rarely absent]. Androecium obdiplostemonous [or of 5 bundles of three stamens], the stamens all fertile or some re- duced to staminodia or scales; filaments connate at the base or distinct: anthers versatile, 2-locular at anthesis, opening lengthwise; nectar glands usually alternating with the petals. Gynoecium of 5 carpels united to an elongated columella, styles united to the columella, forming a beak on the ovary and fruit [or almost absent], 5-fid at the apex; stigmas distinct, ligulate to clavate Jor filiform], ovary superior, 5-lobed 5-loculate, the ovules 2 in each locule, anatropous to campylotropous, superposed and “Prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University made rincipal ood : established in the first paper in the series (Jour. Arnold Arb. 39: 296-346. 1958). The area covered includes North and South Carolina, Geor ia, Florida, Tennessee, Alabama, Mississippi, Arkansas, and Louisiana. The descriptions ar rimarily on the plants in this area, with additional information from extraterritorial taxa in brackets. References I have not seen are ma by an asteris - Duncan M. Porter (Missouri Botanical Garden and Washington University) for carefully reviewing the manuscript. Mrs. i by Dr t for the generic flora project has been consulted. Figures a and c-m were d lapman; figure b is the work of the late Dorothy r. Wood collected most of the plant materials and prepared the dissections for the illustration et has kindly called attention to his nding of Geranium texanum in three parishes in Louisiana and has sent documenting specimens (deposited in a). 1972] ROBERTSON, GENERA OF GERANIACEAE 183 pendulous from axile placentae. Fruit a schizocarp, usually dehiscing elastically into 5 one-seeded mericarps that separate shy the persistent columella with a part of the style forming an awn on the apex of the mericarp, the awn hygroscopic, becoming revolute or spiralled. Seeds embryo development of the Asterad to Onagrad types. (Geranieae of Engler & Prantl, Knuth, and Scholz.) Type GeENus: Geranium L. Five genera and about 700 species widely distributed in temperate and subtropical regions of both hemispheres; essentially restricted to high altitudes in the tropics. Geranium and Erodium, with both cei and naturalized species, represent the family in North America. A n ber at iene of Pelargonium L’Herit. ex Aiton, the horticultural pena) ums,” a predominantly South African genus of about 250 species, are commonly cultivated in our area but are not naturalized although several are ephemeral escapes in California. Sarcocaulon (DC.) Sweet (six spe- cies) is South African, and Monsonia L. (about 30 species) occurs both north and south of the Sahara Desert in Africa and eastward to India. These five genera comprise the Geranieae of Engler & Prantl, Knuth, and Scholz, and the Geraniaceae of Airy Shaw (in Willis), a very natural group with two ovules per carpel and the carpels in fruit separating into one-seeded mericarps, each with a curved or curled awn. The Bieber- steiniaceae, Dirachmaceae, Ledocarpaceae, and Vivianiaceae, which differ in the number of ovules per carpel and in fruit structure, are often included in the Geraniaceae, but the taxonomic affinities of these families need fur- ther study. The Biebersteiniaceae have recently been placed with the Rosaceae (Bortenschlager), the Dirachmaceae in the Tiliales (Hutchinson), and questionably related to the Greyiaceae (Airy Shaw), the Ledocar- paceae in the Malpighiales (Hutchinson), and the Vivianiaceae in the Pittosporales (Hutchinson) or in the Centrospermae (Bortenschlager). There is rather general agreement that the Geraniaceae, sensu stricto, the Oxalidaceae, and the Tropaeolaceae belong together in a single order. The schizocarp of the Geraniaceae is rather complex, and differing and inconsistently applied terminology has created confusion. At the base of the fruit are five mericarps (cocci, carpel bodies) which reflect the five lobes (locules) of the ovary. Arising immediately above is an elongated and more or less cylindrical rostrum (stylar column, beak), the apex of which abruptly tapers into the stylar beak, a narrow cylinder five-fid at the apex. Passing through the center of the fruit from the receptacle to the tip of the rostrum is a persistent columella (central column, stylar column, or rostrum), which has been said variously to be receptacular, ovarian, or stylar in origin. At maturity the fruits dehisce elastically or even explosively with the mericarps and awns separating from the colu- mella. The awns, evidently strips of stylar material that before dehiscence form vertical bands on the rostrum, are hygroscopic and at dehiscence become revolute or spiralled, usually remaining attached to the mericarps 184 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 and either remaining affixed to the apex of the rostrum or not. The seeds are either ejected from the mericarps or retained within them. Reported chromosome numbers are Sarcocaulon Burmannii, 2n = 44; Monsonia senegalensis, 2n = 24, M. longipes, 2n = 38; Pelargonium, 2n = 9 (haploid sporophyte), 14, 16, 18, 20, 22, 32, 36, 38, 44, 45, 54, 56, 60, 66, 72, 77, 81, 88, and 90, with 18 and 22 occurring most frequently; Erodium, 2n = 18, 20, 27, 30, 36, 40, 48, 54 (567), 60; and Geranium, 2m = 18, 20, 22, 24, 26, 28, 32, 34, 36, 38, 42,48, 50, 52, 56, 64, 68, 84, Pollen grains of the Geraniaceae are 3-colpate or 3-colporate with striate to reticulate exine sculpturing. Bortenschlager recognized seven distinct pollen types based on aperture type and sculpturing patterns. Most species of Erodium lack supratectate processes, while most of those of Geranium have mono- or dimorphic processes, although the pollen of certain species in both genera is indistinguishable. Monsonia and Sarcocaulon are palyno- logically identical, and Pelargonium has several distinct pollen types. The studies of Konrad Sprengel, who first observed insect pollination scientifically and proposed the theory of cross-pollination, began with observations on Geranium sylvaticum. Since that time the pollination of numerous species of Geranium, Erodium, and Pelargonium has been in- vestigated. The petals of most species are usually some light to dark shade of red to blue with the more intensely colored veins forming nectar guide lines that converge toward the center of the flower. Some species also have darker-colored spots toward the base of one or more petals, and in Erodium and Pelargonium there is a trend toward irregular corollas. A nectar-secreting spur of the uppermost sepal is adnate to the pedicel in Pelargonium species, and each of the five sepals of Monsonia parvifolia has a small spur. Species of the other genera have five nectar glands alternating with the petals. In some species the odor of the flower is evi- dently more important than corolla markings in the attraction of insects. Hymenoptera, and to a lesser degree, Lepidoptera, Diptera, and Coleoptera, are important pollinators; some large-flowered Pelargonium species of South Africa are visited by sun birds (Nectariniidae). Numerous chemical substances have been identified in the Geraniaceae. Tannins have been reported in Geranium, Pelargonium, Monsonia, and Erodium and are especially abundant in the subterranean parts of peren- nial Geranium species. Gallic and ellagic acids are known from Erodium cicutartum and several species of Geranium, including G. maculatum, and catechins have been isolated from G. pratense and G. palustre. Essential oils occur in the glandular trichomes of a few Geranium species and in many Pelargonium species and hybrids. Most Pelargonium species accumulate tartaric acid, but those of Geranium and Erodium do not. The bark of xerophytic species of Sarcocaulon is rich in aromatic resins and waxes. The numerous species, hybrids, and cultivars of Pelargonium, the widely grown ornamental garden “‘geraniums,” account for the principal economic importance of the family. In addition to the commonly cultivated pelar- goniums, several species of Geranium and Erodium are grown for their 1972] ROBERTSON, GENERA OF GERANIACEAE 185 interesting appearance or showy flowers. Geranium oil used in perfumery, ten as a substitute, extender, or adulterant for more expensive essences such as attar of roses, is extracted from several species of Pelargonium, particularly P. graveolens and P. odoratissimum. Extracts from various species of Geranium are used as astringents and aphrodisiacs and for tanning and dyeing, while extracts from Monsonia ovata and Pelargonium antidysentericum are used for the treatment of dysentery. Erodium cicu- tarium and some other species are important forage crops, but sometimes become rampant weeds. REFERENCES: ANpDREws, H. C. Geraniums: or, A monograph of the genus Geranium: con- taining coloured figures of all the known species and numerous beautiful varieties, drawn, engraved, described, and coloured, from the living plants. 2 vols. 124 pls. Loken: 1805. [Mostly Pelargonium spp. | BaILton, H. Géraniacées. Hist. Pl. 5: 1-41. 1874. [English transl. M. M. Hartoc, The natural history of plants. 5: 1-41. London. 1878.] BENTHAM, G., & J. D. Hooxer. Geraniaceae. Gen. Pl. 1: 269-278. 1862. [Treatment by BENTHAM; family divided into 7 tribes. | BOLKHOVSKIKH, Z., V. Grir, T. MAtveyeEvA, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. Ed. by A. A. Feporov. (Russian and English prefaces.) 926 pp. Leningrad. 1969. [Geraniaceae, 331-333. ] BoRTENSCHLAGER, S. Vorlaufige Mitteilungen zur Pollenmorphologie in der Familie der Geraniaceen und ihre systematische Bedeutung. Grana Palynol. 7: 400-468. 1967. [Comprehensive sampling of 241 taxa in 9 genera; 4 segregate families recognized. Brouwer, W., & A. STAHLIN. Handbuch der Samenkunde. Introd. + 656 pp. Frankfurt am Main. 1955. [Geraniaceae, 200-204 BRUNIEsS, S. Anatomie der eran scoere es 1 in Beziehung zur Systematik der CaNDOLLE, A. P Geraniaceae. ee 1: 637-682. 1824. Caroiin, R. C. The genus Pelargonium L’Hér. ex Ait. in Australia. Proc. Linn. Soc. New S. Wales 86: 280-294. 1962. [Taxonomic treatment; notes and diagrams on habit types. | —. Geraniaceae. Jn: C. G. G. J. vaN STEENIs, ed., Fl. Males. I. 6: 445- 449. 1964. [Includes summary of phytochemistry by R. Hegnauer. | Carotuers, Z. B., Jk. The comparative stem anatomy of some shrubby mem- bers of the Geraniaceae. 192 pp. 1958. Ph.D. thesis, The University of Michigan.* [Cf. Diss. Abstr. 19(8): 1904, 1905. 1959; species of Balbisia, Biebersteinia, Geranium, Monsonia, Pelargonium, Rhynchotheca, Viviania, and Wendtia examined. | . Observations on the procambium and primary phloem of Pelargonium domesticum. Am. Jour. Bot. 46: 397-404. 1959. [The “pericycle” fibers in the family actually are protophloem fibers. | ee J. Geraniopsis eine neue Gattung der Familie Geraniaceae. Novit. Bot. niv. Carolinae Pragensis 1967: 9. 1967. [G. trilopha (Boissier) Chrtek, : yemensis (Deflers) Chrtek. | Currrorp, D. Pelargoniums, including the popular ‘Geranium.’ ed. 2 pp. 15 colored pls. 83 figs. Blandford Press. 1970. [Horticultural infor- mation, brief descriptions of 219 spp. and 850 cultivars. ] 186 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Daker, M. G. Chromosome numbers of Pelargonium species and cultivars. Jour. Roy. Hort. Soc. 94: 346-353. 1969. [47 spp. or named hybrids in 13 sects. and 50 cultivars. | Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. 1966. [Geraniaceae, 127; references. Davis, P. H. Geraniaceae. Fl. Turkey 2: 450-488. ae [Includes 33 Geranium spp. in 5 “groups” and 20 Erodium spp. in 2 s Dawson, M. L. The floral morphology of the Peiuedaan Am. Jour. Bot. = haa 1936. [Includes Geranium maculatum and Erodium rhe ; discussion of 3 theories of obdiplostemony in family. | ce Mesh G. Pollen morphology and plant taxonomy. Angiosperms. Corrected reprint and new addendum. xiv + 552 pp. New York. 1966. [Geraniaceae, 186-189; references. | FEARN, B. Saine succulent Geraniaceae. Natl. Cact. Succ. Jour. 23: 15-18. 1968. [ Geranium, Pelargonium, Sarcocaulon; photographs GAGNEPAIN, F. Contribution a l’étude du pollen des Géraniacées. Bull. Soc. Hist. Nat. Autun 16: 83-97. 1903. [Geranium, Erodiu um, Pelargonium. | Gams, H. Geraniaceae. Jn: G. HEct, Illus. FI. Mittel- -Europa 4(3): 1656-1725. pls. 173, 174. 1924. Gaucer, W. Ergebnisse einer zytologischen Untersuchung der Familie der Geraniaceae. I. Planta 26: 529-531. 1937. [Geranium, 23 spp.; Pelar- gonium, 29 spp.; es 10 spp.; Sarcocaulon biopices, Hanks, L. T., & J. K. Smarty. Geraniaceae. N. Am. F 1. 25(1): 3-24. 1907. HAnnic, E. Uber den ie Rete der Skrelauschidun bei den Drisen- hiacén von Pelargonium. Zeitschr. Bot. 23: 0. Harney, P. M. A chromatographic study of a presumed ancestral to Pelargonium x hortum Bailey. Canad. Jour. Genet. Cytol. 8: 780-787. ee le R. Chemotaxonomie der Pflanzen. Band 4. Dicotyledoneae: Daph- niphyllaceae-Lythraceae 551 pp. Basel & Stuttgart. 1966. [Geraniaceae, 193-201; referen HErmscH, C., Jr. Cie tive anatomy of the secondary xylem in the “Grui- and “Terebinthales,” of Wettstein with reference to taxonomic grouping. Lilloa 8: 83-198. pls. 1-17. 1942. [Geraniaceae, 98-100; wood structure parallels in certain respects that of Oxalidaceae Hess, H. E., E. Lanport, & R. Hirzet. Flora der Schweiz. Vol. 965 pp. _— & eet 1970. [Geraniaceae, 615-632; illustrations ee useful men die W. Uber den Bau des Pistills der Geraniaceen. Flora 47: 401-410. 864. Hutcuinson, J. Evolution and phylogeny of flowering plants. Dicotyledons: facts and theory. xxvi + 717 pp. 2 charts. London & New York. 1969. [ Geraniales, eke JANNICKE, W. rage zur vergleichenden Anatomie der Geraniaceae. Abh. Senckenberg. pete Ges. 14(3): 1-24. 6.* Kenna, G. Das oderm der ee Phyton Austria 6: Kers, L. E. Monsonia parvifolia Schinz loa Sa a species with concealed spurs. Bot. Not. 124: 208-212. 1971. [Each of the five sepals has a small, Ktorzscu, J. F. Bemerkungen zu den Geraniaceen und deren Verwandtschaften. Linnaea 10: 425-439. 1836. 1972] ROBERTSON, GENERA OF GERANIACEAE 187 Knutu, P. Handbook of flower pollination. Vol. 2. (Transl. J. R. A. Davis.) viii + 703 pp. Oxford. 1908. [Geraniaceae, 218-232. In German ed. cf. ] Knutu, R. Geraniaceae. Pflanzenreich IV. 129(Heft 53): 1-640. 1912. [11 genera in 5 tribes; comprehensive treatment of all known species. | . Pelargonium oil. Am. Jour. Pharm. 93: 302-315. i og ———. Geraniaceae. Nat. Pflanzenfam. ed. 2. 19a: 43-66. 1931. Koxwaro, J. O. The family ‘“Geraniaceae” in North-East tropical Africa. Webbia 25: 623-669. 1971. [Erodium, Geranium, Monsonia, Pelargonium. | . Geraniaceae. Fl. Trop. E. Afr. 24 pp : Komaroy, V. A., ed. Geraniaceae. Fl. URSS. (In Russian.) 14: 1-76. 1949. apm and Pelargonium by E. G. Bosrov, Erodium and Biebersteinia by A. I. VvEDENSKIJI; also see vi—viii for classification scheme. | Kucter, H. Bliitendkologie. ed. 2. xi + 347 pp. Stuttgart. 1970. [Erodium, Gera anium, pr at includes photograph of G. pratense taken with ultraviolet light. KUTSCHERA, i: Wurzelatlas mitteleuropdischer Ackerunkrauter und Kultur- pflanzen. xvi + 575 pp. Frankfurt am Main. 1960. [Geraniaceae, 372-378; drawings of root systems of Geranium pusillum, G. molle, Erodium cicu- tarium. Lapse, A. Sur |’ “éperon” de la fleur de Pelargonium. Bull. Soc. Bot. France 111: 321-324. 1964. [Ontogeny and anatomy of the spur in P. zonale. Lecautt, A. Recherches anatomiques sur l’appareil végétatif des Géraniacées. Théses Fac. Sci. con 245 pp. 4 pls. Lille. 1908. [Geranium, Erodium Pelargonium, and Monsonia. See also Compt. Rend. Acad. Sci. Paris 147: 382-384. 1908; vipa and Erodium a L’HERITIER DE BRUTELLE, C. L. Geraniologia, seu Erodii, Pelargonii, Geranii, Monsoniae et Grieli historia ao illustrata. 1 p. pls. 1-44. Paris. 1787-1788 [1792]. [Text unpublished, names validated by Aiton, Hortus Kewensis. Vol. 2. 1789. Lussock, J. A contribution to our knowledge of seedlings. Vol. 1. viii + 608 pp. London & New York. 1892. [Geraniaceae, 294-316. ] MartotH, R. Geraniaceae. Fl. S. Africa 2: 83-91. pls. 30-32. 1925. [Pho- tographs and colored plates of Monsonia, Sarcocaulon, Pelargonium and Geranium. MERXMULLER, Die stammsukkulenten Pelargonien des Liideritz-Distriktes cea andieige Mitt. Bot. Staatssam. Miinchen 5: 229-245. 1964. MririManorr, A. Remarques sur les glandes extra-staminales de quelques Géraniacées. Bull. Soc. Bot. Genéve II. 32: 195, 196. 1941. [Geranium, Erodium, Pelargonium; histochemical study of the secretory tissues. ] sprigs mon, L. Considérations sur l’inflorescence du Pelargonium zonale Mém. Soc. Bot. France 1964: 125-134. 1964. [The determinate umbel ee resulted from ane ache of biparous or uniparous cymes. Moore, H. E., Jr. Pelargoniums in cultivation. Baileya 3: 5-25, 41-46, 70-97. 1955. [Evaluation of names for species and hybrids grown in the U.S.; keys, illustrations. ] Mosepacu, G. Die Fruchtstielschwellung der Oxalidaceen und Geraniaceen. Jahrb. Wiss. Bot. 79: 353-384. 1934. [Includes Geranium dissectum and comments on Geranium, Erodium, and Pelargonium s Mutter, H. The fertilisation of flowers. (Transl. D. W. THOMPSON.) xii + 669 pp. ‘London. 1883. [Geraniaceae, 149-160. ] ~ 188 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Miter, T., J. R. Launpon, & L. F. Bowpen. Geraniaceae. In: A. W. EXELL, A. epee & H. Wixp, Fl. Zambesiaca 2: 130-149. 1963. [Geranium, ; Monsonia, 4; Pelargonium, 8.] Moet POL A seed key to fourteen species of Geraniaceae. Proc. Iowa Acad. Sci. 51: 241-246. 1944 [1945]. [Includes 2 pls. of seeds and mericarps.] Narayana, H. S., & P. K. Aropa. Floral anatomy of Monsonia senegalensis Guill. and Perr. Curr. Sci. Bangalore 32: 184, 185. 1963. Nevupaver, H. F. Uber Blattgrund und Stammknoten von Pelargonien. Osterr. Bot. Zeitschr. 119: 141-153. 1971. [English summary. ] OttMANN, QO. Pollenmorphologische- ctocrere sine Untersuchungen Innerhalb der Geraniales. 183 pp. pls. Lehre. : REIcHE, K. F. Spinone Nat. pte III. 4: 1-14. 1889. (Supple- ment by A. ENG Rickett, H. W. Wildflowers of the United States. Vol. 2. The Southeastern States. Part 1. x + 322 pp. pls. 1-116. New York. 1966. [Geraniaceae, 296-299, pl. 107.] Saver, H. Bliite und Frucht der Oxalidaceen, Linaceen, Geraniaceen, Trope- olaceen und Balsaminaceen. \V Unter- suchungen. Planta 19: 417-481. < pl. 1933. [Floral and fruit anatomy of Geranium sylvaticum, Erodium cicutarium, Pelargonium zonale. | ScHONBECK-TEMEsy, E. Geraniaceae. Jn: K. H. RecHinGER, Fl. Iranica 69: 1-67. pls. 1-8. 1970. [5 genera, 54 spp. ] ScHouz, H. Geraniaceae. Jn: H. Metcuior, Engler’s Syllabus der Pflanzen- familien. ed. 12. 2: 248-250. 1964. ScHtruHorr, P. N. Zytologische Untersuchungen in der Reihe der Geraniales. Jahrb. Wiss. Bot. 63: 707-759. 1924. [Geraniaceae, 708-714; development of pollen, embryo sac, and embryo. | ScHWIEKER, F. Untersuchungen iiber die Postflorationsbewegungen einiger Geraniaceen. Bot. Arch. 6: 206-230. 1924. [Geranium, Erodium.| S@RENSEN, W. Sur la structure du fruit de nos Géraniacées comment il se comporte au moment de la maturité. Danske Vidensk. Selsk. Forh. 1911(Meddel.): 99-137. 1 pl. 1911. [Erodium, Geranium, Pelargonium.] Starrorp, H. A. Distribution of tartaric acid in the Geraniaceae. Am. Jour Bot. 48: 699-701. 1961. [ Geranium, Erodium, Pelargonium. | Sweet, R. Geraniaceae. The natural order of Gerania, illustrated by coloured figures and descriptions; comprising the numerous and beautiful mule- varieties cultivated in the gardens of Great Britain, with directions for their treatment. 5 vols. 500 pls. London. 1820-1830 THURNHILL, J. W., R. K. Matra, & W. H. Woop. Examining three-dimensional microstructures with the scanning electron microscope. Grana Palynol. 6: 1-6. pls. 1-3. 1965. [Pl. 1, Nutmeg variety of Pelargonium, one of the first published SEM photomicrographs of pollen. | TRELEASE, W. A study of North American Geraniaceae. Mem. Boston Soc. Nat. nia . VON KircHner, E. Loew, & C. SCHROTER, Lebensgeschichte der Bliitenpdl. ee 3(3): 1-147. 1926. [Ex- tensive references to early literatu Warsure, E. F. Taxonomy and relationship in the Geraniales in the light of their cytology. New Phytol. 37: 130-159, 189-210. 1938. [Geraniaceae, 138-159, 189-195; chromosome counts sometimes inaccurate. ] Wess, D. A., ed. Geraniaceae. Jn: T. G. Tuttn, V. H. Heywoop, et al., eds., 1972 | ROBERTSON, GENERA OF GERANIACEAE 189 Fl. Europaea 2: 193-204. 1968. [Geranium by D. A. Wess & I. K. Fercuson; Erodium by D. A. Wess & A. O. CHater; Biebersteinia by EBB. D. A. WE Wits, J. C. A dictionary of the flowering plants and ferns. ed. 7. (Revised by H. K. Atry SHAw.) xxii + 1214 pp. + liii pp. (Key to the families of flowering plants.) Cambridge, England. 1966. [Geraniaceae, 473.] ZANKER, J. Untersuchungen iiber die Geraniaceen. Planta 9: 681-717. 1930. [Includes studies on inflorescences and embryos of Geranium, Erodium, and Pelargonium. | KEY TO THE GENERA OF GERANIACEAE IN THE SOUTHEASTERN UNITED STATES General characters: annual to perennial herbs with palmately or pinnately lobed to dissected, stipulate leaves; inflorescences terminal or appearing ec basically cymose, sometimes condensed into determinate umbels; flowers 5- merous, perfect, regular to slightly irregular, sepals mostly distinct, ui Hiate petals distinct, imbricate or convolute ; androecium obdiplostemonous, all stamens fertile or some reduced to staminodia or scales; gynoecium 5-carpellate with an elongated central columella, the ovary supierior, ovules 2 per carpel, only one maturing; fruit a beaked schizocarp dehiscing into 5 mericarps with a strip of stylar tissue forming a hygroscopic awn on each; embryo curved, endosperm absent. Leaves palmately lobed; all stamens with anthers (except in G. pusillum); ultimate peduncles with 2 flowers; mericarps rounded at base, ejecting the seed; awn of mericarp revolute, glabrous or puberulent on the inner surface. Se re ee ee ter ee eh gs oe Le go cee ae 1. Geranium. Leaves pinnately lobed or dissected; only the inner 5 stamens with anthers; peduncles terminating in a several-flowered umbel; mericarps pointed at the wi retaining the seed; awn of mericarp spiraling, pubescent on the inner BU a er rns eer ag ap Es pa eee cre 2. Erodium. 1. Geranium Linnaeus, Sp. Pl. 2: 676. 1753; Gen. Pl. ed. 5. 306. 1754. Annual to perennial herbs [rarely shrubs to small trees]; stems of two kinds, herbaceous and leafy flowering stems with + swollen nodes, and thickened, condensed, often potentially perennial, basal stems or rhizomes underground or at ground level; roots branched or unbranched taproots or numerous adventitious roots originating from the condensed stems or rhizomes. Leaves mostly circular in outline, palmately lobed [infrequently dissected, entire or pinnate], the lobes variously toothed or incised; basal leaves long-petiolate, the petioles often + persistent; cauline leaves op- posite or alternate (especially in the inflorescence), + similar to basal leaves but usually smaller and with shorter petioles. Inflorescences basi- cally cymose, the ultimate peduncles [1]2-flowered and [2]4-bracteate, usually terminal but by continued and sympodial branching within the inflorescences often appearing axillary. Flowers perfect [rarely imperfect], regular. Sepals 5, with obtuse to caudate apices. Petals 5, pubescent at least basally, frequently emarginate at apex, white to some shade of pink or red-purple [blue to blackish-purple], the veins more intensely colored, reticulately veined throughout or only apically. Stamens 10, in two 190 JOURNAL OF THE ARNOLD ARBORETUM [voL, 53 Ficure 1. Geranium. a~j, G. maculatum: a, flowering plant showing erect annual shoot and perennial underground rhizome, « 44; b, flower at anthesis — note inner whorl! of dehiscing stamens and still closed stigmas, X 14%; c, stamen y; : : —n, G. carolinianum: k, tip of flowering and fruiting bra ing one sympodially branched inflorescence, X ¥2; 1, flower with receptive stigmas, anthers mostly fallen, x ae curved) — note persistent filaments and connivent stigmas, X 2 1972] ROBERTSON, GENERA OF GERANIACEAE 191 whorls, the outer whorl opposite the petals and maturing after the inner antesepalous whorl; filaments often of two lengths, basally connate, all bearing anthers (except in G. pusillum); nectar glands 5, alternating with petals, occuring on receptacular area between sepals and the inner whorl of stamens. Ovary frequently pubescent. Schizocarp elastically or explosively septicidal from base upwards, awns usually remaining attached to top of columella; mericarps rounded at base, each with a persistent awn, these glabrous or puberulent on the adaxial surface, recurving abaxi- ally. Seeds mostly ejected from mericarps, the seed coat reticulate to smooth. Cotyledons long petiolate, transversely elliptic and somewhat asymmetrical when expanded, emarginate at apex and cordate at base, entire or slightly 3-lobed. (Including Robertiella Hanks in Hanks & Small.) Lecrorypr species: Geranium sylvaticum L.; see L. T. Hanks & J. K. Small, N. Am. Fl. 25: 4. 1907. (Name from Greek geranion, a crane, in allusion to the beaklike fruit; used by Dioscorides, see Gunther.) — CRANE’S-BILL, WILD GERANIUM. A genus of 250-300 species widely distributed in temperate, montane, and subpolar regions of both hemispheres, extending into the tropics only in the mountains. Some species have amphiatlantic or subantarctic dis- tributions. Major centers of diversification are the mountains of Europe, the Mediterranean region, the Cordilleras of South, Central, and western North America, and Hawaii. About 20 species are native and 12 natural- ized in the continental United States; six of these (including four intro- duced) occur in our area. The genus was divided by Knuth into 30 often unnatural sections based on longevity, habit, floral size, distribution, and the morphology of the rhizome, leaf, and inflorescence. Subsequent authors have rearranged the of floras), or have accepted provisionally Knuth’s sectional names, while admitting the inadequacy of his system (Carolin). The only perennial species of Geranium in our area, G. maculatum L., is of common occurrence in diverse plant communities, especially of the eastern deciduous forests, but also in fields, meadows, forest openings, thickets, along stream banks, and along the edges of cleared land, from Maine and southern Ontario to southeastern Manitoba, northeastern South Dakota and north-central Nebraska, south to central Georgia and eastern Oklahoma. The occasional white-flowered individuals are sometimes sep- arated from the typical rose-purple-flowered plants as forma albiflorum (Raf.) House. Placed by Knuth in his sect. Sylvatica (= sect. Ge- RANIUM), G. maculatum with simple or branched horizontal rhizomes, was closely allied by Jones & Jones to three other species with unbranched, more or less vertical or oblique rhizomes: G. pratense L., 2n = 28 (Eurasia, occasionally adventive in the northeastern United States), G. erianthum DC., 2m = 28, 30 (northern Japan to Alaska and British Columbia), and G. oreganum T. J. Howell (northwestern United States). This group is 192 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 characterized by large flowers, only basally pubescent petals, short stig- mas, and erect fruiting pedicels (sometimes recurved in G. pratense). The six remaining species in our area have been placed in sect. CoLuM- mature fruit). Carolin, in his treatment of the Malesian species, states that species grouped around Geranium carolinianum in North America have some links with “a complex group — containing Knuth’s sections Chilensia, Australiensia, and fragments of Striata and Columbinum — extending from India, through Malesia, temperate saasie so New Zealand, and the Subantarctic islands: to soiithiern America Geranium carolinianum L. (including G. Langloisii Greeine), Qn =» $2; is indigenous throughout temperate North America, mostly in dry dis- turbed habitats. Distinguished from other species in the area by its awned sepals, pedicels less than twice as long as the sepals, and mature carpels hirsute with long antrorse trichomes, plants of this species are quite vari- able, particularly in density of pubescence and structure of the inflores- cence. Plants with the upper internodes greatly shortened and the flowers in compact, many-flowered, determinate umbels are sometimes separated as var. confertiflorum Fernald, confined mostly to the northeastern United States but occurring also in the mountains of North Carolina. Until re- cently known only from south and central Texas, G. texanum (Trelease) Heller, with circular-ovate sepals and an eglandular rostrum, has recently been collected in Louisiana by J. W. Thieret. Geranium Bicknellii (in- cluding G. longipes Goodding non DC., G. nemorale Suksdorf; trans- Canada and northeastern United States), and G. sphaerospermum Fernald (questionably distinct from G. carolinianum; northern states and adjacent Canada) are closely related to G. carolinianum. Possibly native in the western United States, but more likely an early introduction from Europe, Geranium dissectum L. (including G. laxum Hanks in Hanks & Small), 2m = 22, is locally naturalized and abundant in disturbed habitats in the eastern United States from Massachusetts to Michigan south to Georgia and Texas. It is distinguished from the very similar G. carolinianum by the pubescent but not long hirsute mericarps and more strongly reticulate seeds. Three introduced European species of this section are found locally in disturbed habitats in the eastern United States. Geranium columbinum L., 2n = 18 (indumentum eglandular, pedicels twice as long as the awned sepals, mericarps glabrous) occurs from New York west to the Dakotas and south to North Carolina and Georgia; it is also reported from one county in Washington and along the Columbia River in Oregon. Wide- spread in the Pacific Northwest, G. pusillum L., 2n = 26 (sepals acute, several filaments without anthers, fruit rostrum very short, mericarps pubescent) is infrequent eastward to Massachusetts and south to North Contiaas Oklahoma, and Utah. Geranium molle L., 2n = 26 (sepals acute, all stamens fertile, fruit rostrum short, mericarps ‘obliquely wrinkled — or smooth in forma Preuscho fii Abromeit), ranges from Newfoundland 1972] ROBERTSON, GENERA OF GERANIACEAE 193 to British Columbia and the Pacific States and south to North Carolina, Tennessee, and Illinois. Geranium Robertianum L., Herb Robert, 2m = 64 (32, 54?, 56), of sect. ROBERTIANA Boiss., the only species of the genus in North America with completely palmately compound leaves with the middle segment prominently petiolulate, occurs to the north of our area, extending as far south and west as Virginia, West Virginia, Ohio, Indiana, Illinois, and Nebraska. Amphiatlantic in distribution, this species is common through- out Europe, and it has been introduced into many regions of the world, including the Pacific United States. The prevalent somatic chromosome number, especially in the perennial species, is 28. There are evidently two chromosomal races of Geranium phaeum L., 2n = 14 and 28. A polyploid series based on 14 exists, with reports of 2m = 42 (2 species), 56 (numerous species), and 84 (G. san- guineum L.). Other base numbers are found, particularly in the annual species, the most frequent being 2” = 26 and 52; aneuploidy and poly- ploidy are fairly common. Geranium Robertianum has been reported as 2n = 32, 56 (54?), 64, with the last the most frequent; further studies are needed to determine the scope of chromosomal variation in this species. The closely related G. purpureum Vill. is consistently 2n = 32. Geranium anemonifolium L’Hérit. ex Aiton, a species from Madeira and the Canary Islands allied to G. Robertianum, evidently has two races, 2n = 68 and 128 (the highest number known in the genus). Hybridization between Geranium species is infrequent, occurring mostly between closely related species. Documented hybrids involving species indigenous to North America are thus far unknown, but the species in the western United States have not been closely examined in this respect. Naturally occurring hybrids have been found between two European spe- cies of sect. CoLUMBINA with different base chromosome numbers, G. bohemicum L., 2n = 28, and G. lanuginosum Lamarck, 2n = 48, the hybrid chromosome number being 2” = 38. The seeds of these two species do not readily germinate without exposure to heat and plants often appear in areas after forest fires. The pollen of Geranium species is 3-colporate with round to lolongate ora and short colpi. The exine is semitectate with the infrategillar bacula circular in cross section. More or less uniform suprategillar structures are found in the majority of species, with strongly dimorphic processes occur- ring in many circum-Pacific species, including G. Richardsonii of the western United States, and G. Hernandezii of Mexico. The pollen of a few species lacks suprategillar processes and has a striate-reticulate exine like most species of Erodium. Numerous large-flowered species of Geranium, such as G. maculatum, G. palustre, and G. pratense, are strongly proterandrous and mostly cross- pollinated. As the flowers open, the outer whorl of stamens curves out and downward while the inner whorl remains erect, the anthers soon dehiscing (see FicurE 1b). The outer whorl then becomes erect, the anthers dehisce, and the filaments again recurve, as do those of the inner 194 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 whorl. After this the anthers frequently become detached from the fila- ments, and the lobes of the style open, exposing the stigmatic surfaces. In the smaller-flowered species, such as G. carolinianum, that attract few insects, there is a tendency for homogamy with increased probability for self-pollination. Some species of Geranium and Erodium have both per- fect flowers and smaller and functionally carpellate ones that are usually gynodioecious in distribution. (Gynomonoecism, androdioecism, andro- monoecism, and true dioecism are rare in the Geraniaceae. ) There have been several detailed studies of intraspecific variation in European Geranium species. Geranium Robertianum and the closely re- lated G. purpureum (sometimes treated as a subspecies of the former) were studied genecologically by Boécher and by Baker. Bocher & Lewis investigated variations in leaf and petal morphology, height, habit, time and length of the flowering period, longevity, pubescence, and flower color in 36 strains of G. sanguineum from nine European countries. The poly- morphic breeding system of G. sylvaticum has been investigated numerous times, most recently by Putrament (1962) and Vaarama & Jaaskeldinen (1967). Except for Martin’s study of G. maculatum and Russell’s brief account of G. Richardsonii, detailed studies of variability in North Ameri- can species are lacking but would probably be quite informative. Most species of Geranium, including all in North America, have two types of stems. The above-ground floriferous and annual stems are the more prominent. This kind of stem and the basal leaves arise from zome, oblique, or vertical. In many species this basal stem is the porenilatitig organ, but some form of basal stem is present even in annual species. A taproot and/or numerous adventitious roots arise from the basal stem. In some species the schizocarp dehisces explosively with such force that the seeds are thrown out of the mericarps and travel laterally for several hundred centimeters. Observations made by C. E. Wood, Jr. (personal communication), on Geranium maculatum indicate that when undisturbed the fruit opens over a short period with the mericarps separating from the columella and the adaxial side, where the dehiscence suture is located, becoming oriented away from the axis due to the revoluting awns. At this point the seeds are forcibly ejected from the mericarps due to the drying and subsequent contraction of the carpel wall. The awns become even more revolute after dehiscence (see FicurE 1g). Martin, in her study of this species found that the fruits would dehisce explosively if subjected to a slight mechanical stimulation; this violent dehiscence evi- dently occurs spontaneously in Pig aks with the awns and/or mericarps becoming detached from the fru ome species of Geranium are ay ae grown as ornamentals. The most frequently encountered species in eastern North America are G. sanguineum L., 2n = 84, which has several varieties and named culti- vars, and G. grandiflorum Edgew. 1972] ROBERTSON, GENERA OF GERANIACEAE 195 REFERENCES: Under family references see BOLKHOVSKIKH et al., BORTENSCHLAGER, BROUWER ; . Davis, Dawson, Gams, Hanks & SMALL, HEss et al., KenpA, P. Knutu, R. Kwura (1912, 1931), Komarov, KUGLER, LEGAULT, LusBBock, Mir tManorr, MosespacH, H. MULteEr, Murty, RICKETT, SAUER. SCHWIEKER, S@RENSEN, STAFFORD, ge es Wes, and ZANKER Baker, H. G. Geranium purpureum Vill. and G. Robertianum L. in the British flora. I. Geranium purpureum. Watsonia 3: 160-167. 1955; II. Geranium Robertianum. Ibid. 270-279. 1956 Genecological studies in Geranium (section Robertiana). General con- sideration and the races of G. purpureum Vill. New Phytol. 56: 172-192. pls..2, 3. SUF Barton, W. P. C. Vegetable materia medica of the United States; or medical botany. Vol. 1. 273 pp. pls. 1-24. Philadelphia. 1817. [G. maculatum, 149-159, pl. 13.] Bastin, E. S. Structure of Geranium maculatum. Am. Jour. Pharm. 66: 516— 522. 1894. [Includes anatomical details of rhizomes, roots, starch grains. ] BaTE-SmitH, E. C., & P. RrsEREAU-Gayon. Leuco-anthocyanins in seeds. Qual. Pl. Mater. Veg. 5: 189-198. 1959. [G. ibericum, G. pratense, G. sylvaticum. ] BLARINGHEM, L. Hérédité en mosaique du Geranium pratense L. var. album striatum. Compt. Rend. Acad. Sci. Paris 198: 2044-2047. 1934. Bocuer, T. W. Cytogenetic and biological studies in Geranium Robertianum L. Danske Vidensk. Selsk. Biol. Medd. 20(8): 1-29. 1947. [Includes i a tion and intraspecific hybridization studies. | ———— & K. Larsen. Experimental and cytological studies on plant species. IV. Further studies in short-lived herbs. Danske Vidensk. Selsk. Biol. Skrifter 10(2): 1-24. 1958. [G. molle, 19, 20.] . C. Lewis. Experimental and cytological studies on plant species. VII. Geranium inn Ibid. 11(5): 1-25. pls. 1, 2. 1962. [Variation patterns for 7 character. Caro.in, R. C. The genus soho L. in the southwestern Pacific area. Proc Vien Soc. New S. Wales 89: 326-361. pls. 6, 7. 1964. [Morphological notes, taxonomic treatment, plates of seeds and pedicel indumentum. | Ciements, F. E., & F. L. Lonc. Experimental pollination. An niin of the ecology of flowers and insects. vii + 274 pp. pls. 1-17. Washington, D.C. 1923. [G. caespitosum, 39-45, pl. 7; Geranium, 105-110; list of pollinators, DAHLGREN, K. V. O. Chromosomes in Geranium bohemicum and G. lanuginosum with some notes on the hybrid. Hereditas 38: 314-320. 1952. [Seeds of these species do not germinate without exposure to heat. ] Dorocanevskaya, E. A. On the vicarious species of Geranium L. in the basin of the Chon-Kyzyl-Su River. (In Russian; English summary.) Bot. Zhur. 47: 1292-1300. 1962. [Effect of ecological conditions at various altitudes on chemical composition of G. collinum and G. saxatile. Duncan, W. H. Preliminary reports on the flora of Georgia. 4. Notes on the distribution of flowering plants including species new to the state. Castanea 15: 145-159. 1950. [G. columbinum, 154, possibly native. ] Eastwoop, A. The fertilization of Geraniums. Zoé 2: 112. 1891. [G. Fremontii, G. Richardsonii, G. caespitosum, descriptions of pollination. ] 196 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FERNALD, M. L. Geranium carolinianum and allies of northeastern North America. Rhodora 37: 295-301. pls. 371-374. 1935. [9 spp. recognize Fosperc, F. R. The Hawaiian Geraniums. Occas. Pap. Bishop Mus, 12(16): 1-19. 1936. [4 spp. of the anomalous sect. NEUROPHYLLODES Gittort, P., & A. M. Mortsor. Sur la composition glucidique du ‘geranium des prés (Geranium pratense L.). Assoc. Frang. Avanc. Sci. Compt. Rend. 55: 415-418. 1931. Gstirner, F., & K. H. Lisken. Uber Inhaltsstoffe des Rhizoms von Geranium pratense L. Arch. Pharm. 295: 823-833. 1962. [Eight crystalline substances isolated from solutions, and several other compounds identified by chroma- tography. | GunTHER, R. T. The Greek herbal of Dioscorides. xii + 701 pp. Frontisp. Oxford. 1934. [Geranion, 361, 362. Hara, H. Contributions to the study of variations in the Japanese plants closely related to those of Europe or N. America. I. Jour. Fac. Sci. Univ. Tokyo Bot. 6: 29-96. 1952. [G. Porte Saran 81.] Jackson, W. What is Geranium anemonefolium? Heredity 6: 279. 1952. JoHNSON, W. C. Trillium cernuum L. and Geranium maculatum L.: New for South Dakota. Rhodora 72: 554. 1970 [1971]. [Sisseton Co. Jones, G. N., & F. F. Jones. A revision of the perennial species of Geranium of the United States and Canada. Rhodora 45: 5-26; 32-53. 1943. [18 spp.] Knutu, R. Uber die geographische Verbreitung und die Anpassungserscheinun- gen der Gattung Geranium im Verhiltnis zu ihrer systematischen Gliederung. Bot. Jahrb, 32: 190-230. 1903. LerrertovA, I. Zur Chemotaxonomie der Gattung Geranium — freie Saccharide nd Aminosaiiren. Preslia 40: 357-361. 1968. & H. BuéKovA. Zur Chemotaxonomie der Gerbstoff-Merkmale bei der Gattung Geranium. Der Gerbstoffgehalt bei der G. sanguineum III. (In Czechoslovakian; German abstr.) Preslia 40: 60-63. 1968. Exsenerc cHovA. Zur Chemotaxonomie der Gattung Geranium — die Stetigkeit der eaidevetic bei der Art Geranium pyrenaicum L. (In Czechoslovakian; German abstr.) Preslia 42: 267-269. 1970. ; &L. Natienovi. Fat Chemotaxonomie der Gerbstoff-Merkmale in der Gattung pretest t (In Czechoslovakian; German abstr.) Preslia 37: 413-418. 1965. [9 sp Lewis, M. C. Sececaintial differentiation and adaptive significance of leaf structure in relation to xeric environments. XI. Int. Bot. Congr. Abstr. p. 128. 1969. [Energy budgets determined for G. sanguineum leaves of different degrees of dissection. | LunpMAN, B. Some notes on the regional variation of flower colour by Geranium silvaticum L. (In Swedish; English summary.) Sv. Bot. Tidskr. 42: 153- 157. 1948 MacBrype, a Geraniaceae. Jn: R. E. Woopson, Jr., & R. W. Scuery. Flora of Panama. Ann. Missouri Bot. Gard. 54: 201-205. 1967. [Geranium only. | Mik M. C. An ecological life history of Geranium maculatum. Am. Midl. Nat. 73: 111-149. 1965. [Morphology, geographical variation. | MitispaucH, C. F. Geranium maculatum. American medicinal plants 1(32): [2 pp.]. pl. 32. 1887. Moorg, H. E., Jr. A revision of the genus Geranium in Mexico and Central 1972] ROBERTSON, GENERA OF GERANIACEAE 197 America. Contr. Gray Herb. 146: 1-108. pls. 1-5. 1943. [43 spp.; for corrections cf. Gent. Herb. 8: 250-257. 1951.] Munesapa, T., N. Murakami, M. Nisuimoto, & K. Matsumoto. Anatomical characters of a Geranium herb, G. carolinionum L.. (In Japanese ; English summary.) Jour. Jap. Bot. 35: 187-190. 1960. [Naturalized in Japan. ] NacL, W. Uber Endopolyploidie, Restitutionskernbildung und Kernstrukturen im Suspensor von Angiospermen und einer Gymnosperme. Osterr. Bot. Zeitschr. 109: 431-494. 1967. [G. phaeum, 448-452. Necer, F. W. Uber Blatter mit der Funktion von Stiitzorganen. Flora 92: 371- 379. 1903. [Includes G. Robertianum. | Necopr, G. Su di un modo di rappresentazione grafica degli sviluppi quantitativi di organi fiorali: antogrammi quantitativi. Appunti sull’architetture fiorale di Geranium nepalense Sweet. Atti Soc. Nat. Mat. Modena 65: 81-87. 1934. PoHLHEIM, F., & H. J. Tmticu. Verbanderung (Fasziation) bei Geranium pratense L. und ihre title anagem Flora B. 158: 580-593. 1969. [Single recessive-gene fasciation of the PUTRAMENT, A. Some Speen chow on oor sterility in Geranium sylvaticum L. var. alpestre Schur. Acta Soc. Bot. Polon. 31: 723-736. 1962. [Cytological observations and breeding experiments Rarnio, A. J. Uber die Intersexualitit bei Geranium pratense L. und ihre Entstehung durch Bastardierung. I. Ann. Bot. Soc. Zool.-Bot. Fenn. 8(4): 1-28. 1 pl. 1937. Ropertson, C. Flowers and insects. III. Bot. Gaz. 14: 297-304. 1889. [G. maculatum, 299, 300; flower agrees with G. palustre & G. pratense; list of pollinators. | Hepa be Ibid, 18: 267-274. 1893. [G. carolinianum, 272; flowers imperfectly proterandrous; list of pollinators. ] RusseL_, N. H. Quantitative studies in angiosperm taxonomy. X. Valeriana. XI. Geranium. XII. Mimulus. Castanea 29: 138-150. 1964. [G. Richard- sonii, 143-146; 26 characters show plant to plant variation. | SANSOME, F. W. Some experiments with Geranium species. Jour. Genet. 33 359-363. 1936. [G. Endressii, G. striatum; G. sanguineum, G. sanguineum var. Ati G. lancastriense; inheritance in hybrid crosses. SHAW, R. J. A cytotaxonomic study of the genus Geranium in the Wasatch region of Idaho and Utah. Madrofo 11: 297-304. pl. 6. 1952. [G. carolin- ianum, G. pusillum, G. nervosum. G. Richardsonii. | SOUEGEs, R. Em peyouenls des Géraniacées. Développement de l’embryon chez le Geranium molle L. Compt. Rend. Acad. Sci. Paris 17: 556-558. 1923. populations of Geranium sylvaticum L. Ann. Acad. Sci. Fenn. A. IV. Biol. 108: 1-40. 1967. [Also history of the study of floral biology in the genus. ] bina oa O. D. On the morphology and nomenclature of Geranium sanguineum (In Ukrainian; English summary.) Jour. Inst. Bot. Acad. Sci. RSS Ukr. re 137-140. 1935.* 2. Erodium L’Héritier ex Aiton, Hortus Kew. 2: 414. 1789. Annual or biennial herbs [sometimes perennial, caespitose or shrubby], at first acaulescent rosette plants, later with several + jointed elongate, simple or branched stems often arising and radiating from generally un- branched taproots. Leaves longer than broad, pinnately compound [or lobed; rarely simple or palmate], the leaflets toothed, lobed or subdivided, 198 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 sessile or shortly petiolulate; basal and cauline leaves similar, the latter opposite [or all alternate] and tending to be more strongly lobed, divided, or subdivided. Inflorescences determinate umbels, arising from the rosette and/or from nodes of elongated stems; ultimate peduncles several flowered [infrequently with 1 or 2 flowers]. Flowers perfect [rarely imperfect, the plants dioecious], regular to slightly irregular. Sepals 5, often mucronate or aristate. Petals 5, glabrous, equal or the two upper smaller [rarely absent], pink to violet, purple [white, blue or lilac, very rarely yellow or deep red], the veins typically more intensely colored, sometimes with a dark or blackish spot near the base of the two upper [or all] petals. Fertile stamens 5, the outer antepetalous whorl reduced to staminodia [or scales]; filament bases sometimes with two lateral teeth. Ovary pubescent. Schizocarp dehiscing mostly from apex downwards, the meri- carps and awns becoming completely detached from the columella; meri- carps tapering toward base; awns persistent on mericarps [or deciduous], pubescent on adaxial surface, lower portion spirally coiling when free from columella. Seed smooth, retained within mericarp. LECTOTYPE species: E. crassifolium L’Hérit. ex Aiton = E. hirtum (Forsk.) Willd. non Jacq.; see L. T. Hanxs & J. K. Smatz, N. Am. Fl. 25: 21. 1907. (Name from Greek erodios, heron, in comparison of the shape of the fruit to a heron’s bill.) — HERON’S-BILL, STORK’S-BILL, ALFILERIA, FILAREE, PIN-CLOVER. A genus of 60 to 80 species centered in Mediterranean Europe, North Africa, and the Middle East; species have been introduced into most temperate regions of the world. Two sections are usually recognized, Eropium (§ Plumosa Boissier) (mericarp awns quickly deciduous and plumose on the adaxial surfaces with long silvery or yellowish trichomes) and Barsapa Boissier (awns persistent on mericarps, trichomes on adaxial surfaces short, adpressed, and reddish). The latter section was subdivided by Knuth into ten subsections based on leaf morphology, habit, rostrum length, petal size and shape, and density of indumentum. In the United States at least ten species, all belonging to sect, BARBATA, have been reported, several being ephemeral escapes. Two introduced species, Erodium cicutarium (L.) L’Hérit. ex Aiton and E. moschatum (L.) L’Hérit. ex Aiton, both belonging to subsect. CrcuTaria, a taxonomi- cally difficult group centered in the southwestern Mediterranean region, occur in our area. Both are also naturalized in the West, as are E. Botrys Bertol., 27 = 40 (southern Europe and the Mediterranean region) and E. brachycarpum (Godr.) Thell. (including E. obtusiplicatum (Maire, Weiller & Wilczek) J. T. Howell and E. Botrys f. montanum Brumb.; North Africa), 2” = 40, Apparently the western E. macrophyllum Hooker & Arnott (EZ. californicum Greene; related to E. malacoides (L.) L’Hérit. ex Aiton of southern Europe) and E. texanum Gray, 2n = 20, placed by Knuth in subsect. GuttTaTa, otherwise Mediterranean in distribution, are indigenous. 1972] ROBERTSON, GENERA OF GERANIACEAE 199 Erodium cicutarium, characterized by its eglandular apical mericarp pits, aristate sepals, sessile leaflets divided more than half-way to the midrib, and toothless anther-bearing filaments, is found in fields, road- sides, and waste places throughout most of the United States. It probably occurs sporadically in all states in our range but is poorly represented in herbaria. A very plastic species, particularly in habit, size, leaf shape, indumentum, and fruit morphology, E. cicutarium has had a tortuous nomenclatural history with numerous segregate species, subspecies, and varieties described and several classification schemes proposed (see Litar- diére in Briquet, Andreas, Rottgardt, Larsen, Tutin e¢ al., and Guitton- neau). In Europe there are several chromosomal races with the predomi- nant number being 2” = 40, but 2m = 20 is common and 2n = 36, 48 and 54(56?) also occur; hexaploid populations, 2” = 60, have been segre- gated as E. danicum K. Larsen. At least three morphological entities occur in Europe: a mostly tetraploid “field” type, a tetraploid “dune” type, and a mostly diploid ‘‘dune” type. The plants in the United States evidently correspond to the “field” type, E. cicutarium subsp. cicutarium, and the only chromosome number reported from this country is 2n = 40. This species is also naturalized in Australia, Java, East Africa, and prob- ably in most other temperate regions of the world. Erodium moschatum, 2n = 20, characterized by its glandular apical mericarp pits, acute to mucronate sepals, shortly petiolate leaflets divided less than half-way to the midrib, and anther-bearing filaments with two lateral teeth, is common in California and has been reported as an infre- quent adventive in waste places and roadsides in northern Florida and the northeastern United States. It is also naturalized in East and South Africa and Australia. The awn of the mericarp is pubescent inside and has a dextrorsely spiraled coiled portion, the spirillum, and an arching terminal portion, the flagellum. The awns are hygroscopic, the spirals loosening when moist and becoming tighter when dried. In nature, varying weather conditions cause the awns to coil and uncoil repeatedly, and this action evidently forces the seeds into the soil. The awns have been used to make hygrom- eters. The seed is usually retained within the tapering mericarps. At the apex of the mericarp are two flattened areas referred to as pits (foveae, foveoles) which are subdivided in some species by one or more ridges, giving the impression of furrows below the pits. The base chromosome number in the genus is x = 10, with 2 = 20 the most frequently reported number. Tetraploids are common, hexaploids have been recorded in three subsections, and one octoploid is known. Aneuploidy occurs in E. cicutarium (see above) and all reports for species in subsect. ABSINTHIOIDEA are 2n = 18. Most, if not all, species in this latter subsection are dioecious (see Davis & Roberts; illustration in Stefani, et al.). The pollen of Erodium species is 3-colporate and most species lack Suprategillar processes with the exine striate-reticulate. Three subtypes based on sculpturing patterns were distinguished by Bortenschlager. 200 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 few species, including E. texanum, have strongly dimorphic suprategillar processes, like the pollen found in most Geranium species. In the Southwest several species, particularly E. cicutarium, are im- portant spring and winter range forage plants for domestic and game ani- mals. Some species are also grown as ornamentals in rock gardens. Ero- dium cicutarium can be an aggressive weed, competing vigorously with crops such as alfalfa. REFERENCES: Under family references see BOLKHOVSKIKH et al., BORTENSCHLAGER, BROUWER & STAHLIN, P. H. Davis, Dawson, Gams, Hanxs & SMALL, HEss et al., KENDA, H L’HeEritier, Lupzock, MirtmManorr, Mosezacu, H. MULLER, Mur-ey, RICKETT, SAUER, SCHWIEKER, SORENSEN, STAFFORD, WANGERIN, WEBB, and ZANKER. Anpreas, C. H. De inheemsche Erodia van Nederland. (In Dutch; English summary.) Nederl. Kruidk. Arch. 54: 138-229. pl. 1. 1947, [Field, morphological, cytological and hybridization study with emphasis on E. cicutarium and its relatives. | BrucceMan, C. J. M. Overzicht der in Nederland gevonden inheemse en adventieve Erodium-soorten. (In Dutch; English summary.) Gorteria 2: 49-55. 1964. [10 spp.; illustrations. ] BRUMHARD, P. Monographische Ubersicht der Gattung Erodium. 59 pp. Inaug. Diss. Breslau. 1905. (Also see Repert. Sp. Nov. 2: 116-119. 1906.) Carotin, R. C. Notes on the genus Erodium L’Hér. in Australia. Proc. Linn. Soc. New S. Wales 88: 313-319. 1964. [Introduced spp.; for endemic spp. see Ibid. 83: 92-100. 1958. Davis, P. H., & J. Roserts. Materials for a flora of Turkey I. Geraniaceae. Notes Bot. Gard. Edinburgh 22: 9-27, 10955. [Includes comments on IDEA, Dayton, W. A. Alfileria (filaree) “seed.” Rhodora 39: 233-235. 1937. [In- ] E1jx, J. L. van. Phytochemical investigation of Erodium cicutarium. (In N b GUITTONNEAU, G. Contribution a l’étude caryologique [caryosystématique]| du genre Erodium L’Hér. I. Bull. Soc. Bot. France 111: 1-4. 1964; II. Ibid, 112: 25-32. 1 pl. 1965; III. Zbid. 113: 3-11. 1966; IV. Ibid. 114: 32-41. 1967. Sur la répartition de l’Erodium brachycarpum (Godr.) Thell. [= E. obtusiplicatum (Maire, Weiller et Wilczek) Howell] dans le bassin méditer- ranéen. Jbid. 113: 66-68. 1966. . Note sur le texte de la Geraniologia de L’Héritier. Jbid. 114: 42-45. 1967. ———.. Contribution a l’étude biosystématique du genre Erodium L’Hérit. dan le bassin méditerranéen occidental, Thése Fac. Sci. Orsay Univ. Paris. 253 pp. 23 pls. in 2 vols. Orsay. 1970. HANSTEIN, J. Ueber die Einbohrung der Geraniaceen-Friichte in den Boden. Bot. Zeit. 27: 528, 529. 1869. [Seed dispersal mechanism in E. gruinum ; for partial English transl. see I. Sprague & G. L. Goodale, Wild Flowers of America, Vol. 1: 17, 18. 1882.] 1972] ROBERTSON, GENERA OF GERANIACEAE 201 Harpin, J. W. Some notes on weeds of North Carolina. Castanea 24: 22, 23. 1959. [E. cicutarium has completely destroyed alfalfa crops in the Pied- mont. | Irvinc, W. The heronsbill family. Garden London 87: 329-331. 1923. [15 perennial spp. listed for rock garden cultivation. JARRETT, M. P. Arizona’s Wildflowers. Arizona Highways 47(8): 10-39. 1971. [E. cicutarium, E. texanum, 31; colored photographs, 19. Larsen, K. Cytological and experimental studies on the genus gg Ploeg special references to the collective species E. cicutarium (L.) L Danske Vidensk. Selsk. Biol. Medd. 23(6): 1-25. 1958. LITARDIERE, R. pE. Erodium. In: J. Briquet, Prodr. Fl. Corse 2(2): 15-36. 35 Martin, P. S., & C. M. Drew. Scanning electron photomicrographs of south- western pollen grains. Jour. Arizona Acad. Sci. 5: 146-176. 1969. [E. p. 164. Preston, C. E. Two instructive seedlings. Bot. Gaz. 33: 150-154. 1902. [E. cicutarium, phylogeny of trichomes. | RAMALEY, F. The trichome structures of Erodium cicutarium. Bot. Gaz. 34: 140-142. 1902. [Disputes Preston’s findings. | Rottcarpt, K. Morphologische, cytologische and physiologische Untersuchun- i von Okotypen in Schleswig-Holstein. Beitr. Biol. Pflanzen 32: 225-278. 1956. [E. cicutarium vars. immaculatum and pimpinellifolium, 228-243.] SHAw, W. R. Hygrometer made with Erodium awns. Bot. Gaz. 24: 372. 1897. StEcERT, A. Erodium corsicum Léman, ein weiteres Beispiel fiir Blattstieldornen, zugleich ein Beitrag zur Kenntnis der Blattstieldornen tberhaupt. Beitr. Biol. Pflanzen 46: 461-480. 1970. [Thorns are modified petioles. | SovEces, R. Embryogénie des Géraniacées. Développement de l’embryon chez VErodium cicutarium L’Hérit. Compt. Rend. Acad. Sci. Paris 176: 1565- 23. STEFANI, C. pe, C. J. F. Mayor, & W. BarBey. Samos. 99 pp. pls. 1-14. Lausanne. 1891. [Pl. 2 strtaates staminate and pistillate flowers of £. Vatteri. VierHappeER, F. Beitrige zur Kenntnis der Flora Griechenlands 2. Verh. Zool.- Bot. Ges. Wien 69: 102-156. 1919. [Erodium, 112-155. Wacnon, K. A., & H. H. Biswett. Two types of broad-leaf oes: in California. Madrofio 7: 118-125. 1943. [E. botrys and E. bot montanum. | ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 202 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ON THE ORIGIN OF THE ORCHIDACEAE, II? LESLIE A. GARAY The myriad things in nature are completely similar and completely dissimilar. This should be described as a great similarity in dissimilarity. — Hur Sura (4th Century B.C.) In 1960, in discussing my ideas on the origin of the present-day com- plexity of the orchid family, I called attention to the presence of five distinct phyletic lines based on the disparity in their respective endo- morphic and exomorphic features, as well as on the absolute genetic incom- patibility among the members of each line with respect to the others. These kladogenetic lines which have evolved in a parallel manner represent distinct subfamilies known as APOSTASIOIDEAE, CYPRIPEDIOIDEAE, ORCHI- DOIDEAE and NEoTTIOIDEAE. The plants referable to these subfamilies are all terrestrial. The epiphytic mode of life evolved through a secondary differentiation which occurred in the Neorriomear, giving rise to a fifth line or subfamily, the EPIDENDROIDEAE (4, 5). _ This new departure from the commonly accepted classification into two main divisions —MoNANDRAE and DIANDRAE, one versus two fertile anthers — has stimulated responses for the continuation of the old status quo. The fundamental difference between my thesis and those of my critics is the recognition of the evolutionary lines as natural systematic units, whereas in the opposing approaches, where the phyletic lines are randomly accommodated in some arbitrary categories, the taste and flavor of the ancient numerical system of Linnaeus is still cloyingly prevalent. A sum- mary of the main approaches is given in FicureE 1. CRITIQUES AND CRITICAL POINTS Apostasioideae. An attempt has been made by Vermeulen (13, 14) along the lines which were suggested in the past by Schlechter, Ridley, Takhtajan (12), and more recently by Hutchinson (8) to reinstate the APOSTASIOIDEAE to family status. Yet Stebbins and Khush (11) have convincingly shown that the stomatal complex in the APOSTASIOIDEAE is anomocytic, i-e., two guard cells without subsidiary cells, which is char- acteristic of the Orchidales. The Haemodorales or Liliales to which Apostasiaceae has been referred by several is characterized by a paracytic stomatal complex, i.e., two guard cells plus two subsidiary cells. This latter group seems to have arisen only once in evolutionary history of the mo- nocotyledons. *On the origin of the Orchidaceae [I]. Bot. Mus. Leafl. Harvard Univ. 19: 57-96. 1960. AUTHOR MAIN DIVISIONS OF SYSTEM Epiderdr ae AC D/ $i aR Pa Bos / LINDLEY Pebeac | OPHRYDEAE aap a MALAXIDEAE | EPIDENDREAE | VANDEAE BENTHAM || CYPRIPEDIEAE | OPHRYDEAE | NEOTTIEAE EPIDENDREAE | VANDEAE OPERCULATAE REICHENBACH || CYPRIPEDIEAE | OPHRYDEAE oe ae ee ee =| NEOTTIEAE EUVOPERCULATAE og ae 3 Bea! es ~ ACROTONAE ITZ PLEONANDRAE E PFITZER BASITONAE cranthae — Acranthae{ Duplicatae\ Sympodiates Convolutae Articulatae Sto re odiale MONANDRAE DIANDRAE $$ $_______| ROLFE OPHRYDEAE NEOTTIEAE EPIDENDREAE rif VANDEAE MONANDRAE SCHLECHTER ee BASITONAE Polychondreae Marosphaerene THRAUOSPHAEREAE ____ KEROSPHAEREAE MANSFELD OPHRYDEAE NEOTTIEAE EPIDENOREAE VANDEAE j : ORCHIOIDEAE HATCH BASITONEAE ACROTONEAE RL Poe a SSS DRESSLER ee oe _ORCHIDOIDEAE © 2 ae i AND CYPRIPEDIOIDEAE | Bee H DODSON ORCHIDEAE NEOTTIEAE EPIDENDREAE LORCHIDOIDEAE | NEOTTIOIDEAE EPIDENDROIDEAE ae CYPRI- L EPIDENDREAE VANDEAE APOSTASH peniol-| ORCH/DEAE | EPIDOGONEAE -————-—___ ________|_ GARAY OIDEAE| DEAE | D/SEAE Sige. pe : DISPERIDEAE | CRANICHIDEAE Dendrobity es peta vinae Aedes re ditine FIGURE 1. Summary of various systems of classification of ORCHIDACEAE. AVAOVGIHOYO AO NIOIMO ‘AVUVD [7261 £02 204 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 APOSTASIA NEUWIEDIA PHYLOGENY — HOLOGENY FicureE 2. Phylogeny and hologeny of subfamily APosTASIOIDEAE, Vv FIGURE CS). In many ways they resemble the remnants of a grand dinner from which it is no longer possible to reconstruct the complete enu. 1972] GARAY, ORIGIN OF ORCHIDACEAE 205 Cypripedioideae. This taxon, like the APosTASIOIDEAE, is a relic group (5, 12). It is composed of four superficially similar genera: Cypri- pedium, Paphiopedilum, Phragmipedium, and Selenipedium. The generic differentiation must have taken place in the obscure evolutionary past, perhaps shortly after the group as a whole was initiated, because of the absolute incompatibility among the members of each line with respect to the other lines. Lack of incompatibility among members of different genera is the rule rather than the exception in the ORCHIDOIDEAE, NEOTTIOIDEAE and EPIDENDROIDEAE (6). The CyprIPEDIOIDEAE has been accepted generally as a distinct sub- family, although its circumscription varies according to the interpreter. Recently Vermeulen, following in the footsteps of Lindley and Mansfeld, but not necessarily following their reasoning, proposed to elevate this group to family status, thus recognizing the Apostasiaceae, Cypripediaceae and Orchidaceae within the order Orchidales (13, 14). Indeed it makes no difference whether or not we recognize five subfamilies or five distinct families so long as the principles that must govern such steps do not become subservient to arbitrary decisions. Such a guide line, however, is wanting in Vermeulen’s presentation. Orchidoideae. In advancing the thesis that the APosTASIOIDEAE and CYPRIPEDIOIDEAE must be regarded as distinct families, Vermeulen failed to recognize or appreciate the uniqueness of the OrcHipompEAE, for he included it together with a newly proposed subfamily (although with an old and already known name, EPIDENDROIDEAE) in his new concept of the orchid family (13, 14). It is not difficult to recognize the reason for such an action, for the subfamily ORCHIDOIDEAE includes the genus Orchis, the nomenclatorial type of the Orchidaceae. Should one grant family status to each of the five subfamilies, the Orchidaceae will become one of the least significant families in the whole complex, with most of its members showing the strain of ancient isolation and overspecialization (6). The overwhelming majority of the species is in the subfamily EPIDENDROIDEAE. Thus, the interpretation of evolutionary phenomena which led to the idea of the fragmentation of the Orchidaceae into several families has suddenly given way to a completely unrelated aspect of scientific endeavor: nomen- clatorial priorities. The subfamily OrcHIpoIpEAE is composed of well-advanced species, so far without any trace or evidence documenting its origin and progressive differentiation from a lesser evolved stock. The highly evolved pollinia with pollen grains solidly united into massulae which are borne on caudicles with hygrometric behavior (4, 6), the complete basitonic fusion between the anther and the column (4), and the occurrence of polystely in the roots and tubers (15) are all strictly limited to this subfamily. The column in the Orcu1poweae is very dissimilar to that found in other subfamilies; the only fertile anther is formed in such a way that it is completely adnate to, and embedded in, the columnar tissue. To derive the column of the NeotTiowear and EpmipENDROIDEAE from the ORCHT- PHYLOGENET/C TRENDS INTHE COLUMN OF ORCAIDOIDEA E Ceratandra Coryctum Ficure 3. Phylogenetic trends in the column of subfamily ORCHIDOIDEAE, Satyrium Pachites Disa Orchis Habenaria § Mirandae Pterygodium 902 WOLAXOIUVY GIONYV AHL AO TYNUNOL “TOA | es 1972] GARAY, ORIGIN OF ORCHIDACEAE 207 DOIDEAE or vice versa, would require the impossible reversal of the true morphological apex of the anther. Although such a reversal may seemingly come about through inclination of the anther, such a directional movement merely changes the position of the true apex without reversing it, as shown in Ficure 3 depicting that phylogenetic trend (6). It is evident that the various stages in the gyration of the column are intimately correlated with the entomophilous nature of the flowers, each step progressively leading toward irreversible overspecialization, and hence to ultimate extinction. Neottioideae. This kladogenetic line is the most successful one, as well as the most important one in the whole family, for it provides all of the evidence necessary for the interpretation of the various evolutionary pathways that make the present-day complexity understandable. It was within the framework of this subfamily that evolutionary developments (or rather evolutionary innovations) leading to a secondary kladogenetic differentiation, EPIDENDROIDEAE, and opening all avenues toward a suc- cessful expansion in both structural diversity and spatial distribution, occurred. The column in the Neott1o1mEaeE and in the EpIDENDROIDEAE is strik- ingly dissimilar to those of the other subfamilies, although all share in common a certain degree of fusion between the style and the filament or filaments. Such a fusion must have occurred independently many times during the early history of the family, but only four kladogenetic lines have been carried through to the present. The single fertile anther is free and movably connected with the column at its acrotonic end. A closer inspection of the hologeny of this subfamily, FicuRE 4 (3), seems to focus on the presence of two major anagenetic trends. One of these is expressed in the aggregation of free pollen grains into tetrads and the tetrads into pollinia, while the other can be traced in the progressive modification of one of the stigmatic lobes into a new organ called the rostellum. Although the interaction of these two trends seems to provide an inexhaustible reservoir for diversity, the true expansion of the orchid family owes its existence to the coaction of a third evolutionary innovation, the development of the epiphytic mode of life. The main anagenetic lines are designated here as tribe Cranichideae, tribe Neottieae, and tribe Epipogoneae. Epidendroideae. The members of this subfamily are all advanced types. In regard to the development of floral parts, the EPIDENDROIDEAE represents the final stage in the organogenesis of the column, because there is a complete fusion of the reproductive organs in every species. Correlated with this fusion is a fully functional rostellum to which the pollinia are attached. There is also a further correlation between the position of the anther and the rostellum and between the position of the rostellum and structural changes in the pollinia (6). An inspection of longitudinal sections of developing flower buds, which are assembled according to their affinities in Ficures 5 and 6, clearly shows that the 208 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 EPIPOGONEAE NEOTTIEAE CRANICHIDEAE PHYLOGENY HOLOGENY ecm Ficure 4. Phylogeny and hologeny of subfamily NEOTTIOIEAE. various floral types fall within two anagenetic lines or natural evolutionary units: 1, column with an erect anther and an ascending (arrect) rostellum to which the pollinia are attached without appendages, and 2, column with an incumbent anther and a horizontally projecting (porrect) rostellum to which the pollinia are affixed through well-developed stipes. These two evolutionary lines are designated here as tribe Epidendreae and tribe Vandeae respectively (6). The columnar structure in the Epidendreae is very similar to that of the Cranichideae of the subfamily Neorriompear. Anagenetic differentia- tion in the NEOTTIOIEAE is expressed in three distinct lines or natural evolutionary units (5). Because of the identity of the angle of the inser- tion of the anther and also of the ascending rostellum in both the Crani- chideae and Epidendreae, I believe that the subfamily EpmIDENDROIDEAE Owes its existence to a secondary kladogenetic differentiation that took CAysis Octomeria Arpophyl lum Ceratostylis Appendicula Dendrobium Laclia Restrepia Stelis Ficure 5. Diagrammatic median longitudinal sections of column in developing buds of members of tribe EPIDENDREAE. Bulbophyl lum AVAOVGIHIOUO AO NIDIYO ‘AVUVD [ZL61 602 Maxillaria Oncidium Cochlioda Renanthera A campe Ficure 6, Diagrammatic median longitudinal sections of column in developing buds of members of tribe VANDEAR. O1Z WOLAXOTUVY GIONUYV AHL AO TYNYAOL ¢¢ “10A] 1972] GARAY, ORIGIN OF ORCHIDACEAE 211 place in the early history and origin of the NEoTTIOIEAE. Embryological evidence favors such a proposition, for the primitive type of embryo with a single-celled suspensor is found only in members of the Cranichideae and in the less evolved species of the Epidendreae. The evolutionary success of the EPIDENDROIDEAE, however, is due primarily to the develop- ment of the epiphytic mode of life, as has already been mentioned (6). EVOLUTIONARY PERSPECTIVES The family Orchidaceae originated in the area known phytogeograph- ically as Malaysia during the Cretaceous period when most angiospermous families became differentiated. At that time all species were geophytes, for the epiphytic mode of life is a rather recent development, dating back to the Pliopleistocene. While the terms geophytic and epiphytic convey the meaning of ecological habitats, in reality they express distinct evolu- tionary adaptations through morphological modifications in the roots. In the terrestrial roots, which may be thin and fibrous or thick and fleshy, the epidermis consists of one to three layers of cells, the outermost layer being provided with root hairs. In the derived epiphytic roots, the layers of epidermal cells are greatly increased in number into a spongy velamen with the outermost layer cutinized and devoid of root hairs. Hence the structure of the root and not the place of abode differentiates these two types of evolutionary adaptations, the latter one of which is also an evolutionary innovation. here is no other plant family comparable in floral diversity to the Orchidaceae. The basic trimerous pattern common to most monocoty- ledons, displays such a vast array of modifications in this family that without serious study one could question the validity of calling them all orchids. Yet, all of the flowers have three sepals and alternating with them three petals, the unpaired petal greatly modified into a labellum or lip, an evolutionary adaptation providing a landing platform for pollinating insects. In the center of the flower, ordinarily occupied by the style and encircled by alternating rows of stamens, all orchids possess a novel structure called the gynostemium or column. The column is also an evolutionary innovation, the result of fusion of the style with the filaments of various stamens. : Concurrently with the development of the column two additional evolu- tionary innovations became operational. The first is the development of pollinia, in which the pollen grains at the time of maturity are shed either as free single grains, or variously aggregated into tetrads or massulae. Each of these phases is very characteristic for a given subfamily (4, 5.6); The manner in which the individual grains are united into tetrads or massulae is a unique feature of the orchid family, for the interfaces among the grains instead of becoming separated by an exine layer remain con- nected through cytoplasmic plasmodesmata. Because of this phenomenon cytogenetically deficient grains are brought to maturity, and in turn can affect fertilization (1). This condition is amply demonstrated by the 212 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 occurrence of large aneuploid series, a source of variation, especially in the subfamily EpmpENDROIDEAE and the more advanced members of the NEOTTIOIDEAE (5, 6). The kind of pollinia, that is, the stage of the aggregation of pollen grains, is the very factor that determines the population size and the spatial distribution of the individuals of each species. In the subfamilies APOSTASIOIDEAE, CYPRIPEDIOIDEAE and NEOTTIOIDEAE both are rather limited. However, in the EpIpeNpRomear in the final stage of evolution of pollinia into firm, compact massulae, permitting the pollinating vectors to carry them over great distances, spatially extensive populations appear as the rule rather than the exception. Yet, the success of this mechanism in regulating as well as maintaining the population-size is closely tied in with the second additional evolutionary innovation referred to above, namely the alteration of one of the stigmatic lobes into a new organ, the rostellum. The rostellum produces either a viscous secretion or a viscous gland to which the pollinia are attached. It is, therefore, a device that ensures successful cross pollination by firmly anchoring the pollinia to vectors for transportation. This evolutionary modification is observable in the more advanced members of the Nrot- TIOIDEAE and in the subfamily EpmpeNpRoIDEAE. There is no rostellum in the APOSTASIOIDEAE, CYPRIPEDIOIDEAE and ORCHIDOIDEAE (6). In addition to anchorage, the rostellum also serves as a device that prevents self pollination. In a few species, however, if the flower is not visited by a pollinator, at the end of anthesis the rostellum dries up, thus allowing the pollinia to come into contact with the stigma and permitting self pollination (5). Orchids are pollinated by a large variety of insects, especially bees, flies, wasps, butterflies and moths, as well as by hummingbirds. Ordinarily the pollinator is species specific. Undoubtedly, cross pollination is merely a by-product of the insect’s activity when attracted by a flower and stimu- lated to satisfy one of the ever-present dynamic urges of nature: hunger or sex. While the flowers of most orchid species provide either nectar or edible tissue as an attractant for the visiting insect, there are a few which successfully mimic the female species of the pollinator, thus attracting them sexually. Recognition of this unusual orchid-insect association which results in pseudocopulation, is a relatively recent addition to our knowl- edge of orchidology, yet the phenomenon itself must be of long standing in the evolutionary history of the family. Although in the annals of orchidology a large number of descriptive cases of individual pollinating mechanisms are recorded, the value of these observations and their evo- lutionary implications will be first realized when the mutual coaptations etween flowers and insects and the reciprocal modifying influence on their evolution are convincingly demonstrated. The evolutionary innovations discussed above, and their mutual inter- actions to a large extent explain the origin of, and present-day diversity within, the family. However, the dispersal of sundry orchid diaspores has played an equally important role in bringing about this diversity, especially 1972] GARAY, ORIGIN OF ORCHIDACEAE 213 when it was followed by some type of isolation. The common isolating barriers among orchid species are physiological, ecological, ethological, or geographic and relatively seldom genetic. ‘This observation is corroborated by literally thousands of artificial hybrids, often between remotely related species, as well as by the occurrence of natural hybrids. The relative frequency of natural hybrid populations of orchids in the subfamily EPIDENDROIDEAE is an additional indication of the rapid evolutionary expansion that took place within a less rigid framework than would be possible with genetic incompatibility as a controlling force. Hybridity whether natural or artificial can occur only within a given subfamily. The lack of genetic compatibility between species belonging to different subfamilies, even in artificially attempted crosses, seems to support the thesis that the kladogenetic differentiation into five subfamilies must indeed have taken place shortly after the family came into existence (6). Closely associated with the initial kladogenetic differentiation of the orchid family is its early expansion and global dispersal. Continental distribution patterns are usually intimately associated with the geologic, climatic, and edaphic history of a given area, hence their explanation is rarely problematic. Transoceanic dispersals, on the other hand, are more perplexing, for the causes often are hidden in the evolutionary past of our biosphere (7). There are approximately 800 genera recognized in the orchid family, and of these only 32 can now be identified as being involved in some kind of oceanic distribution. It is significant that 27 of the 32 taxa, i.e. 85 percent, are genera whose members are exclusively terrestrial. Epiphytic orchids are less habitat-specific in the microenvironment than terrestrials, yet we know only two such genera, Bulbophyllum and Polystachya, with a pantropical range. This situation, however, does not suggest that ter- restrials are more effectively dispersed than epiphytes; it merely indicates that the initial transoceanic expansion must have taken place shortly after the family became established and before the epiphytic mode of life evolved. Since all of the early species were terrestrials, the transoceanic dispersal of epiphytes, although following the same route as the terrestrials did, was of much later occurrence. The necessity of mycorrhizal association is a well-known factor in orchids, but this association is not species specific. This requirement is more readily met in soil conditions where an ample supply of decaying organic matter is available than on exposed branches of trees. Hence, this is another contributing factor accounting for the rather sparse transoceanic dispersal of epiphytes (7). : Orchid genera, whether restricted to adjacent continents or pantropical have the origin of their phyletic affinities in Malaysia (4, Ss4): this center of origin must not be confused with the actual center of America. The exception to this is the transpacific route between Malaysia and South America as exemplified by the genera Tropidia and Erythrodes. 214 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The migratory routes through which the various orchid dispersals were accomplished are amply documented in the pages of phytogeographical literature. They rather convincingly demonstrate that dispersals took place via oceanic routes by island-hopping rather than through supposed con- tinental highways. There is no definite evidence that land connections of a continental nature were in existence during the Tertiary, when most of the transoceanic dispersals took place. Furthermore, geomorphological and oceanographic evidence indicates that the earth is in a constant flux and was even more so in the past. The subsidence of the ocean floor and the fluctuation of the sea level during the various geological periods, combined with orogenic activities in the earth’s crust, have provided time and again favorable avenues between continents for transoceanic dispersals by means of island-hopping. In addition to these factors, wind currents have also played an important role in the transportation of orchid di- aspores. In that, the “intertropical convection current” (doldrums) must Liparis nervosa, and that of the intertropical convection current coincide remarkably well (7). EPILOGUE progressive evolution, i.e., anagenesis. During the process of specialization, natural selection usually causes progressive adaptation to special conditions The story of the evolution of the orchid family is deeply anchored in this phenomenon of extinction concerning which the present-day species with their great diversity, highly specialized and modified structures, and genetic complexities afford us only a dim view. ACKNOWLEDGMENTS The ideas expressed in this paper are based in great part on the disser- tations submitted by me in partial fulfillment of the requirements for the . 1972] GARAY, ORIGIN OF ORCHIDACEAE 215 degree of Master of Science in June, 1961, and for the degree of Doctor of Philosophy in June, 1964, both at Tufts University. I am greatly indebted to Dr. Herman R. Sweet, my friend and teacher, for his unfailing interest, constant encouragement, and guidance throughout the preparation of these two dissertations Dr. Paul A. Warren, Dr. Reed C. Rollins, and Dr. Kenneth D. Roeder have been more influential than they may realize in helping to precipitate some of the ideas expressed herein. For their kindly help I owe them many thanks. LITERATURE CITED — . BARBER, H. N. 1942. r* pollen-grain division in the Orchidaceae. Jour. Genetics 43: 97-1 . De VocEL, E. F. ae eee of the tribe Apostasieae (Orchidaceae). Blumea 17: 313-350. DreEsSLerR, R. L., & C. H. Dopson. 1960. Bova aerg and phylogeny in the Orchidaceae. Ann. Missouri Bot. Gard. 47: 25-68. Garay, L. A. 1960. On the origin of the Orchidiceke. Bot. Mus. Leafl. Harvard Univ. 19: 57-96. 1961. A phylogenetic study of the Orchidaceae: Apostasioideae, red vicina ae Neottioideae. M.Sc. Thesis. Tufts University. [ Unpubl. bo ies) ma 6. 1964. A phylogenetic study of the Orchidaceae: oe Bpidendroideae Ph.D. Thesis. Tufts University. [Unpu i Evolutionary significance of geographical i er of Sa Proc. 4th World Orch. Conf. Singapore 170-187. 8. Hutcutnson, J. 1959. The Families of Flowering Plants. ed. 2. vol. 2. Oxford University Press. 9. Metcutor, H. 1964. Microspermae. Engler’s Syllabus der Pflanzenfamili- en ed. 12: Z 614-625. 10. RENSCH, B. 1959. Evolution above the species level. Methuen & Co., Ltd., fn oe i. STEBBINS, G. & G. S. Kuusu. 1961. Variation in the organization of the ocak complex in the leaf epidermis of monocotyledons and its bearing on their phylogeny. Am. Jour. Bot. 48: 51-59. 12. TaxutajAN, A. L. 1959. Die Evolution der Angiospermen. Gustav Fischer Verlag 13: ebb ey 1965. tas disor of Epipogium in the system of Orchidales. “ay Bot. Neerl. 14: 966. The eet ae oats Acta Bot. Neerl. 15: 224-253. Wuure, 1. H. 1907. On polystely in roots of Orchidaceae. University of Toronto Studies, Biological Series No. 6 ae OrcH1p HERBARIUM OF OAKES AMES BoTANICAL Museum, HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 216 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 CYTOMORPHOLOGY OF ARBORESCENT MORACEAE A. S. HANs THE FAMILY Moraceae is comprised of 53 genera and 1400 species (Willis, 1966), distributed in tropical and warmer parts of the world with a few taxa in the temperate zone. Kanji Lal et al. (1940) describe 11 genera and about 60 species from Assam alone, one of the areas of the present investigation. The Moraceae exhibit variety in habit from her- baceous climbing and epiphytic to giant arborescent forms. The family is important from the point of view of forestry. The species of Morus L. and Artocarpus Forst. produce timbers in Asia, while Piratinera guianensis Aubl.* is a valuable timber of South America. Other economic products are rubber from Castilloa elastica Cerv. and Ficus elastica Roxb., fruits from Artocarpus incisa L.., Ficus carica L., and species of Morus L., paper from Broussonetia papyrifera (L.) Vent. and species of Streblus Lour., and yellow dye from Chlorophora tinctoria D. Don. The leaves of Morus alba L. are the staple food of silk worms in the silk producing countries. The knowledge of chromosome number, morphology, basic number, meiotic behavior, extent and nature of polyploidy, biology of flowering and fruiting of the desired species, and the probable morphological, eco- geographical, and cytological races of these species is essential for a breeding program for amelioration of forest trees. This paper deals with the cytomorphology of important timbers and their related species of the family Moraceae, and is a part of a larger project dealing with cyto- morphological investigations of Himalayan timbers. MATERIALS AND METHODS * The correct name for this seems to be Brosimum guianense (Aubl.) Huber, since Brosimum has been conserved over Piratinera, — Ed TABLE 1. List of the investigated taxa with their localities, cytological data, and results of previous authors TAXON LOCALITY AND COLLECTOR’S NUMBER ROMOSOME NUMBER AND LEVEL OF PLOIDY PREVIOUS REPORT 1. Artocarpus chaplasha Roxb. Disha: pe og Dehing Reserve Forest n = 28, tetraploid 150m 2. A. gomezianus Wall. Digboi: Upper Dehing Reserve Forest, n = 28, tetraploid 150 m.; 3. A. — Lam. Digboi: Chatel 150 m.; 136 n = 28, tetraploid n = 28 (Nanda, 1962), (cultivated) 2n = 56 (Le Cog, 1963) 4, A. lakoocha Roxb. Digboi: Jeypore, 150 m.; 140 n = 28, tetraploid 2m = 56 (Banerji and Hakim, 1954) 5. Broussonetia papyrifera Darjeeling: Sukna, 150 m.; 52 n = 13, diploid 2n = 26 (Bowden, 1940) (L.) Vent. iced: 6. Ficus altissima Bl. Darjeeling: Rongtong, 300 m.; 3118 2n = 26, diploid 2n = 26 (Krause, 1931) 7. F. benjamina L., var Darjeeling: Rongtong, 300 m.; 3167 2n = 26, diploid osa Kurz 8. F. elastica Roxb Darjeeling: Reyang, 200 m.; 3121 2n = 26, diploid 2n = 26 (Sugiura, 1936; Condit, 1964) 9. F. hookeri Miq Darjeeling: Cart Road, 1900 m.; 3182 2n = 26, diploid 10. F. lepidosa Wall ch roa ied Bo . : 3101 26 = a peat 11. F. mysorensis Heyne Darjee 2n = oid 2n = 26 (Condit, 1928) 12. F. nemoralis Wall. var Dasjestiee eo Bieil ae m.3 207 an = oi pric trilepis (Miq.) King 13. F. religiosa L. Darjeeling: Peshoke, 900 m.; 3139 2n = 26, diploid 2n = 26 (Condit, 1933) 14. F. roxburghii Wall Darjeeling: Ging, 1200 m - 102 2n = 26, diploid 15. Morus alba L. (cultivated) Darjeeling: Sukna, 150 m.; 3116 n = 14, diploid 2m = 28 (Tahara, 1910; Osawa, 1920; Delay, 1947; Senaki Ammal, 1948) 16. M., australis Poir. Darjeeling: Ging, 1200 m.; 3114 n = 14, diploid = 14 (Datta, 1954; Das, 17. M. laevigata Wall. Digboi: Nongpoh, 550 m.; 197 n = 28, tetraploid 2n = 28 (Janaki Ammal, 1948) (cultivated) Darjeeling: Lopchu, 1500 m.; 293 2n = 56 (Datta, 1954), Nainital: Dogaon, 2300 m.; 1217 n = 14, diploid n = 28 (Das, 1961) 18. Pseudostreblus indica Bur, Shillong: Barapani, 1200 m.; 178 n = 14, diploid 19. Streblus asper Lour. Shillong: Shella, 300 m.; 221 n = 13, diploid n = 13 (Gajapathy, 1961) [ZZ61 SNVH “ AVAOVAYONW LNAOSTAOAaAV 218 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 tion of 1350 %. The voucher specimens are deposited in the Herbarium, Botany Department, Panjab University, Chandigarh, India. OBSERVATIONS The exact locality and cytological data of the 19 taxa investigated are summarized in TABLE 1. The genera and the species are arranged alpha- betically. A brief description of important taxa is given below. Artocarpus Forst.: Comprised of lofty evergreen trees with about 60 species in Indo-Malaya and China. Out of five Indian species, A. Airsutus Lam., A. chaplasha Roxb. and A. lakoocha Roxb. are timbers of com- mercial importance (Pearson & Brown, 1932). Artocarpus chaplasha Roxb.: A large tree with an average girth of 4.5 meters and 15 to 18 meter clear bole. It is a handsome tree in Assam and Bengal where the soil is rich. In the tropical evergreen forests of upper Assam it is one of the chief elements in the upper story of the Dipterocarpus-Mesua formation. Its chief associes are Amoora wallichii King, Endospermum chinense Benth., Stereospermum personatum Chatt., Tetrameles nudiflora R. Br., Cinnamomum cecidodaphne Neisn., and Dua- banga grandiflora (Roxb. ex DC.) Walp. It is never gregarious but found scattered in mixed plain forests of Bengal and the foothills of Assam. Bark grayish-brown, exfoliating into pieces. MALE RECEPTACLES: March to April. FRUITING RECEPTACLES: June to September. Artocarpus gomezianus Wall. ex Trecul: A rare tree of Tenasserim and Tavoy (Hooker, 1885; Gamble, 1902). It occurs in Assam and the Andamans (Bor, 1953). This species is quite common in the Lakhimpur district of upper Assam where it grows with a straight and cylindrical stem having at least a 10 meter clear bole. Bark gray, dull red inside, splitting into round flakes. MALE RECEPTACLE: March. FRUITING RECEPTACLE: April to June. Artocarpus lakoocha Roxb.: A large tree, distributed from Kumaon eastwards up to Burma. It is found scattered in Bengal and Assam, where it grows to a height of about 30 meters, with a 9 meter clear stem. Bark dark gray, peeling off into small, round, woody plates. MALE RECEPTACLE: February to May. FRUITING RECEPTACLE: June to August. Broussonetia papyrifera (L.) Vent.: A fast growing moderate-sized tree, indigenous to China and Upper Burma, and introduced in India. It does well on both the aspects of forest from 150 to 1500 meters in the 1972] HANS, ARBORESCENT MORACEAE 219 eastern Himalayas and Khasia and the Jaintia Hills. Brandis (1874) remarks “This useful tree seems to accommodate itself readily to different conditions of climate, and might advantageously be cultivated in North- West India.” FLOWER: February to March. Fruit: April to May. Plants of this species are strictly dioecious. Ficus Tourn. ex L.: The largest genus of the family Moraceae em- braces trees, shrubs, climbers, and epiphytes in the tropics and subtropics of both hemispheres; and is most abundantly represented in the islands of the Indian Archipelago and the Pacific Ocean. The number of species in the genus is estimated to be from 600 to 1500 (Condit, 1964). Out of 60 Indian species, 43 are found in Assam and Khasia and Jaintia Hills. In spite of the large number of species and their widespread distribution, the genus is not of much economic importance. Only three species of some importance are mentioned by Pearson and Brown (1932) from India. Ficus elastica Roxb. ex Hornem.: A gigantic tree in the outer hills of the eastern Himalayas from Nepal eastwards in Assam, Khasia and Jaintia hills, and Burma. Lofty trees grow near Tista (150 m.) in the Darjeeling hills and Pakyong (1000 m.) in Sikkim where they were planted for exploitation of India rubber. RIPE FRUIT: July to September. Morus L.: Includes about 12 species (Rehder, 1940), distributed in the temperate and subtropical regions of the northern hemisphere. Four species are met in India, of which three species produce valuable timbers. Morus alba L. Indigenous to China and now naturalized in western Asia. It is found in tropical and subtropical Himalayan tracts from Kash- mir to Sikkim, and Burma. Extensively cultivated in northern India. Frower: April. Fruit: May. Male and female catkins on the same tree. Meiosis is normal with » = 14. In every pollen mother cell, there are one or two larger bivalents than the rest (FIGURE 4). Morus australis Poir. (= M. acidosa Griff., M. indica L.): moderate- sized, fast growing, deciduous tree; from Kashmir to Sikkim in the sub- tropical and temperate Himalayas. This species has been cultivated for a long time and has become naturalized in many parts of India. Bark brown, blaze dull white. FLOWER: March. Fruit: April to May. This species is strictly dioe- cious, the male trees being rare in comparison to the female. 220 JOURNAL OF THE ARNOLD ARBORETUM [voL.. 53 Ficures 1-10, meiotic and mitotic chromosomes in the Moraceae, all X 1350; (1-5, photomicrographs 6-10, photostats). 1, Artocarpus chaplasha, diakinesis with nucleolus, = 28; 2, seudostreblus indica, early metaphase with nucleolus, m = 14, arrow indicates lightly “er eiis: bivalent; 3, Streblus asper, M- = 13; 4, Morus alba, & wo ) PMC a , % = 14, arrows show larger bivalents; 5, M. laevigat ta, MI, = 28, some bivalents associated with each other. 6, M. aus- tralis, M-I, 2 "4, arrow indicates larger bivalent; 7, Ficus hooker 2n = 26; 8, F. elastic, 2n = 26; 9, F. nemoralis var. trilepis, 2n = 26: 10, F. roxburghii, Meiosis is normal with n = 14. One bivalent in every pollen mother cell is definitely larger than the rest (FIGURE 6). Morus laevigata Wall.: Large trees distributed from Kumaon east- wards to Burma. Quite common in moist deciduous forests of Assam and Bengal. Often cultivated in plantations with Chickrassia tabularis A. Juss. 1972] HANS, ARBORESCENT MORACEAE 221 TABLE 2, Comparison of diploid and tetraploid races of Morus laevigata CHARACTER DreLow (n = 14) TETRAPLOID (” = 28) Habit and Small to medium-sized trees in Middle to large tree in istribution W. Himalayas, not common. E. Himalayas, quite common. Branch Slender, slightly hairy. Stout, hairy. Leaf 100-130 & 50-70 mm., 140-190 80-120 mm., hairy. very slightly hairy. Male catkin Up to 60 mm., sparsely hairy. Up to 110 mm., hairy. at 700 m. and with Alnus nepalensis D. Don at 1500 m. The species grows to 25 m. in the eastern Himalayas and up to 15 m. in the western Hima- layas. FLower: March. Fruit: April to May. Two cytotypes seem to exist, one diploid with » = 14 (Janaki Ammal, 1948; B. S. Gill, personal communication) in the western Himalayas, and the other tetraploid with n = 28 (Datta, 1954; Das, 1961) in the eastern Himalayas. The cytotypes differ morphologically in certain char- acteristics (TABLE 2) except in size of stomata and pollen grains where the differences are negligible. My thanks are due to Dr. B. S. Gill who kindly lent his specimens from Nainital (eastern fringes of the W. Himalayas, or rather the central Himalayas) making possible the com- parison between the cytotypes (FicurE 11). Pseudostreblus indica Bur.: Medium-sized trees with restricted dis- tribution in the evergreen climax forests of Surma Valley of Assam and Khasia and Jaintia hills (Barapani forest, 1200 m.). Bark greenish-gray with a few scattered warts. FLoweErR: May to June. Fruit: November to January. Meiosis is normal. One of the bivalents is invariably more lightly stained than the rest (FicureE 2). Streblus asper Lour.: A medium-sized evergreen tree in the sub- Himalayan tract from Beas eastwards up to Burma, and also in the Anda- man Islands. Bark grayish white and warty. FLower: March to April. Fruit: May to June. This species is strictly dioecious and the frequency of occurrence of male trees is greater than the female ones. DISCUSSION The prevalent base numbers in the Moraceae are 13 and 14 (Darling- ton & Wylie, 1955). A great majority of the species are diploid. Poly- 222 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FIGURE 11. Two cytotypes of Morus laevigata. A, tetraploid (7 = 28), from Khasia and Jaintia hills, and eastern Himalayas. B, diploid (~ = 14), from western Himalayas (male catkin broken). ploidy is of rare occurrence in the genus Ficus. Condit (1964) reports seven tetraploid species of Ficus and one triploid, in the horticultural variety ‘Decora’ of F. elastica. Of the known species, about 8 percent are polyploid. The genus Morus is predominantly diploid except for a few natural or artificially raised polyploid species. All the four presently investigated species of Artocarpus are tetraploid. Perusal of literature reveals only one species (A. cannoni) to be polyploid. Diploid (m = 14) and natural tetraploid (n = 28) races of Morus laevigata occur in the western and eastern Himalayas respectively. A comparison of these cytotypes reveals that the tetraploid is of greater height, more hairy, has larger leaves, and longer male catkins than the diploid. The size of the stomata and pollen grains is almost the same in the two races. Das (1961) reports the occurrence of multivalents, laggards and ir- regular separation of chromosomes in the tetraploid Morus laevigata, thereby inferring an imbalanced state of polyploidy. The present findings show normal meiosis and almost 100 percent pollen fertility in specimens from various localities at different altitudes. Some bivalents have the tendency to associate with each other, but do not necessarily result in multivalent formation. It is possible that induction of polyploidy in this 1972] HANS, ARBORESCENT MORACEAE 223 species is recent and that some of the individuals are still in the “floating” state. The presence of one or two larger bivalents in Morus alba and M. aus- tralis was noted by previous workers. Tahara (1909) found two large chromosomes in these two species and labelled them “alpha” and “beta.” Osawa (1920) after investigating seven species and 85 races of Morus, concluded that there was in every case one large pair per diploid com- plement. The irresistible interpretation of the large chromosomes occur- ring in unisexual trees could be as sex chromosomes. Sinoto (1929) infers that the unequal pair in M. bombycis corresponds to the XY mechanism and consequently the males are heterogametic in regard to sex, but adds that they are not directly related to sex in the sense that the effect of them is not seen in the phenotype. The present investigations reveal in fact quite the reverse situation. The larger bivalent(s) is noticed in the pollen mother cells of M. alba where both male and female catkins were present on the same plant, thus ruling out the possibility of their being sex chromosomes. Schaffner (1936) has reported spontaneous sex re- versal in M. alba. Plants destined to be unisexual, bear catkins of the other sex as well. Progeny obtained by selfing these flowers is either uni- sexual or bisexual. He infers that unisexuality in Morus, or to be pre- cise, in M. alba is not affected by hereditary differentials like the XY allosome set, but by physiochemical or physiological variables, both male and female individuals being potentially bisexual. SUMMARY Nineteen woody species of the Moraceae from the eastern Himalayas, the Khasia and Jaintia Hills, and Upper Assam were investigated cyto- morphologically. Pseudostreblus (P. indica, n = 14) is the genus worked out for the first time. The other taxa reported chromosomally for the first time are Artocarpus chaplasha, A. gomezianus (both with n = 28), Ficus benjamina var. comosa, F. hookeri, F. lepidosa, F. nemoralis var. trilepis, and F. roxburghii (all with 2n = 26). Morus laevigata has two cytotypes, one diploid (” = 14) in the western Himalayas, and the other tetraploid (” = 28) in the eastern Himalayas, which differ in respect to habit, hairiness, and dimensions of leaves and male catkins. The dioecious condition in Morus is not governed by an XY allosome set, but by some yet unknown physiochemical or physio- logical factors. ACKNOWLEDGMENTS This study was completed at Panjab University, Botany Department, Chandigarh, India under the valuable guidance of Professor P. N. Mehra, Head of the Department. The financial assistance by the U.S. government with PL 480 funds in India (Grant A7-FS-12) is gratefully acknowledged. 224 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 LITERATURE CITED BANERJI, I., & A. Haxrm. 1954. A contribution to the “2 pages of Arto- carpus lchaocke Roxb. Proc. Indian Acad. Sci. (B) 39: —132. Bor, N. L. 1953. Manual of Indian Forest Botany. Oxford Panes Pre Bowp DEN, W. M. 1940. Diploidy, polyploidy, and winter hardiness solani: in the flowering plants. Am. Jour. Bot. 27: 357-371. ms ri D. 1874. The forest flora of north-west and central India. Wm. H. n & Co., London, 608 a L af 1928. Cytological and morphological studies in the genus Ficus I. Chromosome number and morphology in seven species. Univ. Calif. Publ. Bot. 11: 233-244. - 1933. Cytological and morphological studies in the genus Ficus II. Chromosome number and morphology in 31 species. Univ. Calif. Publ. Bot. 17: 61-74. . 1964, Cytological studies in the genus Ficus III. Chromosome number n 62 species. Madrofio 17: 153-155, Daeecovce, C. D., & A. P. Wyte. 1955. Chromosome atlas of flowering plants. George & Unwin Ltd. 519 pp. Das, B. C. 1961. Cytological studies of Morus indica L. and Morus laevigata Wall. Caryologia 14: 159-162. Datta, a 1954. Cytogenetical studies on two species of Morus. Cytologia 19: 8 DELAY, @ 1947. Recherches sur la structure des noyaux quiescent chez les phanérogames. Revue Cytol. Cytophysiol. Vég. 9: 169-223; 10: 103-229. GAJAPATHY, C. 1961. teat studies in some Indian medicinal plants. Bull. Bot. Surv. India 3: GAMBLE, J. S. 1902. A see of Indian timbers. Sampson Low, Marston & Co., London, 856 Hooker, J. D. 1885. The ‘flora of British India. Vol. 5. L. Reeve & Co. London. JANAKI ‘AMMAL, E. K. 1948, The origin of the black mulberry. Jour. Roy. Kanji Lat, U. N., P. C. Kanjr Lat, R. N. De, & A. Das. 1940. Flora of Assam, Vol. 4. Govt. of Assam, India. Krause, O. 1931. Cytologische studien bei den Urticales. Planta 13: - LE Coo, C. 1963. Contribution a l’étude cyto- taxonomique des ke a des Urticacées. Revue Gen. Bot. 70: 385-426. NANDA, ‘i C. 1962. Chromosome number of some trees and shrubs. Jour. Ind. Bot. Soc. 41: 271-277. Osawa, A 1920. Cytological and experimental studies in Morus with special reference to triploid mutants. Bull. Imp. Agric. Exp. Sta. Tokyo 1: 318. Pearson, R. S., & H. P. Brown. 1932. Commercial Timbers of India. Vol. 2. Govt. of India, Central Publ. Branch, Calcutta. REHDER, A. 1940. Manual of Cultivated Trees and Shrubs. Macmillan Co., New York, ae pp. SCHAFFNER, J. H. 1936. Offspring of or — reversed carpellate plant of Morus alba. Bot. Gaz. 98: 425-4 Stnoto, Y. 1929. Chromosome ree ee in some dioecious plants with special reference to the allosomes. Cytologia 1: 109-191. Sucrura, T. 1936. Studies on chromosomes in ones plants with special refer- ence to cytokinesis I. Cytologia 7: 544-595 1972] HANS, ARBORESCENT MORACEAE 225 Tawara, M. 1909. On the chromosomes of Morus alba. Bot. Mag. Tokyo 23: 343-353. . 1910. Uber die kernteilung bei Morus. Bot. Mag. Tokyo 24: 281-289. Ty10, J. H., & A. LEvAN. 1950. The use of oxyquinoline in chromosome analysis. Anal. Estac. Exp. Aula Dei. 2: 21-64. nibs x C. 1966. A dictionary of the flowering plants and ferns. 7th ed. Re- d by H. K. Airy Shaw. Cambridge Univ. Press, 1214 pp. TREE IMPROVEMENT RESEARCH CENTRE PAO; Box 1210 KITWE, ZAMBIA 226 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 THE PHRYMACEAE IN THE SOUTHEASTERN UNITED STATES? JoHN W. THIERET PHRYMACEAE Schauer in A. P. de Candolle, Prodr. 11: 520. 1847. (LopsEED FAMILy) A monotypic family ? of herbaceous dicotyledons distinguished by per- fect, hypogynous, zygomorphic flowers; synsepalous calyx with the 3 upper lobes subulate and hooked; sympetalous 2-lipped corolla; 4 epi- petalous stamens; pseudomonomerous gynoecium with a basal ortho- tropous to hemianatropous ovule; and an achene enclosed in the —— calyx. TyPE GENUS: Phryma L. The Phrymaceae are of somewhat uncertain taxonomic status. They are regarded as allied most closely to the Verbenaceae, differing principally in the solitary ovule and the superior radicle. In stem anatomy, Phryma is similar to “certain members” of the Verbenaceae (Smith). According to Lipscomb, in general vegetative form, inflorescence type, and flower and fruit anatomy Phryma has its closest affinities with the tribe Lantaneae of the Verbenaceae and is structurally most similar to the genus Stachy- tarpheta; she favors tribal status for Phryma in the Verbenaceae. Other authors (Baillon; Bentham & Hooker; Hutchinson, 1926) have included Phryma among the Verbenaceae, but it is more commonly treated as the sole genus of a distinct family. The Phrymaceae, along with the Globulariaceae and Littorella and *Prepared for a generic flora of the southeastern United States, a one project of the Arnold Arboretum and the Gray Herbarium of Harvard University made possible nan the support of the National Science Soandatinn (Grant “GB_6459X, principal i sie ee E. Wood, Jr.). This treatment follows the format established i in the first paper in the series (Jour. Arnold Arb. 39: 296-346. 1958). The area pera" includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississip ppi, Arkansas, and Louisiana. References that I have not seen are marked with an asterisk. I am indebted to Dr. Wood for his careful review of the manuscript and for other aid; to Ralph Andrews and Fran Uhler for data on wildlife use of Phryma; and to the curators of herbaria who responded to my request for distributional data on ryma. The illustration was made by Rachel A. Wheeler from dissections by Dr. 2M elananthus Walpers was originally described as a second genus of the Phryma- i ok: tae FS a .o a) 2 i =| i msi : ty between it and Phryma. Finally, on anatomical and morphological grounds, Solereder transferred the genus to the Solanaceae, tribe Salpiglossideae. 1972] THIERET, PHRYMACEAE 227 Bougueria of the Plantaginaceae, provide good examples of pseudomono- merous gynoecia (Eckhardt, Wilson & Just). Such a gynoecium, actually a syncarpous structure, has the appearance of a single carpel; it develops through suppression of all but one of the original carpels. In Phryma, the gynoecium is viewed as being two-carpellate, with one carpel (the pos- terior?) being reduced. Lipscomb found the gynoecium to be “bicarpellate as indicated by two dorsal bundles which extend the length of the pistil.” 1. Phryma Linnaeus, Sp. Pl. 2: 601. 1753; Gen. Pl. ed. 5. 262. 1754. Simple or branched perennial herbs. Leaves opposite, simple, serrate to crenate-serrate, the lower long-petioled, the upper nearly sessile. In- florescences of terminal, long-peduncled, spikelike racemes. Flowers per- fect, zygomorphic, hypogynous, axillary to bracts, horizontal at anthesis, ae Vex 4 ars small plant with flowers and young fruit wer, X 4; ¢, corolla laid open, < 4; d, anther, X 20 gynoecium in vertical section to show ovule, ye 7S, " X 15; f, nearly mature fruiting calyces, X 4; g, fruit, X 5; h, embryo, x 5. Ficure 1. Phryma. a-h, P. leptostachya var. agains a, upper part of x 3/8; b, fo sharply reflexed soon thereafter, opposite or sometimes the lower ones alternate, the pedicels short, bibracteolate. Calyx persistent, synsepalous, cylindrical, strongly zygomorphic, 2-lipped, the upper lip of 3 subulate, hooked lobes about equalling the tube at anthesis, the lower much shorter, of 2 triangular lobes, the tube sulcate, accrescent in fruit. Corolla sym- petalous, the tube cylindrical, the limb 2-lipped, the upper lip upwardly diverging, notched, external in the bud, the lower lip much longer, hori- zontally projecting, 3-lobed, the palate uparching, 2-ridged centrally. An- droecium didynamous, the stamens included, alternate, epipetalous, the 228 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 week a \ | : ~ °° e 5 “- i “ ies . oe. °° soe + — es) sf e 2 | "3 en . . e - we ieee a . . oes oe.oere e® °4 | re bed 5 ee s : : I. Lisianthius sect. Omphalostigma (Griseb.) Weaver, comb, nov. Leianthus sect. Omphalostigma Griseb., pro parte, Gen. & Sp. Gent. 198. 1839 [1838]. Annual or perennial herbs, glabrous or minutely spiculate. Inflorescences terminal and axillary on the primary and secondary shoots, basically of compound dichasia. Corolla salverform, the limb widely flaring, greenish 236 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 or greenish-white, small, the tube 15 mm. long or less. Stamens inserted near the apex of the corolla tube, the filaments equal in one flower, less than 9 mm. long. Sculpturing of pollen grains reticulate. LECTOTYPE SPECIES: L. saponarioides Cham. & Schlecht. The section OMPHALOSTIGMA was defined more or less tentatively by Grisebach (1839) in his genus Leianthus to include the two Mexican species known at that time, L. saponarioides (Cham. & Schlecht.) Griseb. (= Lisianthius saponarioides Cham. & Schlecht.) and L. nigrescens (Cham. & Schlecht.) Griseb. (= Lisianthius nigrescens Cham. & Schlecht.). These species reputedly differed from the West Indian species then known in having peltate rather than capitate stigmas and corollas swollen at the base and narrowed above, rather than narrowed at the base. Stigmatic type has proved to be generally unreliable as an indicator of relationships in Lisianthius. According to my interpretation, L. nigrescens belongs in the typical section of the genus. Therefore L. saponarioides, by exclusion of L. nigrescens, becomes the lectotype species of Grisebach’s section. At any rate, although both L. nigrescens and L. saponarioides have peltate stigmas, only L. saponarioides is characterized by corollas which are swollen at the base and narrowed above, and then only in older flowers. L. saponarioides therefore best demonstrates the characters men- tioned above as supposedly distinctive of sect. OMPHALOSTIGMA according to Grisebach, and, in any event, would be the better choice as a lectotype species. KEY TO THE SPECIES OF LISIANTHIUS SECTION OMPHALOSTIGMA Inflorescences igaeage the dichasia often — and appearing eres calyx 5.5-10 mm. long, the lobes 4-7.5 mm. long, surpassing vary at anthesis: corolla ‘14 2.1 cm. long, the lobes 4-7 mm. long fe cuspidate- acuminate; filaments 3-8.5 mm. long; flowers subsessile or very short pedicel- late, the pedicels not more than 1 mm. long; plants glabrous or minutely spiculate, but the calyx and the midveins of the leaves always glabrous; calyx in fruit 2/3 as long to nearly as long as the capsules. .... 1. L. saponarioides. Inflorescences diffuse, the dichasia always pedunculate and never appearing pleiochasial; calyx 3.5-5 mm. long, the lobes 2-3 mm. long, never surpassing the ovary; corolla 1.2-1.7 cm. long, the lobes 2.5-4.5 mm. long and acute or rarely short acuminate; filaments 2-4 mm. long; flowers distinctly pedicellate, the pedicels 2-4 mm. long: plants minutely spiculate, including the calyx tube and the midveins of the leaves; calyx in fruit usually 1/3 and never more than 1/2 as long as the capsule. 2. L. meianthus. 1. Lisianthius saponarioides Cham. & Schlecht. Linnaea 6: 389. 1831 (as Lisianthus), “Type. Mexico. Veracruz: Ad margines syl- varum Papantlae, Schiede (HAL! ). Leianthus a (Cham. & Schlecht.) Griseb. Gen. & Sp. Gent. 198. 1839 [1838]. 1972] WEAVER, REVISION OF LISIANTHIUS 237 Petasostylis saponarioides (Cham. & Schlecht.) Griseb. in DC. Prodr. 9: 71. 1845 Lisianthus congestus Standl. Carnegie Inst. Publ. 461: 82. 1935. “Type. temala. PreteN: Sabana de San Francisco, La Libertad, Lundell 2479 Rather coarse, erect, monopodial, perennial (?) herbs, glabrous or minutely spiculate. Stems 8-14 dm. tall, green, minutely spiculate (with magnification) or more rarely essentially glabrous, terete below but 4-angled above, the angles usually minutely serrulate. Leaves (1.6)3.2—12 cm. long and (0.7) 1.3—4 cm. broad, dull textured on both surfaces, papil- lose above and with the margins of the upper leaves often very minutely ciliolate, the lateral veins inconspicuous below; the principal foliage leaves narrowly elliptic (or ovate-lanceolate), long-acuminate, long-attenuate at the base, essentially sessile or with a short, winged petiole to 3 mm. long; the upper leaves and those on the secondary branches distinctly ovate or ovate-lanceolate, rounded or subcordate at the base, distinctly short- petiolate. Inflorescences of sessile or pedunculate dichasia, these usually 1-2(—4) times compound but often reduced and appearing simple, borne singly or in groups of 3, the terminal one often sessile and the whole unit appearing pleiochasial; inflorescence branches usually contracted and the dichasia or groups of dichasia dense and compact, more rarely some- what elongate and the dichasia more diffuse, 4-angled, 2 of the angles often expanded into scarious wings, minutely spiculate; bracteoles scari- ous, subulate. Flowers erect or ascending, subsessile or very short pedi- cellate, the pedicels not exceeding 1 mm. in length. Calyx 5.5-10 mm. long, fused for 2/10-3/10 its length; the lobes 4—7.5 mm. long and 1-1.5 mm. broad, lanceolate, long-acuminate, scarious margined, ridged abaxially; the tube 1.5—2.5 mm. long. Corolla 1.4—2.1 mm. long, salverform, greenish-white or green; the tube at anthesis narrow, 10-15 mm. long and ca. 1.5 mm. in diameter, conspicuously white-striate at least at the base; the lobes 4-7 mm. long and 2—4.5 mm. broad, flaring and recurved, ovate- lanceolate or more rarely somewhat elliptic, long-acuminate, conspicu- ously longer than broad, usually 1/3 as long as the tube but varying from 3/10-4/10 as long. Stamens inserted near the apex of the corolla tube, in the upper 1/6-1/3, at or just above the apex of the calyx lobes; fila- ments subequal, (3)4.5—7.5(8.5) mm. long, surpassing the corolla tube but not the lobes, usually surpassing the style; anthers 1.5-2.5 mm. long and ca. 1 mm. broad, oblong, becoming almost sagittate upon dehiscence. Style 6-10 mm. long, surpassing the corolla tube; stigma peltate. Cap- sule oblong-ellipsoid, 5-10 mm. long and 3-4 mm. broad; the marcescent corolla apparently not persisting until the seeds are shed. FLOWERING: All year. DisTRIBUTION: From 600 to 1200 meters on rocky limestone hillsides and in secondary scrub from the State of Veracruz, Mexico, south and east to Chiapas, the departments of Huehuetenango and Petén, Guatemala, and western British Honduras (Map 1). 238 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 REPRESENTATIVE COLLECTIONS exico. CHIAPAS: Laguna Ocotal Grande, ca. 25-30 km. se. of Cerro Libano, Dressler 1481 (13, MEXU, US). VERACRUZ: ad margines sylvarum Papantlae, Schiede & Deppe s.n. (BM). Guatemala. HUEHUETENANGO: cafetal of Finca Soledad, 5 mi. se. of Barillas, Sierra de los Cuchumatanes, Steyermark 49547 (F). PeTtEN: La Cumbre, w. of km. 139 of Cardenas Road, Contreras 6205 (F); Lake Petén Itza, 2 km. s. of Santa Elena, on La Cueva Road, Contreras 2060 (LL); Sabana San Francisco, La Libertad, Lundell 2479™(MicH, isotype of L congestus); Tikal, on top of Temple IV, Contreras 166 (¥, 1J, LL). British Honduras. Cayo Dist.: hill se. of Millionario, Proctor 29820 (DUKE, IJ); near Camp 6, Gentle 2363 (A, F, LL, NY). Lisianthius saponarioides is a rather poorly collected plant of which only 28 specimens were available for study. With their small, greenish, salverform corollas and short stamens inserted near the apex of the corolla tube, L. saponarioides and L. meianthus form a very distinctive group unlikely to be confused with any other species in the genus. Despite its distinctiveness, Lisianthius saponarioides has had a rather confused history. Until recently, the name usually was applied incor- rectly to plants recognized here as L. acuminatus and L. quichensis, both morphologically very different from L. saponarioides. The reason for the confusion is unclear. Chamisso and Schlechtendal, in the original descrip- tion of L. saponarioides, describe the corolla as “. . . hypocrateriformis, tubo tertia circiter parte calycem superans ...” Certainly no other species in the genus, known then or now, is characterized by having a salverform corolla with the calyx reaching to 2/3 of the length of the corolla tube. Grisebach (1839, 1845) evidently applied the name correctly (as Leian- thus saponarioides, 1839, and Petasostylis saponarioides, 1845). Hemsley (1882) cited Sumichrast 1558, a fragment of this collection now the lecto- type of L. acuminatus Perk., as representative of L. saponarioides. Gilg (1895) included a figure of a plant (p. 92, Figure 41-A) which he called L. saponarioides, but which undoubtedly represents L. acuminatus Perk., a fact recognized and pointed out by Perkins (1902, p. 493). However, Perkins (loc. cit., p. 491) cited Heyde & Lux 2921, from Dept. Quiché Guatemala, as representative of L. saponarioides. H eyde & Lux 2921 super- ficially resembles L. acuminatus much more than it does L. saponarioides. The specimen of Heyde & Lux 2921 (us) was later selected by John Donnell Smith as the holotype of L. guichensis. Although Williams (1968, 1969) considered Lisianthius congestus Stand]. to be synonymous with L. saponarioides, he did it with some reservation. I have seen the holotypes of both L. congestus and L. saponarioides; they are similar in all respects and undoubtedly represent the same taxon. There was no correctly determined material of Lisianthius saponarioides in American herbaria in 1935 when Standley described L. congestus. There- fore, in view of the confused application of the name at that time, it is quite possible that Standley did not consider L. saponarioides when de- scribing his new species. He did not compare L. congestus to any known 1972] WEAVER, REVISION OF LISIANTHIUS 239 species but rather stated that it is “distinguishable at once” from other Central American species “by the densely flowered, congested inflores- cence.” 2. Lisianthius meianthus Donn. Sm. Bot. Gaz. 52: 51. 1911 (as Lisianthus). “Tyrer. Guatemala: Sacolal, bei Pansamala, von Tuerckheim 1436 (us!). Rather coarse, annual herbs, minutely pubescent. Stems 7-14 dm. tall and to 5 mm. in diameter, erect, green, terete below but 4-angled above, minutely and densely spiculate in the upper portions. Leaves (0.7)1.5—9 cm. long and (0.3)1.3-3.5 cm. broad, dull textured on both surfaces, densely papillose above and with very minutely ciliolate margins, the midvein and the principal lateral nerves conspicuous below, minutely spiculate; principal foliage leaves lanceolate or ovate-lanceolate, rarely tending toward elliptic, long-acuminate, attenuate at the base, essentially sessile or with short, winged petioles to 2 mm. long; the upper leaves and those on the secondary branches distinctly lanceolate, rounded or sub- cordate at the base, distinctly short-petiolate. Inflorescences of distinctly pedunculate dichasia, these usually 1-2 times compound but often reduced and appearing simple, borne singly or in groups of 3, the ultimate branches often scorpioid; inflorescence branches elongate and the dichasia or groups of dichasia loose and open, 4-angled, densely spiculate. Flowers ascending or horizontal, the pedicels 2-4 mm. long and minutely spiculate; bracteoles subulate, scarious. Calyx 3.5-5 mm. long, fused for 3/10 — nearly 1/2 its length; the lobes 2-3 mm. long and ca. 1 mm. wide, lanceolate, acumi- nate, scarious-margined, ridged abaxially; the tube 1-2 mm. long, minutely spiculate. Corolla 12-16.5 mm. long, salverform; the tube at anthesis nar- row, 9-12.5 mm. long and ca. 1.5 mm. in diameter at the base, greenish, conspicuously white-striate at least at the base; the lobes flaring, 2.5—4.5 mm. long and 1.5—2.5 mm. broad, oblong-oblanceolate or oblong-elliptic, acute or rarely short-acuminate, longer than broad, 1/3-1/5 as long as the tube. Stamens inserted near the apex of the corolla tube, in the upper 1/5, well above the apex of the calyx lobes; filaments subequal 2—4 mm, long, surpassing the corolla tube, equalling or slightly exceeding the style; anthers ca. 1.5 mm. long and 1 mm. broad, oblong, becoming almost sagittate upon dehiscence. Style 5-8 mm. long, surpassing the corolla tube; the stigma peltate. Capsule oblong-ellipsoid, beaked, 6-9 mm. long, not including the 1-3.5 mm. long beak, and 2-4 mm. in diameter ; the marcescent corolla not persisting until the seeds are shed. FLOWERING: Oct.—Apr. DISTRIBUTION. From near sea level to 1500 meters in pine forests, limestone thickets, and along roadsides in central, southern, and eastern Guatemala (Map 1). SPECIMENS EXAMINED Guatemala. ALTA VeRAPAz: Saquija, 45 km. ne. of Coban, Standley 70166 (Ny, F); between Coban and Finca Chimoté, near Rubeltein, Steyermark 44179 240 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 (F); woods between Finca Gubilguitz and Hacienda Yaxcabanal, Steyermark 44821 (F, NY). IzaBAL: along Highway CA-9 near turnoff to Mariscos, Weaver 2186 (DUKE); between milla 49.5 and ridge 6 mi. from Izabal, Montafa del Mico, Steyermark 38534 (GH, F). SUCHITEPEQUEZ: Finca Moca, Skutch 1568 (F, US). A very poorly collected species, Lisianthius meianthus is the only one in the genus definitely known to be an annual. Lisianthius meianthus is similar to L. saponarioides and was combined with it by Williams (1968, 1969). As outlined in the key, L. meianthus differs from L. saponarioides in having the principal veins of the leaf (lower surface) and the calyx tube spiculate; an open loose inflorescence; pedicels at least 2 mm. long; calyx shorter and never equalling the ovary; and the corolla lobes shorter, and acute or short-acuminate. These differ- ences are constant and diagnostic, and all of the available specimens are easily assignable to one taxon or the other. Whether the two taxa, L. meianthus and L. saponarioides, are specifi- cally or only varietally distinct is perhaps open to question. Until more complete collections are available for both, I prefer to treat L. meianthus and L. saponarioides as distinct species. II. Lisianthius section Lisianthius. Subshrubs, shrubs, or slender trees, more rarely suffrutescent, perennial herbs, very rarely annual, glabrous or puberulous, Inflorescences terminal and axillary on primary and secondary shoots, or on specialized determi- nate flowering branches, of compound or apparently simple dichasia, more rarely reduced to single flowers. Corolla tubular or funnelform, the limb erect, spreading, or recurved, the tube yellow, or more rarely red or black. Stamens inserted in the lower 1/3-1 /2 of the corolla tube, the filaments of unequal length in a given flower, more than 10 mm. in length, usually much longer. Type spectes: L, longifolius L. This, by far the larger section, may be divided into two subsections, based primarily on differences in habit. KEY TO THE SUBSECTIONS OF LISIANTHIUS SECTION LISIANTHIUS Perennial, suffrutescent, very rarely annual, herbs, dying back to the woody base after each flowering season; plants monopodial, the main axis determinate; branching excurrent: inflorescences typically of compound dichasia, terminal and axillary on the primary and secondary shoots, never reduced to solitary axillary flowers; capsules ovoid or more rarely oblong-ellipsoid; principal foliage leaves sessile and usually clasping; corolla tube yellow, red, or black. i a A. subsect. HERBACEI. 1972] WEAVER, REVISION OF LISIANTHIUS 241 solitary, axillary flowers; capsules oblong-ellipsoid or fusiform; principal foliage leaves petiolate (or subsessile), never clasping; corolla tube yellow or WErY Wares POI oe). Baka ee, ee B. subsect. FRUTICOsI. Lisianthius sect. Lisianthius subsect. Herbacei Weaver, subsect. nov. Herbae perennae suffrutescentes, grossae monopodiales, raro herbae annuae, omnino glabrae. Ramificatio excurrens. Folia sessilia plerumque amplexicauliaque. Dichasia vulgo composita, terminalia axillariaque. Cap- sulae ovoideae vel raro oblongo-ellipsoideae. SPECIES TyPICA: L. nigrescens Cham. & Schlecht. L 1380/236 The species in subsect. HERBACEI, in contrast to those in subsect. FRu- TICOSI, are suffrutescent herbs with terminal inflorescences. One of them, Lisianthius silenifolius, is possibly an annual. Six of the seven species are distributed in Mexico and Guatemala, the remaining one, L. silenifolius, in Cuba. KEY TO THE SPECIES OF SUBSECTION HERBACEI Corolla tube dull red; corolla lobes obtuse and somewhat erose, very rarely acute; calyx fused for 1/3-1/2 its length, fleshy, the lobes narrowly ovate, acute or short-acuminate, completely ecarinate; anthers purple. .......... Sc piel nes tyes ey a ee ee eS ae caren ee 9. L. viscidiflorus. Corolla tube yellow or black; corolla lobes acuminate or more rarely acute; calyx fused for 1/4 its length or less, not fleshy, the lobes lanceolate, long- acuminate or more rarely acute, ridged or carinate abaxially; anthers yellow. B. Corolla lobes ovate to squarish or suborbicular, less than twice as long as broad, 2-4 mm. long, 1/10-1/8 as long as the tube; corolla narrowly tubular; inflorescences congested or rarely diffuse, the dichasia often sessile and appearing pleiochasial; i ; winged; stamens inserted on the corolla tube ca. 1/2 the distance from. the bese to. the ANGE. os 6 i 8. L. Corolla lobes ovate-lanceolate or lanceolate, more than twice as long as broad, (3)5-24 mm. long, more than 1/5 as long as the tube; corolla broadly tubular or funnelform; inflorescences diffuse, the dichasia never sessile or appearing pleiochasial; inflorescence branches terete or indis- tinctly 4-angled, very rarely with narrow, scario ings; stamens in- serted on the corolla tube ca. 1/3 of the distance from the base to the > = wo apex. C. Corolla black; capsules ovoid or fusiform. ae D. Capsules fusiform; inflorescence branches distinctly ‘ angles expanded into narrow, scarious wings, these minutely ser- rulate; inflorescences of apparently simple dichasia, very rarely of compound dichasia; leaves dull green and papillose on the upper surface, weakly clasping at the base; corolla lobes more than 1/2-3/4 as long as the tube; stamens inserted below the apex of the calyx lobes; styles seldom surpassing the @ 3 > 8 =) S g | vl =] th 5 o wn ra) o =] a © 242 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 narrow scarious wings; inflorescences of compound dichasia, rarely of apparently simple dichasia; leaves glossy and not papillose on the upper surface, strongly clasping at the base; corolla lobes usually 1/6-1/2 as long as the tube, very rarely more than 1/2 as long; stamens inserted above the apex of the calyx lobes; styles nearly always equalling or surpassing the corolla lobes. .... ieee tee as one hace ile ie ers eres ee 4. L. nigrescens. C. Corolla yellow; capsules ovoid and usually broadly so. E. Foliage leaves rounded or truncate at the base, clasping; corolla lobes spreading or recurved; corolla tube not at all inflated, widest at the apex; styles not surpassing to conspicuously sur- passing the corolla lobes. F. Corolla lobes recurved, 11-20 mm. long, more than 1/3 as long as the tube; calyx 9-15 mm. long; style equalling to conspicuously surpassing the corolla lobes; plants of Mexico. eed ee eee Oa ee 5. L. oreopolus. F. Corolla lobes spreading, 5.5-9 mm. long, less than 1/3 as long as the tube; calyx 6-11.5 mm. long; style never surpassing the corolla lobes; plants of Cuba. ......... 6. L. silenifolius. E. Foliage leaves narrowed at the base, not clasping; corolla lobes erect; corolla tube slightly inflated, widest below the apex; styles always surpassing the corolla lobes. .......... 7. L. quichensis. 3. Lisianthius cuspidatus Bertol. Nov. Comm. Bonon. 4: 408. ¢. 37. 1840 (as Lisianthus). “Typr. Guatemala, without exact locality, Veldsquez, presumably at Bologna (BoLo) but not seen. Leianthus cuspidatus (Bertol.) Griseb. in DC. Prodr. 9: 82. 1845. Lisianthus nigrescens var. cuspidatus (Bertol.) L.O. Wms. Fieldiana Bot. 31: 408. 1968. _Rather coarse, perennial herbs. Stems 0.7—3 m. tall and to 1 cm. in diameter, erect or ascending, green or often purplish, terete below and 4-angled above, the angles expanded into narrow, scarious wings. Leaves (0.9)2.4-10.1 cm. long and (0.1)0.4—2.7 cm. broad, firmly membranaceous, the margins minutely papillose, the principal foliage leaves lanceolate or more rarely ovate-lanceolate, long-acuminate, narrowly tapered to rounded at the base, weakly clasping, the base usually less than twice the diameter of the stem; upper leaves and those on the secondary branches lanceolate, the bases rounded or cuneate, short-petiolate or sessile but not clasping. Inflorescences composed of dichasia, usually 2—3-flowered and appearing simple, rarely 1-compound, long-pedunculate; inflorescence branches strongly 4-angled, elongate and the inflorescences appearing loose and open. Flowers drooping, long-pedicellate, the pedicels 0.5—-2.2 cm. long; bracteoles subulate, scarious. Calyx 10-14.5 mm. long, fused for 1/6—-1/5 its length; the lobes 8-12 mm. long, lanceolate, long-acuminate, scarious- margined, ridged abaxially; the tube 1.5-2.5 mm. long. Corolla funnel- form, uniformly black, 4-5.6 cm. long; the tube 2.6-3.3 cm. long and 7-8 mm. in diameter at the apex; the lobes 1.4~-2.4 cm. long and 3—5 mm. broad, spreading and slightly recurved at the tips, lanceolate or oblong- 1972] WEAVER, REVISION OF LISIANTHIUS 243 lanceolate, long cuspidate-acuminate, 1/2-3/4 as long as the tube. Stamens inserted on the corolla tube ca. 1/3 of the distance from the base to the apex; filaments 1.8—2.9 cm. long, only slightly surpassing the corolla tube; anthers oblong, 2—2.5 mm. long and 1—1.5 mm. broad. Style 2.1-2.9 cm. long, surpassing the corolla tube but not reaching to more than the midpoint on the corolla lobes; stigma peltate. Capsule fusiform, beaked, 1.1-1.5 cm. long, not including the 1-2 mm. long beak, and 3-5 mm. in diameter; the marcescent corolla seldom persisting until the seeds are shed. FLOWERING: Aug.—Nov DIsTRIBUTION: In pine forests about 2000 meters elevation in central Guatemala (Map 2). SPECIMENS EXAMINED Guatemala: without nar pea 320 Ryle without locality, Bernoulli & Cario 1804 (s); San Martin, Johnston s.n, (F). Et Procresso: Sierra de las Minas, Finca el Se. 1 km. from: ‘ highway from Salama to El Rancho, Hawkes, Hjerting & Lester 1968 (¥, s); between Calera and middle slopes of quebradas on Volcan - a hae 43001 (F, NY). GUATEMALA: in the Sapoie barranca near Guatemala, Hayes s.n. (GH); San Juan Sacatepéquez, Lewis 895 (F). QUICHE: Sten locality, pie 1400 (F). Lisianthius cuspidatus is a poorly collected species; only eleven speci- mens were available for study. It is quite similar to L. migrescens var. chiapensis, especially in its black corollas. I have not seen Bertoloni’s type, or any authentic material known to have been in his possession when the original description was made. How- ever, his diagnosis and the plate accompanying it are perfectly adequate and leave no doubt as to the plant he had at hand. The relationship of Lisianthius cuspidatus to L. nigrescens, the other black-flowered species, has been the subject of considerable debate since L. cuspidatus was originally described. Grisebach (1845) and Robinson (1910) considered the two as distinct species. In fact Grisebach even placed them in different genera, cuspidatus in Leianthus and nigrescens in Petasostylis. Hemsley (1882), Kuntze (1891), and Perkins (1902) reduced L. cuspidatus to synonymy under L. nigrescens. Williams (1968, 1969) treated L. cuspidatus as a variety of L. nigrescens. His L. nigrescens var. cuspidatus, as discussed later, included plants here treated as L. nigrescens var. chiapensis and L. cuspidatus. Williams (1968) felt that the characters used by Robinson (1910) to separate Lisianthius cuspidatus from L. nigrescens, i.e. the subcuneate leaf bases, the more deeply lobed corolla, and the corolla lobes distinctly sur- passing the style “begin to lose their significance with more adequate col- lections.” My observations would lead me to disagree with Williams; although there is a small overlap in lobe/tube ratio between L. cuspidatus and L. nigrescens (var. chiapensis), the other characters hold quite well. In addition the 3-flowered dichasia, serrulate stem angles, dull-textured, papillose leaves, and especially the fusiform capsules clearly set L. cuspi- gS, circles). Map 3, L ! AP 2, L. tus (t , L. nigrescens var. chiapensis (half - Or ntatus var. brevidentatus (dots . brevidentatus var. collinus (half-circ ) evi- I e), 1972] WEAVER, REVISION OF LISIANTHIUS 245 datus apart from L. nigrescens. I think that the differences between the two taxa are of sufficient magnitude to warrant maintaining L. cuspidatus as a distinct species. 4. Lisianthius nigrescens Cham. & Schlecht. Linnaea 6: 388. 1831 (as Lisianthus). Coarse, perennial herbs. Stems 0.5—2 m. tall and to 2 cm. in diameter, erect or ascending, green except at base, terete below but indistinctly 4-angled above. Leaves firmly membranaceous, the lateral veins promi- nent, the principal foliage leaves 3-19 cm. long and 1-6.4 cm. broad, ovate-lanceolate or elliptic or sometimes rather oblong or even somewhat pandurate, gradually or abruptly acuminate, the bases truncate and auric- ulate, distinctly clasping, much broader than the stem, the upper leaves and those on the secondary branches smaller, lanceolate or ovate, rounded at the base, short-petiolate or sessile but not clasping. Inflorescences of long-pedunculate dichasia, these 1 to 4 times compound, the lateral branches often scorpioid, or sometimes reduced and appearing simple; inflorescence branches terete or indistinctly 4-angled, elongate with the dichasia appearing loose and open. Flowers horizontal or nodding, the pedicels (0.3)0.6—-3.0(4.0) cm. long; bracteoles subulate, scarious. Calyx (6)7.5-13.5 mm. long, fused for 1/8—1/4 its length; the lobes (4.5) 7-11 mm. long, and 1.5—2.5 mm. broad, lanceolate, long-acuminate, scarious- margined, carinate or at least ridged abaxially. Corolla funnelform, uni- formly black, glossy outside and dull within, 3.2-5.4 cm. long; the tube 2.7—3.9 cm. long and 4-7 mm. in diameter at the apex; the lobes 0.5— 1.6(2.0) cm. long and 2-4 mm. broad, 1/6—1/2(3/5) as long as the tube, lanceolate or ovate-lanceolate, gradually short-acuminate or rather long cuspidate-acuminate, spreading, recurved only at the tips if at all. Sta- mens inserted on the corolla tube ca. 1/3 of the distance from the base to the apex; filaments 2.1-3.4 cm. long, exceeding the corolla tube and fre- quently the lobes, the longest ones equalling the style or not; anthers 2-3.5 mm. long and 1-2 mm. broad, broadly oblong, truncate or obtuse at the apex, the connective very slightly if at all exserted. Ovary ovoid; style 2.1-4.5 cm. long, exserted, not at all to conspicuously exceeding the corolla lobes. Capsules ovoid and frequently broadly so, 7-13 mm. long and 3-6 mm. broad; the marcescent corolla strongly persisting nt after the seeds are shed. Key TO THE VARIETIES Corolla lobes 9 mm. or less in length, gradually short-acuminate; filaments, at least the longest ones, usually surpassing the co si lobes; styles always surpassing the corolla lobes, and often al teaes Pe a eal S Pee Fe ee a) a 4a. L. nigrescens var. nigrescens. Corolla lobes 9-20 mm. long, but usually more than 11 mm., long cuspidate- acuminate; filaments never surpassing the stand lobes; aro at most pawl 246 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ling or barely surpassing the corolla lobes. OA od SU Ne U OER SS a SR NS Sep i eral: 4b. L. nigrescens var. chiapensis. 4a. Lisianthius nigrescens Cham. & Schlecht. var. nigrescens. Lisianthius nigrescens Cham. & Schlecht. Linnaea 6: 388. 1831 (as Lisianthus). YPE. Mexico. VERACRUZ: in sylvis Papantla, Schiede (HAL!). Leianthus nigrescens (Cham. & Schlecht.) Griseb. Gen. & Sp. Gent. 199. 1839 [1838]. Petasostylis nigrescens (Cham. & Schlecht.) Griseb. in DC. Prodr. 9: 71. Corolla 3.24.6 cm. long, the lobes lanceolate, gradually short-acuminate, 5—8(9) mm. long and 2-3 mm. broad, 1/6—1/4 as long as the tube. Fila- ments 2.1—3.3 cm. long, at least the longest ones in each flower equalling or surpassing the corolla lobes. Styles 2.7-3.5(4.2) cm. long, exceeding the corolla lobes (often conspicuously), always surpassing the styles. FLOWERING: Jan._Aug. CHROMOSOME NUMBER: n = 18. DisTRIBUTION: From 800 to 1400 meters on dry rocky hillsides and on roadside banks in Veracruz, Hidalgo, and Oaxaca, Mexico (Map 2). REPRESENTATIVE COLLECTIONS Mexico: without exact locality, Sessé, Mocino, et al. 5097 (Fr). H1atco: alrededores de Huejutla, Maury 6013 (mexvu). Oaxaca: I km. e. of Tonaguilla, Schultes 591 (GH); vicinity of Choapan, Nelson 893 (GH, US). VERACRUZ: otoa, Galeotti 1473 (¥, GH, NY, US); Orizaba, Purpus 1182 (F, GH, MO, NY, uc); Cerro de Escamela, on road to Ojo de Agua, just e. of Orizaba, Weaver 2134 (A, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, S, UC, us); Papantla, Liebmann 10787 (GH, F, MO, s, UC, Us); Zacuapan, Purpus 5994 (cH, Mo, UC); from El Rancho de Sacalica to El Municipio de Zongolica, Santos 2075 (MIcH); Cérdoba, Miranda 632 (MExv), Williams (1969) considered Cook 65 (us) the only known Guatemalan specimen that represents the typical variety. This specimen is inter- mediate in character between var. nigrescens and var. chiapensis, but is closer to the latter especially in the shape of the corolla lobes and the fact that the style does not exceed the corolla lobes. 4b. Lisianthius nigrescens Cham. & Schlecht. var. chiapensis Weaver, var. nov. Type. Mexico. Cutapas: 7.8 mi. e. of Chiapa de Corzo on Highway 190, Weaver 2168 (puxE! ye Habitu et foliis varietati typicae similis, differt lobis corollae longiori- bus, longe cuspidato-acuminatis, longitudine vulgo 1/2 tubi aequantibus, et stylum vix vel paulo superantibus. Corolla 3.7—5.4 cm. long, the lobes lanceolate or ovate-lanceolate, rather long cuspidate-acuminate (9)11-16(20) mm. long and 3-4 mm. broad, 3/10-1/2(3/5) as long as the tube. Filaments 2.3-3.4 cm. long, never surpassing the corolla lobes, sometimes equalling the style. Styles 2.5-4.5 cm. long, at most equalling or very slightly surpassing the corolla lobes. FLOWERING: Apr.—Oct. CHROMOSOME NUMBER: n = 18 1972] WEAVER, REVISION OF LISIANTHIUS 247 DISTRIBUTION: From 500-1800 meters in pine and/or oak forests, secondary scrub, and along roadsides in Chiapas, Mexico, and Huehue- tenango, Guatemala (Map 2). REPRESENTATIVE COLLECTIONS Mexico. CHIAPAS: without exact locality, Ghiesbreght 702 (BM, GH, MO, NY); 7.8 mi. e. of Chiapa de Corzo on Highway 190, Weaver 2168 (A, BM, DS, F, IJ, LL, MICH, MO, NY, Ss, UC, Us; isotypes); along Mexican Highway 190 in the Zinacantan paraje of Muctajoc, Municipio de Ixtapa, Breedlove 11839 (ps, MEXU); at junction of Mexican Highway 190 and road to Bochil & Simojovel, Municipio de Ixtapa, Breedlove 9600 (ps, F); ca. 30 mi. e. of Tuxtla Gutiérrez, Webster, Miller & Miller 11699 (DUKE, MICH, MEXU); 17 km. n. of Tuxtla Gutiérrez along road to El Sumidero, Municipio de Tuxtla Gutiérrez, Breedlove 14006 (ps, LL, MicH); along Mexico 190 at paraje Granadilla, Municipio de Zinacantan, Laughlin 1074 (ps, LL); in paraje ’Apas, Municipio de Zinacantan, Laughlin 1213 (ps, LL, MEXU); 2 mi. sw. of Aguacatenango junction on road to Las Rosas, Weaver 2183 (A, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, S, UC, Breedlove 6571 (ps, MIcH); between San Ricardo & Ocozucuantla, Nelson 2986 (us); along Mexican Highway 190, 7 mi. s. of La Trinitaria, Municipio de La Trinitaria, Breedlove 10038 (Ds, LL, MEXU). emala. HUEHUETENANGO: Finca Candelaria, Cook 65 (us); between San Ildefonso Ixtahuacan and Cuilco, Steyermark 50682 (¥F, US). Lisianthius nigrescens is by far the most common of the Mexican species, rather frequently encountered on roadside banks, especially in the state of Chiapas. With its black corollas, ovoid capsules, and sessile leaves with broad, clasping bases, L. nigrescens is easily distinguishable from most of the other species of Lisianthius. The only other species with black flowers is L. cuspidatus, which differs from L. nigrescens in that its leaves, though sessile and more or less clasping, are papillose above and on the margins and the bases are usually narrowed and not auriculate; its corollas are more deeply lobed; and its capsules are fusiform. L. oreopolus resembles L. nigrescens strongly in aspect, vegetative characters, and cap- sule shape, but differs in having uniformly yellow corollas, with the tube expanding more or less abruptly above the basal constriction, and the lobes strongly recurved. Williams (1968, 1969) treated L. cuspidatus Bertol. and L. oreopolus Robins. as varieties of L. nigrescens. He pointed out (1968) that var. cuspidatus differs from typical L. nigrescens in having larger and more deeply lobed corollas, but warned that, with more adequate collections, the varieties may eventually be shown to be the same. Williams (1968) considered var. oreopolus to be almost identical with var. cuspidatus except for its yellow corollas. In this treatment L. oreopolus, and L. cuspidatus, have been retained as full species. Williams apparently misinterpreted Lisianthius cuspidatus. He relied too heavily on the degree of corolla lobing as the primary character for separating it from L. nigrescens; he thought the differences in the leaf 248 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 bases were insignificant and was apparently unaware of the difference in capsule shape as well as the other characters listed above. Bertoloni described Lisianthius cuspidatus as having oblong capsules. However, Williams (1969) included in his var. cuspidatus plants with the oblong (or fusiform) capsule of L. cuspidatus and the broadly ovoid capsules of L. nigrescens. According to Williams (loc. cit.) his var. cuspidatus is distributed in the Mexican state of Chiapas and the Guate- malan departments of Huehuetenango, Guatemala, Chimaltenango, El Progresso, and Quiché. The plants from the last four Guatemalan de- partments are characterized by having fusiform capsules and therefore belong to L. cuspidatus. The Mexican plants and those from Huehue- tenango are characterized by having ovoid capsules and therefore belong to L. nigrescens. Williams’ var. cuspidatus is therefore a heterogeneous taxon, including elements of both L. nigrescens and L. cuspidatus. His figure of var. cus- pidatus (1969, p. 325, figure 88) represents both elements, although this fact is not particularly obvious from the drawing. Figure 88—A, drawn from Steyermark 50682 (¥F), represents L. nigrescens, and figure 88—C, drawn from Hawkes, Hjerting & Lester 1968 (F), represents L. cuspidatus. The specimens of Lisianthius nigrescens from Chiapas, Mexico, and Hue- huetenango, Guatemala, mostly included in L. nigrescens var. cuspidatus by Williams (1968, 1969), differ in a number of ways from typical L. nigrescens from the Mexican states of Veracruz, Hidalgo, and Oaxaca. The flowers are generally larger, the corolla lobes are longer in absolute length and longer in relation to the tube, and the styles only equal or very barely exceed the corolla lobes. Although there is a small amount of overlap in these critical characters, the two forms are readily distin- guishable in most cases. On the basis of available collections, it appears that their ranges are distinct. These two forms are not perfectly differ- entiated but the tendencies are strong indeed, and I think failure to draw a formal distinction between them would be taxonomically unsound. Therefore I have designated the large-flowered populations from Chiapas and Huehuetenango as a new variety of L. nigrescens. Since the new variety is nearly restricted to the state of Chiapas, I have named it L. nigrescens var. chiapensis. Hemsley (1882) and Perkins (1902) cited an Oersted specimen (s.n.), supposedly of Lisianthius nigrescens, from Pacaca, Costa Rica. Both authors included L. cuspidatus in their concept of L. nigrescens. I have not seen the Oersted specimen, nor any other specimen of a black-flowered species from Costa Rica, and I could not make even a reasonable guess as to its true identity, 5. Lisianthius oreopolus Robins. Proc. Am. Acad. 45: 398. 1910 (as Lisianthus). Type: Mexico. Cutapas: without exact locality, Ghiesbreght 702 bis (cu!). Lisianthus nigrescens var. oreopolus (Robins.) L.O. Wms. Fieldiana Bot. 31: 408. 1968. e 1972] WEAVER, REVISION OF LISIANTHIUS 249 Coarse, perennial herbs. Stems 1-2 m. tall and to 2 cm. in diameter, erect, green except at the base, terete below but indistinctly 4-angled above. Leaves firmly membranaceous, the lateral veins prominent, 2.5— 21.4 cm. long and 0.5—4.9 cm. broad, the principal foliage leaves sessile, lanceolate or more or less oblong, short-acuminate, the bases truncate and auriculate, broad, distinctly clasping, much broader than the stem; upper leaves and those on the secondary branches smaller and narrower, the bases narrower and less distinctly or not at all clasping. Inflorescences of long-pedunculate dichasia, these 1-4 times compound, the lateral divi- sions often scorpioid, or rarely appearing simple; inflorescence branches terete or indistinctly 4-angled, elongate and with the inflorescences loose and open. Flowers horizontal or nodding, the pedicels 3—9.5 mm. long; bracteoles subulate, scarious. Calyx 9-15 mm. long, fused for 1/8-1/4 its length; the lobes 6.5-12.5 mm. long and 1.5—2.5 mm. broad, lanceolate, long acuminate, scarious-margined, indistinctly carinate abaxially; the tube 1.5—3 mm. long. Corolla 4—6 cm. long, tubular-funnelform, uniformly bright yellow; the tube 2.74.1 cm. long and 5—9 mm. in diameter at the apex; the lobes 11-20 mm. long and 3-5 mm. broad, (1/3)1/2-2/3 as long as the tube, conspicuously recurved, lanceolate or oblong-lanceolate, long cuspidate-acuminate. Stamens inserted on the corolla tube ca. 1/3 of the distance from the base to the apex. Filaments 2.5—-5.1 cm. long, at least the longest ones surpassing the corolla lobes, very rarely equalling the style; anthers 2—3.5 mm. long and 1—1.5 mm. broad, broadly oblong, the apex acutish, the connective slightly exserted. Ovary ovoid; style 3.7-6.2 cm. long, exceeding the corolla lobes or rarely only equalling them; stigma peltate. Capsules broadly ovoid, short-beaked, 8-12 mm. long and 5—7 mm. broad; the marcescent corolla persisting until the seeds are shed. FLOWERING: May—Aug. CHROMOSOME NUMBER: 7” = 18. DIsTRIBUTION: From 1000 to 1800 meters in dry or rather moist pine or mixed forests or on roadside banks in eastern and southern Chiapas, Mexico (Map 3). REPRESENTATIVE COLLECTIONS Mexico. CHIAPAS: slopes along the Ala Shashib River below Habenal, paraje of Mahben Chauk, Municipio de Tenejapa, Breedlove 6477 (ps, F); in the paraje of Mahosik, Municipio de Tenejapa, Breedlove 14852 (Ds, DUKE) ; near schoolhouse of Pokolum, paraje of Sibanilhd, Municipio de Tenejapa, Breedlove 11015 (ps, F, LL, MICH); in the paraje of Kulak’tik, Municipio de Tenejapa, Breedlove 10996 (ps, F, LL, MICH); in the barrio of Ti Ha’, paraje of Mahbenchoak, Municipio de Tenejapa, Alush Shilom Ton 2267 (DUKE); near the Colonia Choro, Municipio de San Pedro Chendlho, Alush Shilom Ton 2457 (DUKE); 3.1 mi. n. of Bochil on road to Simojovel, Weaver 2174 (A, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, S, UC, US). Lisianthius oreopolus has a rather restricted distribution, and has until recently been poorly collected. With its numerous, large yellow flowers it is a very striking and handsome plant. With its long, recurved corolla 250 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 lobes, ovoid capsules, and sessile, clasping leaves, it is not likely to be confused with any other yellow-flowered member of the genus. In the original description of Lisianthius oreopolus, Robinson noted that the new species resembled L. nigrescens in habit but differed in its yellow corollas with longer and more spreading lobes. Williams (1968) reduced L. oreopolus to varietal status under L. nigrescens, stating that it is an almost exact duplicate of his L. nigrescens var. cuspidatus in corolla size and lobing. L. oreopolus is quite similar to L. nigrescens, especially var. chiapensis, but its flowers tend to be larger in all ways and the corolla is in general more deeply lobed. However, there is a complete overlap in all calyx and corolla measurements between the two species. The most striking difference between the two species is the color of the corolla. In addition the corolla tube of L. oreopolus is more or less abruptly ex- panded above the basal constriction and the lobes are conspicuously re- curved, while the corolla tube of L. nigrescens is gradually expanded above the basal constriction and the lobes are merely spreading. It seems possible that the breeding systems of the two taxa are quite different. The difference in corolla color is very striking and very prob- ably has a marked effect on the type of pollinator attracted to the re- spective taxa. In addition, the anthers of L. oreopolus are longer exserted than those of L. nigrescens. The corolla lobes of L. nigrescens are merely spreading and the anthers are generally held within them. Access to the throat of the corolla is blocked by the anthers, and a bee seeking nectar would certainly brush against them in the process. The corolla lobes of L. oreopolus are recurved and the filaments project rather far from the throat of the corolla; access to the throat of the corolla is not blocked by the anthers. Unfortunately, nothing is known about the pollinators of these plants. At any rate, Lisianthius oreopolus does differ consistently from L. nigrescens, even though the most striking difference, the color of the corolla, may be due to only a single gene difference. Until more is known of the biology of these plants, particularly their pollination systems and crossability, I prefer to treat them as distinct species. 6. Lisianthius silenifolius (Griseb.) Urb. Symb. Antill. 3: 334, 1902 (as Lisianthus). Type. Western Cuba, Wright, presumably at Gottingen (GoOET) but not seen. Leianthus silenifolius Griseb. Cat. Pl. Cub. 180. 1866. Rather coarse, annual (?) herbs. Stems 0.5—1 m. tall and to 1 cm. in diameter, erect, green above and gray-brown below, terete. Leaves mem- branaceous, the lateral veins prominent; the principal foliage leaves ses- sile, 3.0-13.3 cm. long and 1.2—-4.5 cm. broad, oblong-lanceolate to lan- ceolate or more rarely elliptic, gradually acuminate at the apex, the base rounded or truncate, more or less auriculate, clasping; upper leaves and those on the secondary branches smaller, lanceolate or ovate-lanceolate, rounded or cuneate at the base, subsessile or short-petiolate, not clasping. 1972] WEAVER, REVISION OF LISIANTHIUS 251 Inflorescences of pedunculate dichasia, these typically once-compound, but rarely 2-4 times compound, and then the lateral branches often scorpioid, or occasionally apparently simple; inflorescence branches terete or indistinctly angled. Flowers nodding, the pedicels relatively stout, 3-10 mm. long; bracteoles subulate, scarious. Calyx 6-11.5 mm. long, fused for 1/8-1/5 its length; the lobes 5-10 mm. long and 1.5—2.5 mm, broad, lanceolate, acuminate, scarious-margined, ridged abaxially; the tube 1-1.5 mm. long. Corolla 2.7-4.0 cm. long, tubular-funnelform, uni- formly bright yellow; the tube 2.2—3.1 cm. long and 4-6 mm. in diameter at the apex; the lobes 5.5—9 mm. long and 2-3 mm. broad, 2/10-3/10 as long as the tube, spreading, ovate-lanceolate, short-acuminate. Stamens inserted on the corolla tube ca. 1/3 of the distance from the base to the apex; filaments 1.5—2.7 cm. long, exceeding the corolla tube, the longest ones usually equalling or nearly equalling the style; anthers 1-2 mm. long and 1-1.5 mm. broad, broadly oblong, almost square, truncated at the apex. Ovary ovoid; style 1.8-3.0 cm. long, surpassing the corolla tube; stigma peltate. Capsule ovoid, usually broadly so, short-beaked, 5-11 mm. long, not including the 1 mm. long beak, and 3-6 mm. in diameter; calyx in fruit usually equalling or surpassing the capsule, rarely only 2/3 as long; marcescent corolla usually but not always persisting until the seeds are shed. FLOWERING: Jan.—Oct. DISTRIBUTION: On limestone or other soils, in pinelands, grassy slopes, or on cliff faces in western Cuba (Pinar del Rio) (Map 8). REPRESENTATIVE COLLECTIONS Cuba: without exact locality, Wright s.n. (GH; possibly an isotype). PINAR DEL Rio: Bahia Honda, Loma de Cajalbana, Ekman 12722 (s); pinelands, Cajalbana, Alain & Acuia 1106 (NY, US); base of Sierra Guane, Shafer 10548 (Ny); San José de Sagua to San Marcos, Shafer 11967 (Nv, US); vicinity of Sumidero, Shafer 13389 (BM, F, US, NY); Loma Pelada de Buenavista, Caya- jabos, Leén 13204 (cH); Sierra de los Organos, hills around Arr. Rosario, w. of Cayajabos, Ekman 12970 (s). Lisianthius silenifolius is the only member of subsect. HerBacer found in the Greater Antilles. It is easily distinguishable from L. glandulosus (sect. LisIaNTHIUS subsect. Fructicost), the only other Cuban species, by its sessile, clasping leaves, its terminal inflorescences, and its ovoid capsules. In addition, L. silenifolius is restricted to Pinar del Rio Province in western Cuba, while L. glandulosus is found only in Oriente Province on the eastern end of the island. : ae No unquestionably authentic material of Lisianthius silenifolius has been seen. However, the characters described in the original diagnosis leave no doubt as to the plant Grisebach had at hand. In addition, no other species is found in western Cuba. A Wright specimen of L. sileni- folius, collected between 1860 and 1864, is present in the collections at the Gray Herbarium. This specimen, which I have seen, is a possible iso- type. 252 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 7. Lisianthius qyichensis Donn. Sm. Bot. Gaz. 52: 51. 1911 (as Lisianthus).” Type. Guatemala. Quicnt: Rio Negro, Heyde & Lux 2921 (us!). Coarse, perennial herbs. Stems 0.8—2 m. tall and to 1 cm. in diameter, erect or ascending, green, terete or nearly so. Leaves firmly membrana- ceous, the lateral veins prominent; principal foliage leaves sessile, (4.2) 6.8-21.5(24) cm. long and (1.2)2.0-5.7 cm. broad, broadest at or above the middle, narrowly elliptic or somewhat oblanceolate, long-attenuate and narrow at the base; upper leaves and those on the secondary branches smaller, 1.2-10.5 cm. long and (0.6)1.5-3.3 cm. broad, lanceolate, cuneate, short-petiolate. Inflorescences of long-pedunculate dichasia, these 1 to 3 times compound and the lateral branches often scorpioid, or occasionally appearing simple; inflorescence branches terete, elongate and the inflo- rescences loose and open. Flowers nodding or horizontal, the pedicels 2-13 mm. long; bracteoles subulate, scarious. Calyx (6)7.5-13 mm. long, fused for ca. 1/5 of its length; the lobes (4.5)6—-11 mm. long and 1.5—2.5 mm. broad, lanceolate or subulate-lanceolate, long-acuminate, scarious-mar- gined, ridged or indistinctly carinate; the tube 1.5—-3 mm. long. Corolla (2.6)3-4.2 cm. long, tubular, uniformly yellow; the tube (2.3)2.7-3.5 cm. long and 3-7 mm. in diameter at the broadest point, somewhat inflated in the midparts and constricted slightly at the apex; the lobes 3—6(8) mm. long and 1.5-3 mm. broad, more than 2-times as long as broad, 1/8-1/4 as long as the tube, erect, not at all spreading, ovate- lanceolate to almost triangular, gradually acuminate. Stamens inserted on the corolla tube 1/3-2/5 of the distance from the base to the apex; filaments 1.83.4 cm. long, at least the longest and frequently all exceeding the corolla lobes, the longest one usually equalling or slightly exceeding the style; anthers 1.5-2.5 mm. long and 1-1.5 mm. broad, oblong, truncate or rounded at the apex, the connective sometimes very slightly exserted. Ovary ovoid; style 2.1-4 cm. long, always surpassing the corolla lobes, often conspicuously; stigma peltate. Capsule broadly ovoid, 6-12 mm. long and 3.5-6 mm. broad; calyx in fruit equalling or surpassing the capsule; the marcescent corolla persisting until long after the seeds are shed. FLOWERING: Mar.Sept. CHromosomE NUMBER: ” = 18. DIsTRIBUTION: From 900 to 1600 meters on open, rocky slopes, in secondary scrub, and on roadside banks in eastern Chiapas, Mexico, and north central Guatemala (Map 3). REPRESENTATIVE COLLECTIONS exico, CHIAPAS: Los Lagos, 3 mi. nw. of Rancho San José (34 mi. se. of 25-30 km. se. of Monte Libano, Dressler 1626 (BM, F, GH, MEXU, MICH, NY, Uc, us); San Carlos 4 Santa Rita, Miranda 7130 bis (MEXU). Guatemala: Atta Verapaz: along Quiché Highway about 12 km. w. of San Cristébal, Standley 89711 (¥); 7 mi. w. of San Cristébal Verapaz on Highway 7W, Weaver 1972] WEAVER, REVISION OF LISIANTHIUS 253 2177 (A, BM, DUKE, F, LL, MICH, MO, NY, US); 3.5 mi. w. es San te a Vannes on Highway IW, Weaver 2178 (A, BM, DS, DUK , LL, MICH, MO, NY, S, UC, US). HUEHUETENANGO: between Las Palmas mie Chaculé ariel de los Cuchumatanes, Steyermark 51743 (8); vicinity of Maxbal, ca. 17 mi. n. of Barillas, Sierra de los Cuchumatanes, Steyermark 48695 (F). QuIcHE: Rio Negro, Heyde & Lux 292(cu, MO, NY, US; isotypes). Lisianthius quichensis is a rather uncommon plant and until recently has been infrequently collected. Its closest relatives are probably L. nigrescens and L. oreopolus. It differs from both of these species by having leaves which are narrowed at the base and not clasping. It differs further from L. nigrescens in its yellow corollas, inflated corolla tube, and gen- erally shorter corolla lobes. From L. oreopolus, L. quichensis differs also in its much shorter, erect corolla lobes. Lisianthius quichensis has frequently been confused with L. sapona- rioides Cham. & Schlecht., a very dissimilar plant. Perkins (1902) referred to L. saponarioides the collection (Heyde & Lux 2921) later designated as the type of L. quichensis, a fact noted by Williams (1969), and dis- cussed in this treatment under L. saponarioides. 8. Lisianthius brevidentatus (Hemsl.) O. Ktze. Rev. Gen. Pl. 2: 429. Coarse, perennial herbs. Stems 0.5—1.5 m. tall and to 1 cm. in diameter, terete below but 3—4-angled or -winged above, green. Leaves sessile, clasping, 3.2—-20.4 cm. long and 1.2—5.1 cm. broad, the principal foliage leaves narrowly oblong-elliptic with a tendency to be widest slightly above the middle, broadly attenuate at the base, the apices acuminate or cus- pidate-acuminate, the upper leaves and those on the lateral branches lanceolate to ovate or more rarely elliptic, rounded or subcordate at the base, sessile but not clasping, or short petiolate. Inflorescences usually of dichasia, these 1-4 times compound, borne singly or in groups of 3 or more rarely 5, with the terminal one often sessile or nearly so and the whole unit appearing pleiochasial, the ultimate branches monochasial and resembling secund racemes; inflorescence branches short and the dichasia or groups of dichasia consequently dense and compact, or elongate and the dichasia loose and open, terete or indistinctly angled or more com- monly 2-3-winged, if 3-winged, 2 of the wings decidedly broader than the third, the wings to 1 mm. broad. Flowers sessile or the pedicels to 3 mm. long, the bracteoles scarious, subulate, 1-3 mm. long and 0.5-1 mm. broad. Calyx 6-9.5 mm. long, fused for 1/5-1/3 its length; the lobes 4.5-6 mm. long and 1-2 mm. broad, lanceolate, long acuminate, ridged abaxially, scarious-margined; the tube 1.5-3 mm. long. Corolla tubular, somewhat fleshy, 2.4-3.4 cm. long; the tube yellow, 2.2-3.0 cm. long and 3-4 mm. broad, conspicuously white-striate at base, especially with age; the lobes 2-4 mm. long and 2-3 mm. broad, greenish-yellow, spreading or slightly recurved at the apex, squarish to suborbicular or ovate, abruptly short-acuminate, auriculate at the base or not. Stamens 254 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 inserted on the corolla tube just below the middle, well above the apex of the calyx lobes; filaments 1.4-1.7 cm. long, equalling or barely ex- ceeding the corolla tube; anthers 2-3.5 mm. long and ca. 1 mm. broad, narrowly oblong, 2-3 times as long as broad. Style 2.3-3.3 cm. long, surpassing the corolla lobes in mature flowers, always exceeding the stamens in relative position; stigma peltate. Capsule oblong-ellipsoid, 8-13 mm. long and 3-5 mm. in diameter; the persistent calyx 2/3 to nearly as long as the capsule; the marcescent corolla strongly persisting. KEY TO THE VARIETIES 8a. Lisianthius brevidentatus (Hemsl.) O. Ktze. var. brevidentatus. Leianthus brevidentatus Hemsl. Biol. Centr. Am. Bot. 2: 344. 1382. Type, Guatemala: Sierra del Mico, Bernoulli 942, photograph (pUKE!), holotype K). Lisianthus elatus Standl. & Steyerm. Fieldiana Bot. 22: 267, 268. 1940YT YPE. Guatemala. IzaBaL: between milla 49.5 and ridge 6 mi. from Izabal, Mon- tafia del Mico, Steyermark 38562 (¥!). Inflorescence branches contracted and the dichasia or groups of dichasia dense and compact. Calyx 6-9.5 mm. long, fused for 1/4—1/3 its length, the lobes 4.5-6 mm. long and 1-2 mm. broad, the tube 1.5-3 mm. long. Corolla lobes squarish to suborbicular, about as broad as long, 2-3 mm. long and 2-3 mm. broad, auriculate at the base. FLowrrinc: Feb—Aug. CHROMOSOME NUMBER: n = DistriBuTion: In pine forests, open rocky slopes, and along roadsides from near sea level to 1000 meters in the Departments of Alta Verapaz and Izabal, Guatemala (Map 3). SPECIMENS EXAMINED Guatemala. ALTA VeRAPAz: on Guat. 4, 45 km. ne. of Coban, Stone 2765 (A, BM, Ds, DUKE, F, IJ, LL, MICH, MO, NY, s, UC, US); vicinity of caves, sw. of Lanquin, Steyermark 44140 (F, US); sobre el paredén del camino entre San Pedro Carcha y Sacouyou, Molina & Molina 12115 (F). IzaBaL: 0.8 mi. from Highway CA-9 on road to Mariscos, Weaver 2184 (A, BM, DUKE, F, LL, MICH, MO, s, US); along Highway CA-9 near turnoff to Mariscos, Weaver 2186 (A, DUKE, F, MICH, Us); trail from Los Amates to Izabal, Blake 7795 (cu, Us); 1972] WEAVER, REVISION OF LISIANTHIUS 255 8b. Lisianthius brevidentatus (Hemsl.) O. Ktze. var. collinus (Standl.) Weaver, comb. nov. Lisianthus collinus Standl. Carnegie Inst. Publ. 461: 81. 1935. Type. British Honduras: Jacinto Hills, Schipp 1205 (F!). Inflorescence branches elongate and the dichasia or groups of dichasia appearing loose and open. Calyx 6.5—7.5 mm. long, fused for ca. 1/4-1/3 its length, the lobes 5—6 mm. long and 1—1.5 mm. broad, the tube ca. 1.5 mm. long. Corolla lobes ovate, 2/3 mm. broad. DISTRIBUTION: Known only from the type locality on a cliff face in peaty soil at 300 ft. elevation (Map 3). SPECIMENS EXAMINED British Honduras. Totepo (?) Dist: Jacinto Hills, Schipp 1205 (GH, MICH, MO, S, UC, isotypes). Williams (1968, 1969) included under Lisianthius brevidentatus a rather diverse assemblage of plants, including, as I have interpreted them, L. quichensis Donn. Sm., L. calciphilus Stand]. & Steyerm., and a hybrid between L. axillaris and L. saponarioides (previously known as L. pete- nensis Standl. & Steyerm.). Williams (1968, p. 407) did note that there are differences, in the inflorescence, calyx, and corolla, in the material he placed under L, brevidentatus. He remarked, however, that these differ- ences are ones which may depend on the stage of growth of the plants, or that they are not easily defined. The following table compares Wil- liams’s treatment and mine of the same plants. WILLIAMs (1968, 1969) WEAVER L. brevidentatus (Hems!l.) L. brevidentatus (Hemsl.) O. Ktze. O. Ktze. var. brevidentatus (L. elatus Standl. & (L. elatus Standl. & Steyerm.) Steyerm.) (L. collinus Standl.) L. brevidentatus var. collinus (Standl.) Weaver (L. quichensis Donn. Sm.) L. quichensis Donn. Sm. (L. calciphilus Standl. & L. calciphilus Stand. & Steyerm.) Steyerm. (species non satis nota} (L. petenensis Standl, & A hybrid (no formal Steyerm. designation) Lisianthius quichensis, the range of which L. brevidentatus in central Guatemala, is rath than twice as long as broad, and long-acuminate; | 2 tinctly inflated in the middle; flowers distinctly “seacaaeed Sia branches angled but not winged; dichasia never sessile an st cies groups never appearing pleiochasial; capsules broadly ovoid; 256 JOURNAL OF THE ARNOLD ARBORETUM LvoL. 53 filaments, at least the longer ones, exceeding the corolla lobes; and leaves narrowed at the base and not clasping. Lisianthius calciphilus, which is known from a single specimen from Dept. Alta Verapaz, Guatemala, also differs strikingly from L. breviden- tatus in a number of characters: much longer corolla lobes, nearly twice as long as broad, gradually short acuminate; the corolla distinctly funnel- form; the flowers distinctly pedicellate; and the bracteoles conspicuous, oblanceolate, to 8 mm. long, and green at least in part. This plant is here treated as a species non satis nota. Lisianthius elatus Standl. & Steyerm., known only from the holotype, Steyermark 38562 (¥) from Dept. Izabal, Guatemala, is synonymous with L, brevidentatus. With the original description of L. elatus, Standley and Steyermark speculate that the most closely related species is L. collinus Stand]. and note that L. collinus differs in having “a lax in- florescence with scattered flowers, some of which are borne on greatly elongate pedicels” (measurements of all available specimens of L. collinus, however, show that the pedicels are never more than 3 mm. long, hardly elongate). The holotype of Lisianthius brevidentatus, Bernoulli 924 (Kk) from the Sierra del Mico, Guatemala, was not available for study, but I have seen a photograph of it. There is a specimen of Bernoulli 924 at the New York Botanical Garden, and it undoubtedly represents the same taxon as the holotype. Therefore I have used this specimen as typical of L. brevi- dentatus for comparison with L. elatus. Even a cursory examination leaves no doubt that the holotype of L. elatus is an entirely typical specimen of L. brevidentatus var. brevidentatus. Actually, it is possible that at the time they described L. elatus, Standley and Steyermark were unfamiliar with typical L. brevidentatus. Indeed in 1940 there were no more than three specimens of L. brevidentatus in American herbaria, and none were at the Field Museum. Lisianthius collinus Standl. is known from a single collection, Schipp 1205, from the Jacinto Hills in southern British Honduras. These speci- mens resemble L. brevidentatus in all respects except that the inflorescences are loose and open and the corolla lobes are slightly longer in relation to their width, distinctly ovate, and without auricles at the base. Standley, in describing it, compared L. collinus only with L. axillaris. However, it is undoubtedly closely related to L. brevidentatus, but certainly distinct. In view of its apparent distinctness and the fact that typical L. breviden- tatus is absent from British Honduras, but with an eye to caution because of the extremely limited number of specimens available, I prefer to treat L. collinus as a variety of L. brevidentatus, rather than to submerge it entirely as did Williams or to treat it as a distinct species as did Standley. Perhaps more adequate collections will necessitate a change in its status. 9. Lisianthius viscidiflorus Robins. Proc. Am. Acad. 45: 398. 1910 (as Lisianthus).7T yer, Guatemala. ALTA VERAPAz: Coban, von Tuerckheim I1-1308 (cu!). 1972} WEAVER, REVISION OF LISIANTHIUS 257 Coarse, perennial herbs. Stems 0.5—2.5 m., tall and to 1.5 cm. in diam- eter, erect, green and terete above, gray-brown and somewhat quadrate below. Leaves firmly membranaceous, sessile, clasping, (3.1)4.9-18 cm. long and 1.3—-7.2 cm. broad, oblong, oblong-elliptic, or tending toward oblanceolate, more rarely ovate, the apices short-acuminate, the bases broadly rounded or subcordate or rarely narrow and tapering; upper leaves and those on the lateral branches smaller, 1.3-7.2 cm. long and 0.3-2.5 cm. broad, narrowly ovate or lanceolate. Inflorescences of long- pedunculate dichasia, these 1-5 times compound or apparently simple, the lateral branches often scorpioid and secund. Flowers horizontal, weakly ascending or weakly nodding, the pedicels stout, 2-14(21) mm. long; bracteoles scarious, subulate. Calyx 4.5-8 mm. long, fused for 1/3—1/2 its length; the lobes 1.5—5 mm. long and 1.5—2.5 mm. broad, narrowly ovate to lanceolate and ovate, short-acuminate or more rarely acute, becoming somewhat blunt with age, scarious-margined, completely ecari- nate; the tube 1.5—-4 mm. long. Corolla tubular, glutinous, 2.9-3.9 cm. long; the tube 1.9-3.4 cm. long and 4-5 mm. in diameter at the apex, dull red; the lobes 3.5-5.5 mm. long and 2-3.5 mm. broad, 1/8-1/6 as long as the tube, spreading, greenish with a dull purple base, narrowly ovate, the apex blunt and somewhat undulate and erose, very rarely the apex acute. Stamens inserted on the corolla tube 1/3—2/5 of the distance from the base to the apex; filaments 1.5—2.9 cm. long, the shortest some- times barely surpassing the corolla tube, the longest ones always sur- passing the corolla lobes in mature flowers and frequently equalling or exceeding the style; anthers 2-3 mm. long and 1~1.5 mm. broad, oblong, the apex truncate or retuse. Style 1.8—-3.4 cm. long, always exceeding the corolla lobes in mature flowers; stigma peltate. Capsule narrowly ovoid or oblong-ellipsoid, lustrous, glutinous, short-beaked, 9-14 mm. long, not including the 1-2 mm. long beak, and 3-6 mm. in diameter; calyx in fruit 1/3-2/5(1/2) as long as the capsule; the marcescent corolla not persisting until the seeds are shed. FLOWERING: Mar.—Sept. CHROMOSOME NUMBER: DisTRIBUTION: From 150-1300 meters, along roadsides and in second- ary scrub in central Guatemala and Chiapas, Mexico (Map 4). SPECIMENS EXAMINED Mexico. Curtapas: Laguna Ocotal Grande, ca. 25-30 km. se. of Cerro Libano, Dressler 1448 (MEXU, MICH, NY, US). Guatemala. ALTA VERAPAZ: 3.5 mi. w. of San Cristébal Verapaz on Highway 7W, Weaver 2179 (A, DUKE); Coban, Johnson 718 (, US); 1.5-2 mi. s. of Gubilguitz, Steyermark 44431 (F, NY, us); Coban Road, between km. 284-285, between Chiracté and Chapultepec Farm, Contreras 4725 (LL); along road, between San Cristébal Verapaz and Chixoy, Steyermark 43885 (r); Cerro Chinajé, between F inca Yalpemech and Chinajé, above source of Rio San Diego, Steyermark 45672 (F); Coban, von Tuerckheim 1308 (¥, MO, Us; isotypes). QuicHE: 24.1 mi. e. of Cunén on Highway 7W, Weaver 2182 (A, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, S, UC, US). Although Lisianthius viscidiflorus is a rather poorly collected species, 258 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 it is quite common along roadsides in the vicinity of Coban, Dept. Alta Verapaz, Guatemala. The flowers are predominantly red but the plant is not particularly conspicuous even when in full flower. L. viscidiflorus is anthers. It is easily distinguishable by these characters and the red corolla tube from the other species of sect. LIstANTHIUS subsect. HERBACEI. Of these species it most closely resembles L. brevidentatus in corolla and leaf form. L. axillaris (sect. LIsIANTHIUS subsect. FruTICOosI), the only other species in the genus with red corolla tubes, differs strongly from L. viscidiflorus in having acuminate corolla lobes, yellow anthers, and solitary axillary flowers. As the specific epithet implies, and as pointed out by Robinson in the original description, the flowers of Lisianthius viscidiflorus are sticky, particularly L. longifolius, and the young buds and capsules of most of the species are viscid to some degree. Lisianthius viscidiflorus is a very distinct species, not particularly variable in most of its characters. However, the single known population from Chiapas, Mexico (Dressler 1448), differs from the Guatemalan populations in having acute rather than obtuse corolla lobes. This is the only consistent difference between the populations and is, I think, not of sine Pha amt to warrant the recognition of the two as formal varietie Lisianthius sect. Lisianthius subsect. Fruticosi Weaver, subsect. nov. Suffrutices, frutices vel arbores graciles, glabrae vel puberulae. Plantae ramosae, ramis indeterminatis. Folia petiolata vel raro subsessilia. In- . oe solitarii, axillares. Capsulae oblongo ellipsoideae. “Species Typica: L. longifolius L. KEY TO THE SERIES > Inflorescences not densely compacted, obviously dichasioid, or reduced to single flowers and subtended by foliaceous or scarious bracts, but these not involucrate. . Filaments and styles shortly exserted, the exserted a much less than the total length of the corolla; pollen grains reticulate. .......... ee ay ongifolit. B. Filaments and styles greatly exserted, the exserted portion ee ae equal to the total length of the corolla; pollen grains appearing smooth or nearly so with the light microscope. .................... b. Exserti. Inflorescences densely compacted, a appearing age or ave and sub- tended by 2 pairs of foliaceous involucral bracts. .......... . Umbellati. @ Lisianthius sect. Lisianthius subsect. Fruticosi ser. Longifolii eaver, ser. nov. 1972] WEAVER, REVISION OF LISIANTHIUS 259 Suffrutices vel frutices, omnino glabri vel puberuli. Inflorescentiae dichasiales, diffusae, bracteis foliaceis vel scariosis non involucratis sub- tentae, ramulis elongatis, flores raro solitarii. Lobi calycis saepe carinati vel alati. Styli filamentaque breviter exserta. Pollinis grana exinio reticu- lato.~Sprectes typica: L. longifolius L. With its 15 species, series Longifolii is the largest species group in the genus. Seven of the species are distributed in Mexico and Central America, with one species ranging barely into northwestern Colombia. The remainder are distributed throughout the Greater Antilles, with a majority in Jamaica (five species) and one species on each of the other three islands. The group is rather diverse morphologically, but all of the species are shrubs or subshrubs with inflorescences arranged diffusely on determinate, lateral flowering branches or rarely reduced to solitary, axillary flowers. KEY TO THE SPECIES OF SERIES LONGIFOLII A. ag tube red; flowers typically borne singly in the axils oe the foliage WONG. i ee ee . axillaris. A. soll tube yellow; flowers borne in dichasia, these wees be or on deter- minate, lateral flowering branches, very rarely the flowers borne singly in the axils of the foliage leaves. B. Corolla tube distinctly inflated, widest above the middle but constricted at the apex; corolla lobes and the apex of the tube often dark green; leaves coriaceous or Eig aay eae broadest at or above the middle. C. Corolla lobes 1.2-2.0 cm. long, ca. 1/2 as long as the tube; leaves coriaceous; corolla lobes ecu e calyx lobes completely ecarinate; corolla strongly persistent on the mature capsules; plants OF JAMA cs Vo eee. ee 3. L. adamsii. Corolla lobes 1 cm. or less in length, less than 1/4 as long as the tube; leaves coriaceous or membranaceous; corolla lobes greenish- yellow or dark green; calyx lobes carinate or at least ridged abax- ially; corolla not persisting on the mature capsules; plants of Cen- tral America. Stamens inserted well above the apex of the calyx lobes; calyx in fruit 1/5-2/5 as long as the capsule; corolla lobes less than 2 times as long as broad, usually less than 1/8 as long as the tube; calyx 5-8 mm. long, the lobes narrowly ovate and abruptly short-acuminate, more rarely ovate-lanceolate and POUIDET BCUUNSNRIG ay het ce sos 10. L. skinneri. Stamens inserted at or just above the apex of the calyx lobes; calyx in fruit 1/2 as long to nearly as long as the capsule; corolla lobes 2 times as long as broad or longer, more than 1/6 as long as the tube; calyx 7-13.5 mm. long, the lobes lanceolate and long-acuminate. E. Leaves membranaceous; corolla (4.2)4.7-5.9 cm. long, the lobes 7-10 mm. long, the tube 4-5 times as long as the calyx; styles barely surpassing the corolla lobes, the fila- ments never surpassing the corolla lobes. .............. .: Dl ee ee Ge OGRE Gn ee Nee ee Pe 11. L. peduncularis. E. Leaves more or less coriaceous; corolla 2.5—4.2 cm. long, O u 260 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 the lobes 4-6.5 mm. long, the tube less than 4 times as long as the calyx; the longer filaments and the styles surpassing the corolla lobes, the latter conspicuously 40. ...o.5 scx. PORTE Ba oe SP a ate SPCR eager COR ere L. jefensis. B. Corolla tube not inflated, widest at the apex; corolla lobes yellow or 1 ally broadest at or below the middle. F. Plants glabrous, or very rarely minutely spiculate; calyx lobes never alate; flowering branches very rarely with leaves above the first division and never on the secondary axes; corolla seldom persisting until the seeds are shed. G. Inflorescences congested, of compound dichasia, the lateral 0 branches scorpioid or helicoid; corolla paler inside than out; flowering branches with 3 to many pairs of leaves below the mWwermey Giving | 22. L. seemannii. Inflorescences diffuse, mostly of apparently simple dichasia, the lateral branches not scorpioid or helicoid ; corolla the same color inside as out; flowering branches rarely with more than 1 pair, and never more than 3 pairs of leaves below the lowermost division. H. Filaments, at least the longest ones in each flower, exceeding the corolla lobes. I. Corolla lobes ca. 1/2 as long as the tube, and acute or abruptly short-acuminate; leaves usually broadest above the middle, gray-green, thick and somewhat coriaceous, the lateral veins indistinct; plants of Jamaica. ....... Se ee es ea re 13. L. adamsii. I. Corolla lobes less than 1/3 as long as the tube and long cuspidate-acuminate; leaves broadest at or below the middle, gr in-membranaceous, the lateral veins prominent; plants of Central America. .............. H. Filaments not exceeding the corolla lobes. J. Corolla lobes 5-6 times as long as broad, more than 3/5 as long as the tube and gradually long-acuminate; calyx ca. 1/2 as long as the corolla tube. ............ ee a a es a 18. L. laxiflorus. J. Corolla lobes less than 4 times as long as broad, 1/2 or less as long as the tube, acute or abruptly short-acumi- nate, very rarely gradually acuminate; calyx 1/3 or less as long as the corolla tube. K. Corolla lobes less than 1/3 as long as the tube. L. Styles surpassing the corolla lobes in mature fl filaments 2.1-2.7 cm. long, inserted on the corolla tube ca. 1/3 of the distance from the L. Styles not surpassing the corolla lobes in mature flowers; bracteoles below the pedicels scarious; 1972] WEAVER, REVISION OF LISIANTHIUS 261 flowers horizontal or nodding; calyx fused for ee its length, stigma capitate; filaments . long, inserted on the corolla tube just sears the middle; ee : Cade es panes sie. eet ER eee . L. glandulosus. K. Corolla lobes more than 1/3 as i as the tube. M. Inflorescence branches terete or indistinctly angled, the angles never expanded into narrow, scarious wings; dichasia, at least the terminal ones, 2—3-flowered; leaves more than 2.5 times as long as broad, long-petiolate; plants of Ja- maica. N. ue tube tightly consisicted 3 in the lower oe OL, Corolla tube more or less gradually expand- ing from the base; styles never surpassing the corolla lobes; capsule long-beaked, the beak 1/3-1/2 as long as the capsule. ..... ee ead verannt stg Mame ge ae 19. L. latifolius. M. Inflorescence branches strongly angled, the angles often expanded into narrow, scarious wings; dichasia reduced to 1 or rarely 2 flowers; bracteoles below the pedicels green; leaves less than 2.5 times as long as broad, subsessile or petiolate; plants of Hispaniola. ............. 20. L. domingensis. F. Plants, especially the inflorescence, puberulous, if nearly glabrous, the calyx lobes broadly alate abaxially; flowering branches with leaves above the first division and often on the secondary axes; corolla persisting long after the seeds are shed O. o Calyx lobes alate or very rarely merely carinate; leaves on the flowering branches obtuse or rounded or very rarely subcordate at the base; branching opposite or alternate below and usually opposite above; inflorescences, especially those at the apex of the flowering branches, 2-3-flowered. ........ 3. L. longifolius. Calyx lobes carinate abaxially; leaves on the donee branches, at least the upper ones, distinctly cordate at the base; branching alternate or bifurcate, very rarely opposite; inflorescences 2- flowered or the flowers solitary. ........... 24. L. cordifolius. 10. Lisianthius skinneri (Hemsl.) O. Ktze. Rev. Gen. Pl. 2: 429. 1891 (as Skinneri). Leianthus skinneri Hemsl. Biol. Centr. Am. Bot. 2: 345. 1882.“Lectotypr. Guatemala: without exact locality, Skinner (K); photo. (mo!). Lisianthus arcuatus Perk. Bot. Jahrb. 31: 492. 1902.~Type. None designated; kg 9 (Costa Rica et Veragua) and Hoffmann 832 etic mae ted, both presumably at Berlin and destroyed during World W bias scopulinus Robyns & Elias, Ann. Mo. Bot. Gard. 55: ore "1968, 262 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 YType, Panama. VeRAGUAS: Mouth of the Rio Concepcién, Lewis, Croat & Hawker 2799 (mo). Shrubs or more rarely subshrubs, glabrous throughout. Stems 0.5—3 m. tall and to 2 cm. in diameter, erect or ascending, rarely sprawling, gray- brown below and green above, terete. Leaves membranaceous, glossy, dark or rather pale green above and distinctly paler beneath, the blades 3.5-20(24.5) cm. long and 1.5—5.5(7.7) cm. broad, narrowly elliptic to narrowly obovate (broadly elliptic), generally widest above the middle, abruptly acuminate, the bases long-attenuate into the winged petiole; the upper leaves essentially sessile or the petioles to 2 cm. long. Inflorescences of pedunculate, apparently simple (1-compound or rarely 3-compound) dichasia, occasionally reduced to 1 or 2 flowers; flowering branches ascend- ing, usually ternately divided, each division terminated by a dichasium or group of dichasia; axes of the flowering branches and the dichasial branches indistinctly 4-angled to essentially terete, elongate and the in- florescences appearing loose and open. Flowers horizontal or nodding, the pedicels 0.7—2.4 cm. long; bracteoles subulate, scarious. Calyx 4.5-8 mm, long, united for 3/10—-3/5 of its length, 1/10-1/6(1/5) as long as the corolla tube; the lobes 2-5 mm. long and 1.5—3 mm. broad, narrowly ovate (lanceolate, broadly ovate, or suborbicular), abruptly short-acumi- nate (long-acuminate or acute), scarious-margined, ridged or carinate abaxially; tube 1.5-3.5 mm. long. Corolla tubular, more or less fleshy or thinner, 2.76.3 cm. long; tube 2.4—5.7 cm. long and 0.4—1.1 cm. broad at the widest point, bright yellow but with a green or yellow-green apex, constricted in the lower 1/4—-1/3, inflated in the mid-portions; lobes 2.5-6 mm. long and 2.5-4 mm, broad, (1/6)1/8-1/12 as long as the tube, yellowish-green, or rather dark green with yellow green margins, spreading and with the tips often recurved, triangular-ovate and abruptly short-acuminate (ovate-lanceolate and rather long-acuminate). Stamens inserted in the lower third of the corolla tube; filaments 2~—5 cm. long, at least the longest ones equalling or surpassing the corolla lobes, the longest ones sometimes surpassing the style; anthers 1.5-3.5 mm. long and 1—1.5 mm. broad, broadly oblong, the connective slightly exserted. Style 2.3-5.3 cm. long, exserted, usually surpassing the corolla lobes; stigma peltate. Capsules narrowly oblong-ellipsoid, conspicuously beaked, 1-2.2 cm. long, not including the 1-6 mm. long beak, and 4—6 mm. in diameter; the marcescent corolla falling off before the seeds are shed. FLOWERING: Jan—Sept. CHROMOSOME NUMBER: 7 = 18, DIsTRIBUTION: From sea level to 1300 meters in moist forests, generally along the edge, or along roadsides, on the Atlantic Coastal Plain of southeastern Guatemala and adjacent Honduras and again in Panama, and at middle elevations in the interior of Costa Rica and Panama (Map 4). REPRESENTATIVE COLLECTIONS Guatemala. IzABaL: south shore of Lake Izabal, west of the village of Izabal, Proctor, Jones & Facey 3049 (¥, 13, LL, MEXU, Us). Honduras, ATLANTIDA: 1972] WEAVER, REVISION OF LISIANTHIUS 263 south of San Alejo near Rio San Alejo, Standley 7829 (F); Cuyamel, Carleton 584 (F). Costa Rica. ALAJUELA: Villa Quesada, Smith 1884 (F, MO, NY); Cata- ratas de San Ramon, Brenes 13575 (¥, NY); ca. 13.5 mi. e. of Arena ae 6.5 mi. w. of Fortuna, Wilbur & Stone 10257 (A, BM, DUKE, DS, F, IJ, H, MO, NY, S, UC, Us). GUANACASTE: Tilaran, Valerio 1151 (F). rein ‘Gaeenieul Isla Bonita, Pittier s.n. (Ny, Us); Finca Hnos. Vargas, Puerto Viejo de Sara- oe Pai roadsides 5-7 mi. se. of Portobello, Weaver & Wilbur 2249 (A, BM, Ds, F, IJ, LL, MICH, MO, NY, S$, UC, us); summit of Cerro Santa ia ‘Allen: 5104 (BM, Mo). PANAMA: slopes of Cerro Jefe, along road from Cerro Azul, ~ be- low the turnoff to La Ripa Weaver & Wilbur 2244 (A, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, S, Us); VERAGUAS: mouth of the Rio Concepcién, pons Croat & Hawker 2709) ee isotype of L. scopulinus). CANAL ZONE: forest along telephone cable trail between Splice S16 & S49, Rio Indio, Steyermark & Allen 17427 (BM, Ds, MO, s); headwaters of Rio Arenal, near IWTC pump, Johnston 1502 (A, MO). Lisianthius skinneri is the most widespread species, ranging from south- eastern Guatemala to central Panama; with L. jefensis and L. peduncularis it forms a rather distinctive, closely interrelated species group. These three species differ from one another in the length of the calyx, the shape and length of the corolla lobes, and the ratio between the lengths of the corolla lobes and the corolla tube. The differences are not striking, especially in view of the variation within L. skinneri, but they are con- stant. The close sympatry of L. skinneri with both of the other species and the absence of intermediate forms supports the maintenance of all three as distinct. Lisianthius skinneri is a rather variable species, but no population is sufficiently distinct to warrant recognition as a subspecies or variety. There is no perceptible pattern to the variation. The Guatemalan plants are characterized by long, narrow, thin-textured corollas; long-acuminate, relatively long and narrow calyx and corolla lobes; and styles usually longer than the filaments. Wilbur & Stone 10257, from Volcan Arenal, Costa Rica, is characterized by long, narrow, rather fleshy corollas; merely acute calyx lobes; broad and very short-acuminate corolla lobes; and styles shorter than the filaments in a given flower. Other specimens from Costa Rica and those from Honduras and the interior of Panama resemble the Guatemalan plants but have a broader corolla, with calyx and corolla lobes broader and shorter acuminate, and styles and stamens various. The plants from the Atlantic Coastal Plain of Panama have a short, rather slender, fleshy corolla; broad, short-acuminate calyx and corolla lobes; and the styles conspicuously longer than the filaments. Although no original material has been seen and duplicates have not been located, according to the original description L. arcuatus Perk. differs from L. skinneri in having arcuate rather than straight corollas and subelongate rather than short internodes. Internode length, especially in shrubs, depends largely on environmental factors and has proved of minimal taxonomic utility in this genus. Many dried specimens of L. 264 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 skinneri are indeed characterized by having flowers with arcuate corollas. Actually the corollas appear merely swollen on one surface and would more correctly be termed “gibbous” rather than “arcuate.” This condition Lisianthius scopulinus Robyns & Elias reputedly differs from L. skinneri in having narrowly ovate, long-acuminate, and longer calyx lobes. I have seen an isotype of L. scopulinus, and it strongly resembles typical L. skin- neri except that the corolla lobes are generally longer than usual. The longest calyx lobes (5 mm.) are indeed longer than any seen in specimens of L. skinneri. However, the shorter lobes (4 mm.) on this specimen are no longer than those, for instance, on Proctor, Jones & Facey 3049 (F, IJ, LL, MEXU, US), perfectly good L. skinneri, from Lake Izabal, Guatemala, geographically rather far removed from the type and only known locality of L. scopulinus at the mouth of the Rio Concepcion in Veraguas, Panama. In addition, L. skinneri is not infrequently charac- terized by having rather long-acuminate, narrowly ovate calyx lobes. Therefore, it seems that L. scopulinus does not really differ in any sig- nificant way from L. skinneri. When Hemsley (1882) described Leianthus skinneri, he cited two specimens as representative of the new taxon: a Skinner specimen, without a collection number (Guatemala, without exact locality) and Fendler 137 11. Lisianthius peduncularis L.O. Williams, Fieldiana Bot. 31: 408. 1968 (as Lisianthus).“Typr. Panama. Cocrk: El Valle de Anton, Allen 3410 (mo!). Slender shrubs or subshrubs, glabrous throughout. Stems 1—2.5 m. tall and to 2 cm. in diameter, erect or ascending, terete, gray-brown and woody below, green and herbaceous above. Leaves glossy grass-green above, paler and dull beneath, the lateral veins prominent, the blades (3.2)5.2-13.5(18.0) cm. long and (1.1)2.0-5.4(8.4) cm. broad, narrowly obovate or oblanceolate (elliptic), widest at or above the middle, abruptly short-acuminate (acute), the bases long-attenuate into the winged, 0.4—- 2.0(3.2) cm. long petiole. Inflorescences of pedunculate, apparently simple 1972] WEAVER, REVISION OF LISIANTHIUS 265 (1-compound) dichasia, these sometimes reduced to 1 or 2 flowers; flowering branches often with several pairs of reduced leaves below the first division, once- or rarely twice-ternately divided, each division ter- minated by a dichasium; dichasial branches and stalks 4-angled, elongate and with the inflorescences loose and open. Flowers nodding, the pedicels 6-14 mm. long; bracteoles subulate, scarious. Calyx 8—-13.5 mm. long, fused for 1/8—1/5(1/4) its length; the lobes 6.5-11.5 mm. long and 1-2.5 mm. broad, lanceolate, long-acuminate, scarious-margined, weakly carinate abaxially; the tube 1.5-3 mm. long. Corolla tubular (4.2)4.7-5.9 cm. long; the tube (3.4)4-4.9 cm. long and 5.5-9 mm. broad at the broadest point, bright yellow, constricted in the lower third, inflated in the mid- portions; the lobes (7)8-10 mm. long and 3-4 mm. broad, 1/5-1/4 as long as the tube, spreading and recurved at the tips, greenish-yellow or rather dark green, lanceolate or ovate-lanceolate, long-acuminate. Stamens inserted on the corolla tube ca. 1/4 of the distance from the base to the apex; filaments (2.9)3.5—4.2 cm. long, the shorter ones barely surpassing the corolla tube, the longest ones not surpassing the corolla lobes; anthers 2-3 mm. long and 1—1.5 mm. broad, yellow, narrowly oblong, acutish, the connective slightly exserted. Style (3.2)4.0-5.2 cm. long, usually slightly exceeding the corolla lobes, always surpassing the filaments; stigma peltate. Capsule fusiform, long-beaked, 1.1—-1.6 cm. long, not including the 2-3.5 mm. long beak, and 4-5 mm. in diameter; the marcescent corolla not persisting until the seeds are shed. FLOWERING: Jan.—_June. CHROMOSOME NUMBER: n = 18. DistripuTion: Along roadsides and in dense moist forests at 1000 meters, just north of El Valle de Antén in central Panama (Map 5). SPECIMENS EXAMINED Panama. Cocteé: El Valle de Anton, Allen 3410 (om, MO; isotypes); El Valle, north rim, Allen 1793 (GH, Mo, NY); inside the crater of El Valle de Antén at La Mesa, Weaver, Wilbur & Correa 2247 (A, DUKE, F, LL, MICH, MO, NY, US). Lisianthius peduncularis is known only from the type locality where it is a rather rare plant. Although the species was only described in 1968, specimens of L. peduncularis have been present in collections since 1939. All of the known specimens had been determined as “L. Jatifolius Sw.”, a Jamaican species resembling L. peduncularis primarily in its large, broad leaves. The affinities of Lisianthius peduncularis are clearly with L. skinneri and L. jefensis. From the former it may be distinguished by its longer calyx and corolla lobes, its calyx fused for 1/4 its length or less, its corolla tube no more than 5 times as long as the calyx, its corolla lobes more than twice as long as broad and longer in relation to the tube, and its stamens inserted at or above the apex of the calyx lobes. L. peduncularis may be distinguished from L. jefensis by its membranaceous leaves, its longer corolla, its longer corolla lobes, and its filaments and styles not, or only barely exceeding the corolla lobes. 266 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 I have collected Lisianthius peduncularis at El Valle, where it certainly is not common. Only a single small population was found. It is entirely possible that all the known specimens are from this population, and the population is merely an aberrant one of L. skinneri. However the presence of a specimen of L. skinneri from El Valle [Allen 2369 (mo)], with no specimens intermediate toward L. peduncularis known from the area suggests otherwise. Although I do so with some reservation, I prefer to treat L. peduncularis as a species distinct from L. skinneri. Perhaps more adequate collections in the future will lead to a reconsideration. 12. Lisianthius jefensis Robyns & Elias, Ann. Mo. Bot. Gard. 55: 60. 1968 (as Listanthus).VTypr. Panama. Panama: Cerro Jefe, Elias & Hayden 1798 (mo). so shrubs, glabrous throughout. Stems 1.3—2.5 mm. tall and to diameter, erect or the secondary branches ascending, terete, Santen below sad green above. Leaves glossy dark green above, paler and dull below, the lateral nerves prominent, the blades (1.5)2.7-11.8 (20.6) cm. long and 0.8—5.0(7.7) cm. broad, narrowly obovate to oblan- ceolate and elliptic, widest at or above the middle, abruptly acuminate, the bases long-attenuate into the winged petiole. Inflorescences of pedun- culate dichasia, these apparently simple (1-compound) but sometimes reduced to 1 or 2 flowers; flowering branches ascending, the axes terete or slightly flattened, once- or more rarely twice-ternately divided, each division terminated by a dichasium; branches and the stalks of the dichasia indistinctly 4-angled, elongate and the inflorescences loose and open. Flowers horizontal or nodding, the pedicels 0.7-2.5 cm. long; bracteoles subulate, scarious. Calyx 7-13 mm. long, fused for 1/3-2/5 its length; lobes 4.5-9 mm. long and 1.5-3 mm. broad, lanceolate to ovate-lanceolate, long-acuminate, scarious-margined, ridged or weakly carinate abaxially; tube 3-4 mm. long. Corolla tubular, 2.5-4.2 cm long; tube 2.1-3.5 cm. long and 5-8 mm. broad at the broadest point, bright yellow with a greenish-yellow or dark green apex, constricted in the lower third, inflated in the midparts; lobes 4-6.5 mm. long and 1.5-3.5 mm. broad, 1/5-1/6 as long as the tube, spreading and recurved at the tips, greenish-yellow or dark green with paler margins, triangular-ovate ovate-lanceolate, rather long acuminate. Stamens inserted on the corolla tube ca. 2/5 of the distance of the base to the apex; filaments 1.3-3 cm. long, exserted, the longest ones surpassing the corolla lobes; anthers 2.5-3.5 mm. long and 1—1.5 mm. broad, yellow, narrowly oblong, acutish, the connective slightly exserted. Style 2.2-3.6 cm. long, con- spicuously surpassing the corolla lobes, always surpassing the filaments; stigma peltate. Capsule fusiform, conspicuously beaked, 1—1.8 cm. long, not including the 1.54.5 mm. long beak; the marcescent corolla not per- sisting until the seeds are shed. FLOWERING: Jan.Sept. CHROMOSOME NUMBER: n = 18 1972] WEAVER, REVISION OF LISIANTHIUS 267 DistTRIBUTION: From 850-900 meters in cloud forest near the summit of Cerro Jefe on the Continental Divide in central Panama (Map 5). REPRESENTATIVE COLLECTIONS Panama. PANAMa: slopes of Cerro Jefe on the road to La Eneida, Weaver & Wilbur 2250 (A, BM, ps, DUKE, F, IJ, LL, MO, NY, S, UC, Us); slopes of Cerro Jefe, along road from Cerro Azul, just before the turnoff to La Eneida, Weaver & Wilbur 2243 (a, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, S, UC, US); Cerro Jefe, Elias & Hayden 1798(us; isotype). Lisianthius jefensis is known from only one, low shrubby mountaintop in central Panama where it is one of the more conspicuous elements of the vegetation. It is closely related to both L. skinneri and L. pedun- cularis. It differs from L. skinneri in its coriaceous leaves, its longer calyx lobes, its corolla lobes more than twice as long as broad and longer in The apparent distinctness of the populations on Cerro Jefe, in spite of their close sympatry, strongly supports the separation of the two as distinct species. 13. Lisianthius adamsii Weaver, Brittonia 22: 11. 1970. Type. amaica. St. ANN: Douglas Castle, ca. 2 mi. n. of the Mason River Savannah, Weaver 1293 (puxKE!). Spindly shrubs or subshrubs, glabrous throughout. Stems erect, 0.2-2 m. tall and to 1.5 cm. in diameter, terete, green and herbaceous above, gtay-brown and woody below, leafy nearly to the base. Leaves dull gray- green, thickish, the venation indistinct, the blades (5.8)8—-15( 17.6) cm. long and ( 1.9)2.5-5.5(6.6) cm. broad, broadly oblanceolate to obovate (elliptic), abruptly short-acuminate (acute), the bases attenuate into the winged, 0.5-2 cm. long petiole. Inflorescences of pedunculate, apparently simple dichasia, these occasionally reduced to 1 or 2 flowers; flowering branches 1-—2-ternately divided, each division terminated by a dichasium or group of dichasia; inflorescence branches and stalks elongate and the inflorescences loose and open. Flowers nodding, the pedicels 0.8-4.5 cm. long; bracteoles scarious, subulate. Calyx 7-13 mm. long, fused for 1/6-1/3 its length; the lobes 5-9 mm. long and 1.5-2.5 mm. broad, lanceolate, acuminate, scarious-margined, ecarinate; the tube 1.5-4 mm. long. Corolla tubular-funnelform, 3.9-5.6 cm. long; the tube 2.7-3.6 cm. long and 8-11 mm. in diameter at the broadest point, constricted in the lower third, somewhat inflated, frequently unevenly, in the midparts, 08 = 4 eh p Maps 4-7. Distribution of Lisianthius species. Map 5 fet * ~~ ~ as A. oA, aoe 4, L. skinneri (dots), L. viscidiflorus (triangles). Map 5, L. auratus (dots), L. jefensis (half-circle), L. peduncularis (triangle). Map 6, L. acuminatus (triangle), L. axillaris (dots). Map 7, L. seemannii. 892 WOLAXOdUVY GIONUYV AHL AO TYNANOL S$ “I0A] 1972] WEAVER, REVISION OF LISIANTHIUS 269 glossy clear yellow; the lobes 1.2-2 cm. long and 5-7 mm. broad, greenish-yellow, spreading and conspicuously recurved, ovate-lanceolate, abruptly short-acuminate. Stamens inserted on the corolla tube ca. 1/3 of the distance from the base to the apex; filaments 3—4 cm. long, exserted, at least the longest ones conspicuously surpassing the corolla lobes; anthers 2-4 mm. long and 1-2 mm. broad, oblong, acute, the connective exserted. Style 3.5-5.5 cm. long, surpassing the filaments; stigma peltate. Capsule oblong-ellipsoid, 14.5-16.5 mm. long and 5—7 mm. in diameter; the marcescent corolla persisting until the seeds are shed. FLOWERING: Feb.—Oct. CHROMOSOME NUMBER: 7” = 1 DISTRIBUTION: From 500-750 meters in thickets and on rocky, lime- stone hillside in the central and western parishes of Jamaica (Map 12). SPECIMENS EXAMINED Jamaica. CLARENDON: upper w. slope of Croft’s Mountain, Proctor 29234 (13). HANoveER: interior summit slopes of Dolphin Head, Proctor 10037 (13). MANCHESTER: Somerset Woods, ca. 5 mi. nw. of Mandeville, Proctor 16182 (a, BM, IJ). St. ANN: Douglas Castle, ca. 2 mi. n. of the Mason River Savan- nah, Weaver 1293 (GH, UcwI; isotypes). Lisianthius adamsii, the most recently described species in the genus, is one of the more distinct and attractive of the species. With its coria- ceous, gray-green, oblanceolate or obovate leaves; rather broad, inflated corolla tube and long recurved corolla lobes; and ecarinate calyx lobes, it is not easily confused with any other. Vegetatively Lisianthius adamsii resembles somewhat L. umbellatus and L. capitatus (Series Umbellati), also with thickish, oblanceolate leaves; but it differs strikingly from them in its loose, open inflorescences and much longer corolla lobes. Although the relationships of L. adamsti are clearly with Series Longifolii, its specific relationships within the group are obscure. Its closest relatives are possibly the Jamaican L. /ati- folius, or, more probably, the Central American L. skinneri. 14. Lisianthius acuminatus Perk. Bot. Jahrb. 31: 493. 1902 (as Lisianthus). “Lectotype. Mexico, without exact locality, Sumi- chrast 1558 (us!; a fragment). Slender subshrubs, glabrous throughout. Lower portions of the stems unknown, the upper portions green, terete or indistinctly 4-angled. Leaves membranaceous, papillose above, the lateral veins prominent, the margins papillose, the blades 7.7-19.6 cm. long and 2.5-5.6 cm. broad, elliptic, abruptly long-acuminate, shortly attenuate and tapering into the winged petiole (to 4 mm. long). Inflorescences of pedunculate, apparently simple (1-compound) dichasia; flowering branches 1—2-ternately divided, each division terminated by a dichasium; axes of the flowering branches and the stalks and branches of the dichasia 4-angled, elongate and with the inflorescences diffuse; bracts subtending the dichasia green, linear. Flowers ascending, the pedicels 3-7 mm. long; bracteoles linear, scarious-margined 270 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 but usually green along the midline. Calyx 6.5~7.5 mm. long, fused for 1/5—1/3 its length; the lobes 4-6 mm. long and 1.5-2 mm. broad, lan- ceolate, acuminate, scarious-margined, distinctly carinate; the tube 1.5—2.5 mm. long. Corolla tubular, 3.4~-3.7 cm. long; the tube 2.9-3.2 cm. long, yellow; constricted in the lower third, the lobes 4.5-5.5 mm. long and 2.5-3 mm. broad, 1/8-1/6 as long as the tube, slightly spreading, green- ish-yellow, ovate, acute or abruptly short-acuminate. Stamens inserted on the corolla tube ca. 1/2 of the distance from the base to the apex; filaments 2.1-2.7 cm. long, surpassing the corolla tube but not the lobes; anthers ca. 2.5 mm. long and 1.5 mm. broad, broadly oblong, the con- nective very barely exserted. Styles ca. 3 cm. long, exceeding the corolla lobes, always exceeding the filaments; stigmas peltate. Capsule not seen. DISTRIBUTION: Near Orizaba in south-central Mexico (Map 6). SPECIMENS EXAMINED Mexico: without exact locality, Sumichrast 1558 “x: lectotype collection), photograph (DUKE). VERACRUZ: Orizaba, Botteri (cH); Monte San Cristobal, Orizaba, Mohr & Botteri (us). Lisianthius acuminatus is one of the most poorly understood and rarely collected of the species. Only 2 specimens, in addition to a fragment of the lectotype and a photograph of an isolectotype, the most recent collected in 1866, were available for study. L. acuminatus appears to be most closely related to the Jamaican L. latifolius or perhaps to the alliance of West Indian species that includes L. troyanus, L. laxiflorus, and L. glan- dulosus. Of the Mexican and Central American species L. acuminatus might possibly be confused with L. quichensis in subsect. HERBACEI. How- ever that plant is a suffrutescent herb with a determinate main axis. In addition, the corolla tube of L. guichensis is somewhat inflated and broad- est below the apex, the corolla lobes are erect, and the filaments exceed the corolla lobes. In the original description of Lisianthius acuminatus, Perkins cited two specimens as representative of the new taxon. These specimens (Sumichrast 1558 and 1856), which must be treated as syntypes, were presumably at Berlin and destroyed during World War II. There is however a dupli- cate of Sumichrast 1558 at Kew, a photograph of which (DUKE) I have seen. In addition Perkins sent a fragment of Sumichrast 1558 to John Donnell Smith according to a letter from her dated 31 March 1910 and presently attached to Heyde & Lux 2921 (us), the holotype of L. qui- chensis. This fragment includes a leaf, a bud, and the upper portion of the corolla of a nearly mature flower. The fragment, along with the orig- inal description, serves to identify Perkins’ plant; since it is all that remains of the original material, the fragment of Sumichrast 1558 (US) is designated the lectotype of L. acuminatus Perkins. 15. Lisianthius auratus Standl. Trop. Woods 37: 29. 1934 (as Lisian- thus). “Type. Honduras. Comavacua: open mountain forests, Siguatepeque, Edwards P-556 (¥1). 1972] WEAVER, REVISION OF LISIANTHIUS 271 Slender subshrubs, glabrous throughout. Stems 0.5-2.5 m. tall and to 1 cm. in diameter, erect or reputedly sprawling, green or brown at base, terete below but indistinctly 4-angled above. Leaves thin-membranaceous, dull dark green above and paler beneath, the principal lateral veins evi- dent on both surfaces, on drying the veinlets occasionally conspicuous, the blades 2.1-11.5 cm. long and 0.8-3.6 cm. broad, ovate-lanceolate or ovate or tending toward elliptic, long-acuminate, rounded, or short-atten- uate at base, the upper ones and those on the secondary branches more distinctly lanceolate and rounded at base; petioles 3-8 mm. long. In- florescences of pedunculate, apparently simple dichasia, these frequently reduced to 1 or 2 flowers. Flowering branches with a terminal and 1-3 pairs of axillary dichasia; axes of the flowering branch and the dichasial branches 4-angled. Flowers nodding or almost pendent, the pedicels 0.6-1.7 cm. long; bracteoles scarious, subulate. Calyx 5-10 mm. long, fused for 2/10-3/10 its length; the lobes 3.5—7.5 mm. long and 1-2 mm. broad, lanceolate, long-acuminate, scarious margined, weakly carinate abaxially; the tube 1.5-2.5 mm. long. Corolla tubular-funnelform, uni- formly yellow, 2.9-4.0 cm. long; the tube 2.2-3.0 cm. long and 4.5—7.5 mm. broad; constricted in the lower half, the lobes 5-11 mm. long and 3-4.5 mm. broad, 1/4~3/8 as long as the tube, spreading but not recurved, ovate-lanceolate, cuspidate-acuminate. Stamens inserted in the lower 1/4-1/3 of the corolla tube; filaments 2.0-3.0 cm. long, the longer ones exceeding the corolla lobes, and often equalling the style; anthers 2-2.5 mm. long and 1-1.5 mm. broad, narrowly oblong, the connective slightly exserted. Style 2.4-3.6 cm. long, surpassing the corolla lobes; stigma capitate. Capsule 8-12 mm. long and 3-5 mm. broad, oblong-ellipsoid, short-beaked; the marcescent corolla not persisting until the seeds are shed. FLowerinc: All year. CHROMOSOME NUMBER: 7 = 18. DISTRIBUTION: From near sea level to 1800 meters in pine forests or savannahs at scattered localities from western British Honduras to south- eastern Guatemala, and west-central Honduras to northeastern Nicaragua Map 5). SPECIMENS EXAMINED (us; isotype); Barranco El Soccoro, mountains about 8 km. w. of Siguatepeque, Williams & Williams 18382 (Gu, F); trincheras, 20 km. n. of Siguatepeque, Howard et al. 628 (a). InTIBUCHA: paredones humedos de la Cascada de Yamaranquila, Molina 6358 (us); Barranco Yamaranquila, cerca de Yashse, Molina 6524 (us). Nicaragua. CoMARCA DE EL Cabo: Rio Leicus 28 km. so. de Waspan, Molina 15206 (Fr); bosque de la Quebrada Cuyu, Molina 15037 (F); matorrales de Rio Leicus cerca del campo de aviacién de Tronquera, 35 km. so. de Waspan, Molina 15171 (F). 272 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Of the Central American species of Lisianthius, L. auratus appears to be most closely related to L. axillaris, a fact recognized by Standley when he described L. auratus. Although they have occasionally been confused, L. axillaris differs strikingly from L. auratus in having red flowers almost invariably borne singly in the axils of the foliage leaves. L. auratus strongly resembles certain of the West Indian species especially the L. troyanus alliance, and is probably quite similar to the stock from which they evolved. [To be concluded| ARNOLD ARBORETUM 1672-1072 CENTENNIAL JOURNAL OF THE ARNOLD ARBORETUM Volume 53. Number 3 a. _HARVARD UNIVERSITY our! BoTranical » Journal of the Arnold Arboretum ‘Published smarter in January, April, July, and October by the Arnold — Arboretum, Harvard University. 2 en roe eta bse 1 price $16. sr year, beginning g January, 1972. Pee es = tied = ‘Subscriptions and remittances should be ‘sent to Ms. Ellen Bernst ein, “Arnold Arbor g ge, Massa achu: gaeect he, ‘ofttim * Volumes I-XLYV, reprinted, and [o ae available a ‘the "Kraus = Reprint ‘Corporation, 16 East 46th Street, rk, 2 ne hoo? JOURNAL OF THE ARNOLD ARBORETUM VoL. 53 Jury 1972 NUMBER 3 A REVISION OF THE NEOTROPICAL GENUS LISIANTHIUS (GENTIANACEAE) #¢ 006849 RICHARD E. WEAVER, JR.* 16. Lisianthjus troyanus Urb. Symb. Antill. 6: 32. 1909 (as Lisian- thus). LECTOTYPE. Jamaica: near Troy, Harris 8789 (Nv!). Slender subshrubs, glabrous throughout. Stems 0.5—1.5 m. tall and to 1 cm. in diameter, erect or ascending, terete, green above and gray-brown below. Leaves membranaceous, the principal lateral veins prominent, glossy grass-green above, paler and dull beneath, the blades 2.9-12 (16.8) cm, long and 0.7-4.9 cm. broad, 3—4 times as long as broad, elliptic (usually narrowly so) to oblanceolate, broadest at or above the middle, the apices short-acuminate (long-acuminate or acute), the bases long- attenuate and tapering into the narrowly winged, 0.5—2.5 cm. long petiole. Inflorescences of apparently simple dichasia, these sometimes reduced to single flowers; flowering branches 1—2-ternately divided; inflorescence branches and stalks terete, elongate and the inflorescences diffuse. Flowers nodding, pedicels slender, 3-19 mm. long; bracteoles subulate, scarious. Calyx (4)5.5-10.5(12) mm. long, fused for 1/8-1/4 its length; the lobes 3-9(11) mm, long and 1-1.5 mm. broad, lanceolate or subulate-lanceolate, long acuminate, scarious-margined, ridged abaxially; the tube 1-3 mm. long. Corolla 3.3-4.6 cm. long, tubular-funnelform, uniformly bright yel- low; tube 2.2-3.3 cm. long and 5—7.5 mm. in diameter at the apex, con- stricted in the lower 1/3-2/5; lobes 9-14 mm. long and 2.5—4.5 mm. broad, widely spreading, ovate-lanceolate or lanceolate, abruptly short- acuminate (to gradually acuminate), 1/3-2/5(1/2) as long as the tube. Stamens inserted on the corolla tube 1/3—2/5 of the distance from the base to the apex; filaments 1.9-3 cm. long, not exceeding the corolla lobes but rarely almost equalling them; anthers 1.5—2.5 mm. long and 1-1.5 mm. broad, narrowly oblong, acutish, the connective slightly exserted. Styles 2.4-3.7 cm. long, exceeding the filaments, slightly surpassing the corolla lobes or not; stigmas capitate, distinctly bilobed. Capsules oblong-ellipsoid, 8-12 mm. long, not including the 1-2.5 mm. long beak, and 2.5-4 mm. * Continued from volume 53, p. 272. 274 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 in diameter; the marcescent corolla not persisting on the mature capsule. FLOWERING: All year. CHROMOSOME NUMBER: 7 = 18. DISTRIBUTION: From 90 to 600 meters on rocky limestone hillsides or roadside banks in western Jamaica (Map 10). REPRESENTATIVE COLLECTIONS amaica. near Troy, Harris 87389“ (ucwr: isotype); Cockpit Country, Troy, Britton 509 (Ny); Cockpit Country, near Balaclava, Marble 923 (Ny). HANOVER: Cocoa Walk District, W side of the Hanover Hills, near the Great Morass, Farr s.n. (IJ, NY); Houghton Mtn. District, 0.5 mi. S of Cave Valley, Proctor 11295 (IJ). WESTMORELAND: 0.5 mi. S of Moreland Hill School, Weaver & Weaver 1272 (A, BM, DUKE, F, LL, MICH, MO, S, UCWI, US). Lisianthius troyanus, a rather uncommon species, is probably most closely related to the Puerto Rican L. laxiflorus and the Cuban L. glandu- losus. It resembles these species strongly in most respects, particularly in its capitate, distinctly bilobed stigmas, but it should not be confused with them since the ranges of all three species are distinct. Of the Jamaican species, L. troyanus most closely resembles L. latifolius. That species differs however in its peltate stigmas, its corolla tubes not conspicuously constricted in the basal portion, its styles and filaments never exceeding the corolla lobes, and its broader leaves; in addition L. latifolius is re- stricted to the cloud forests and elfin woodland in the mountains of eastern Jamaica while L. troyanus is found on dry limestone hillsides of western Jamaica at an elevation below 700 meters. The holotype of Lisianthius troyanus, along with most of Urban’s type material was destroyed during the bombing of Berlin during World War II (H. Scholz, in litt.). However, I have seen 2 isotypes (NY, UCWI). Of these, the specimen at the New York Botanical Garden is more com- plete and better exhibits the characters described by Urban. Therefore, Harris 8789 (Ny) is designated the lectotype of L. troyanus Urb. The specimens of Lisianthius troyanus from the vicinity of Troy in the Cockpit Country, including the type collection and two additional ones, differ slightly from those specimens from the far western parishes of Jamaica. These latter specimens are distinguished from the typical element by having corolla lobes abruptly acuminate and slightly shorter in relation to the tube, generally shorter calyx lobes, and styles surpassing the corolla lobes. Although there is an overlap in all the critical characters, plants of the two groups of populations look slightly different from herbarium speci- 17. Lisianthius glandulosus A. Rich. in Ramon de la Sagra, Hist. Isla de Cuba 11: 102. 1850 (as Listanthus). Tver. Cuba: Sierra Maestra, St. Yago de Cuba, Linden 2014 (P!). 1972] WEAVER, REVISION OF LISIANTHIUS 275 L. stenophyllus Urb. Symb. Antill. 9: 237. 1924. Type. Cuba. ORIENTE: Sierra Maestra, Pinar del Papayo, Ekman 9269 (s!). Slender subshrubs, glabrous or minutely spiculate. Stems 0.2—1 m. tall and to 5 mm. in diameter, slender, erect or ascending, glabrous or rarely minutely spiculate, especially above, brown and terete below, green and 4-angled above. Leaves membranaceous, the venation not particularly conspicuous; foliage leaves subsessile or short-petiolate, the blades (1.5) 2.3-10.5 cm. long and (0.2)0.8—4.2 cm. broad, elliptic to very narrowly elliptic to ovate-lanceolate, widest at or below the middle, 2-10 times, but usually less than 6 times, as long as broad, gradually acuminate at the apex, the base obtuse or shortly attenuate into the short, winged Map 8. Distribution of Lisianthius species. L. domingensis (squares), L. glandulosus (half-circles), L. silenifolius (dots). petiole; petioles to 7 mm. long. Inflorescences of long-pedunculate, appar- ently simple (1-compound) dichasia, occasionally reduced to 1-2 flowers; flowering branches 1—2-ternately divided, each division terminated by a dichasium or a reduced dichasium; axes of the flowering branches and in- florescences strongly angled, the angles frequently expanded into narrow, scarious wings. Flowers nodding, frequently strongly, the pedicels rela- tively stout (3)8-13(20) mm. long; bracteoles subulate or linear, scarious. Calyx (5)6.5-12 mm. long, the lobes (4.5)5.5-10.5 mm. long and 1-2 mm. broad, narrowly lanceolate, long-acuminate, scarious-margined, weakly carinate abaxially. Corolla (2.4) 2.9-3.8 cm. long, tubular-funnel- form, uniformly bright yellow; the tube (1.9)2.2-3.2 cm. long and 3-6 mm. in diameter at apex, tightly constricted in the lower half, 3 times (or more) as long as the calyx; the lobes 5-9 mm. long and 2-4 mm. broad, spreading, narrowly ovate, abruptly or gradually short-acuminate (acute), 2/10-3/10 as long as the tube. Stamens inserted above the apex of the calyx lobes; filaments 1-2.2 cm. long, never exceeding the corolla lobes, the longer ones usually equalling the style; anthers 1.5—2 mm. long and ca. 1 mm. broad, oblong, obtuse. Style 1.5—-2.5 cm. long. Cap- sule fusiform or oblong-ellipsoid, short-beaked, 8-15 mm. long, not in- cluding the 1-2 mm. long beak, and 2—4 mm. in diameter ; calyx in fruit 1/3-2/3 as long as the capsule; the marcescent corolla sometimes per- sisting until the seeds are shed. FLOWERING: 276 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 DISTRIBUTION: From 300 to 1500 meters in pinelands or in moist lime- stone forests in the mountains, particularly the Sierra Maestra, in Oriente Province in eastern Cuba (Map 8) REPRESENTATIVE COLLECTIONS Cuba: without locality, Linden 2014 (BM; isotype). ORIENTE: Sierra Maestra, Pinar del Papayo, Ekman 9269 (s; type of L. stenophyllus), supra Daiquiri, Ekman 8174 (micu, s), Loma del Gato, Ekman 7004 (s), Gran Piedra, Shafer 9058 (NY, Us), Pico Turquino, Seifriz 1075 (us), w. of Rio Yao, Morton & Acuna 3429 (us), foothills on n. side of La Bayamesa, Webster 4078 (MICH, us); Pico del Cristal, Lopez 81 (us); Baracoa, pinelands on way to Florida, ater 3992 (F, MICH, s, UC, US); at the top of El Yunque, Ekman 3976 (2, S, US); minas de Iberia ad Taco Bay, Ekman 3807 (s); on Via Azul, between Suhenifin and Cajobabo, at km. 22 from Sabanilla, Morton & Alain 9000 (us); ; Arriba, Shafer 8409 (Ny, us); vicinity of El Cuero, Britton & Cowell 12768 (ny); Arralo de Medio, Shafer 3258 (nv); Sierra del Nipe, Loma Mensura, Ekman 5743 (s); Corojo, “Treinta Pinos,” Ekman 5178 (s). Lisianthius glandulosus is probably most closely related to the Jamaican L. troyanus and the Puerto Rican L. laxiflorus. It differs from both of these species in its corolla lobes which are less than 1/3 as long as the tube. It differs further from L. troyanus in its strongly 4-angled inflorescence branches and its style never exceeding the corolla lobes. It differs further from L. laxiflorus in that its corolla tube is 3 times or more longer than the ca Although Perkins (1902) considered it a synonym of Lisianthius glandu- losus, L. silenifolius, the only other Cuban species, differs strongly in that it is an annual or perennial suffrutescent herb with inflorescences of com- pound dichasia, sessile leaves, and ovoid capsules. L. silenifolius is here placed in subsect: HERBACE! of sect. LISIANTHIUS. In the protologue of Lisianthius stenophyllus, Urban stated that his new species is distinguished from the other Antillean species by its narrow leaves. L. stenophyllus is known from only the two specimens of the type collection, Ekman 9269, and these specimens indeed have leaves narrower than any found elsewhere in the genus. The leaves are at least 8 times longer than broad. However, except for being more densely spiculate, Ekman 9269 is identical to L. glandulosus in its other features. Since leaf shape is a widely variable character throughout the genus, L. stenophyllus is not maintained as a distinct species. Its apparent sym- patry with L. glandulosus would prevent me from treating it as a variety of that species. Therefore, at least until more material is available, I consider it to be a slightly aberrant population of L. glandulosus. 18. perro laxiflorus Urb. Symb. Antill. 3: 332. 1902 (as Lisian- hus).“LectotyPr. Puerto Rico: prope Pefiuelas ad montem Las pone Sintenis 4386 (ucwt!). } Lisianthus gracilis sensu Perkins, Bot. Jahrb, 31: 493. 1902, not Leianthus (= Lisianthius) longifolius 8 gracilis Griseb. 1972] WEAVER, REVISION OF LISIANTHIUS 277 Spindly subshrubs, glabrous throughout. Stems 0.3-1 m. tall and to 5 mm. in diameter, erect or ascending, or reputedly climbing, brown and more or less terete below, green and angled above. Leaves thin-membrana- ceous, the principal lateral veins and often the veinlets evident, at least upon drying; leaves 2.6-9.7 cm. long and 0.4-3.1 cm. broad, lanceolate (or ovate-lanceolate), broadest below the middle, the apices gradually long-acuminate, the bases obtuse or short-attenuate and tapering into the short, winged petiole, subsessile or the petioles to 5 mm. long. In- florescences of long-pedunculate, apparently simple (1-compound) di- chasia, often reduced to 1 or 2 flowers; flowering branches typically 1—2-ternately divided, the divisions terminated by dichasia, the lateral ones 1—(2—3)-flowered, the terminal one ( 1)2-3-flowered; axes of the flower- ing branches and inflorescences strongly 4-angled, the angles typically expanded into narrow, scarious wings. Flowers nodding, the pedicels relatively slender (0.3)1.1-2.2(2.8) cm. long; bracteoles foliaceous or more rarely nearly scarious, linear to narrowly elliptic, 2-8(10) mm. long. Calyx 8-13.5 mm. long, fused for 1/8 its length or less; the lobes 7-12 mm. long and 1-2 mm. broad, narrowly lanceolate, long-acuminate, ridged abaxially, essentially green but with narrow, scarious margins; the tube 0.5—2 mm. long. Corolla 3.0—4.5 cm. long, deeply cleft, funnelform, uni- formly bright yellow; the tube 1.8—2.5 cm. long and 4~7 cm. in diameter at the apex, slightly constricted in the lower 1/3-1/2; the lobes 1.2-2 cm. long and 2-4 mm. broad, 6/10 as long to nearly as long as the tube, spreading or perhaps slightly recurved, elliptic-lanceolate or lanceolate, gradually long-acuminate. Stamens inserted on the corolla tube one half or just less than half the distance from base to apex; filaments 1.4-2.5 cm. long, reaching to about the midpoint on the corolla lobes, occasion- ally equalling the style; anthers 1.5—2.5 mm. long and ca. 1 mm. broad, oblong, obtuse or acutish. Styles 1.9-2.7 cm. long; stigmas capitate, dis- tinctly bilobed. Capsule fusiform, short-beaked, 7-15 mm. long, not including the 1 mm. long beak, and 2.5-4 mm. in diameter ; calyx in fruit nearly equalling capsule; the marcescent corolla occasionally per- sisting until the seeds are shed. FLOwERING: Dec.—July. DIstRIBUTION: From 300 to 1300 meters in moist forests or along roadsides in the mountains of Puerto Rico (Map 9). REPRESENTATIVE COLLECTIONS Puerto Rico: Pefiuelas, inter frutices ad Las Cruces, Sintenis 4386 (F, GH, MO, NY, S, UCWI, us; isolectotypes); Maricao, Sintenis 202 (BM, GH, S, US; cited collection); Yauco, Garber 64 (cH, Ny; cited collection) ; 14 mi. ne. of Mayagiiez, Heller 4474 (¥, GH, MICH, MO, NY, US; cited collection) ; Lares, ad Buenos Aires, Sintenis 6077 (BM, Us; cited collection); prope Utuado, Sintenis 6424 (BM, GH, NY, US; cited collection); south side of El Yunque Mtn., etd quist 13195 (DUKE, MO); km. 15 on El Yunque Road from Palmer ie Flori a, Wagner 2 (a); Route 191, km. 17.2 on south side of the Luquillo Mountains, Wagner 1802 (DUKE); Rio de Maricao, Britton & Cowell 4228 (¥, MO, NY, US); Maricao Forest, near radar site, Howard & Nevling 15349 (a); Cerro Gordo, 278 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 San German, Vélez 1426 (GH); mountains of Toro Negro, Cerro de Punta, Alain 9480 (13); Sierra de Naguabo, Shafer 3528 (F, MO, NY, US). Lisianthius laxiflorus, the only species in the genus known from Puerto Rico, is most similar to the Jamaican L. troyanus Urb., from which it differs in its 4-angled inflorescence branches, its lanceolate leaves, its corolla lobes 3/5—-9/10 as long as the tube, and in its distribution. In her treatment of Lisianthius, Perkins (1902) included a plant which she called Lisianthus gracilis (Griseb.) Perk., citing as the basionym Leianthus longifolius (L.) Griseb. var. gracilis Griseb. (Fl. Brit. W. Ind. Is]. 421. 1862). From the specimens she cited, all Puerto Rican, it is obvious that this plant is the same as Lisianthius laxiflorus Urb. In fact, all of these specimens were cited slightly later by Urban as representative of L. laxiflorus. In his treatment of Leianthus longifolius (= Lisianthius longifolius L.) Grisebach (1862) recognized 3 varieties: a (the typical one), B gracilis, and y cordifolius. No authentic material of Grisebach’s has been seen; his description of 8 gracilis is brief and could equally well apply to some specimens of the Jamaican Lisianthius longifolius, as well as to L. laxiflorus or the Cuban L. glandulosus. However, in his comprehensive description of Leianthus longifolius, Grisebach (loc. cit.) states that the corolla tube is “twice or thrice” as long as the lobes. Lisianthius laxi- florus is thus excluded, since in this species the corolla tube is less than twice as long, and usually less than 1.5 times as long, as the lobes. In addition, Grisebach (loc. cit.) gave the distribution of Leianthus longi- folius as Jamaica and Cuba; Lisianthius laxiflorus is restricted to Puerto Rico From the description, then, Leianthus longifolius (including var. gracilis) as circumscribed by Grisebach (1862) does not include the Puerto Rican material. Therefore, since Perkins’ Lisianthus gracilis was based on Grisebach’s Leianthus longifolius var. gracilis, Urban was correct in applying a new name to the Puerto Rican Lisianthius. In describing Lisianthius laxiflorus Urban cited 9 specimens as repre- sentative of his new species. Although most of these were destroyed dur- ing World War II (H. Scholz, in litt.), there is reason to believe that at least one specimen of L. laxiflorus seen by Urban is extant. Among the collections at the University of the West Indies in Mona, Jamaica, is a specimen of Sintenis 4386, the label of which, including the word “typus,” is written in Urban’s hand as determined by Dr. C. D. Adams of that university. The locality data on this specimen, “Prope Pefuelas ad montem Las Cruces,” compares favorably with an entry, “Pefiuelas in monte Las Cruces,” in Urban’s protologue of L. laxiflorus. In addi- tion, this specimen was annotated by Perkins, as Lisianthus gracilis (Griseb.) Perk., in 1901, the year before the original description of L. laxiflorus was published. Since both Perkins and Urban worked at Berlin- Dahlem, it is quite possible that Urban saw this specimen. Twelve additional specimens of Sintenis 4386 have been seen, The labels 1972] WEAVER, REVISION OF LISIANTHIUS 279 of these were not written in Urban’s hand and the locality data (“Pefiuelas, inter frutices ad Las Cruces”) differ from those of the above-mentioned specimen. Until proved otherwise, I consider the specimen of Sintenis 4386 (ucw1) from “Prope Pefuelas ad montem Las Cruces” to be the only cited specimen probably seen by Urban, and I designate it the lectotype of Lisianthius laxiflorus Urb. 19. Lisianthius latifolius Sw. Prodr. 40. 1788; Fl. Ind. Occ. 1: 348. 1797 (as Lisianthus)./LEcTOTYPE. Jamaica, without exact locality, Swartz (s!). Leianthus latifolius (Sw.) Griseb. Gen. & Sp. Gent. 199. 1839 [1838]. Slender shrubs or subshrubs, glabrous throughout. Stems 0.3-2 m. tall and to 1.5 cm. in diameter, erect or ascending, terete, green and herbaceous above, gray-brown and woody below. Leaves firmly membranaceous, the principal lateral veins usually prominent, glossy grass-green or rather dark green above, paler and dull beneath, the blades (3.6)5-17(21.6) cm. long and (1.5)2—7 cm. broad, 2.5 to 4 times as long as broad, elliptic or oblong elliptic (obovate), widest at or above the middle, long-acuminate, the bases attenuate into the winged (0.3) 1—4.2 cm. long petiole. Inflores- cences of pedunculate, apparently simple dichasia, or occasionally reduced to solitary, axillary flowers; flowering branches 1—2-ternately divided, each division terminated by a dichasium, with 1-3 pairs of reduced leaves below the first division; inflorescence branches terete, elongate, and the inflorescences diffuse. Flowers nodding, pedicels relatively stout (0.6) 1-2.6 cm. long; bracteoles subulate, scarious. Calyx 6-14 mm. long, fused for 1/5-1/4 its length; lobes 4.5-10.5 mm. long and 1.5-2 mm. broad, lanceolate, long-acuminate, pale green, scarious-margined, ridged abaxially; the tube 1.5-3 mm. long. Corolla funnelform, 3.5—5.5 cm. long; the tube 2.2-3.9 cm. long and 7-10 mm. broad at the apex, bright yellow, slightly constricted in the lower third; lobes 11-18 mm. long and 3.5-6 mm, broad, greenish-yellow, spreading or weakly recurved, ovate-lan- ceolate or oblong-lanceolate, acute or abruptly short-acuminate, 2/5-1/2 as long as the tube. Stamens inserted on the corolla tube 1 /3-2/5 of the distance from the base to the apex; filaments 1.9-2.9 cm. long, not exceeding the corolla lobes; anthers 2—3.5 mm. long and 1-2 mm. broad, broadly oblong, obtuse, the connective not at all exserted. Ovary fusi- form; style 2.5-4.2 cm. long, never equalling the corolla lobes; stigma peltate. Capsule oblong-ellipsoid to fusiform, long beaked, 10-17 mm. long, not including the 3-7 mm. long beak, and 4-6 mm. in diameter; the marcescent corolla never persisting on the mature capsule. FLOWER- ING: All year. CHROMOSOME NUMBER: ” = 18 DistTRIBUTION: From 600-2100 meters in cloud forests and elfin wood- land in the Blue Mountains and the John Crow Mountains in eastern Jamaica (Map 10). 280 JOURNAL OF THE ARNOLD ARBORETUM [vor. 53 REPRESENTATIVE COLLECTIONS ig Jamaica. “Ind. occ.”, Swartz (s); without locality, Swartz (pm; isolectotype) ; without locality, MacFadyen (cH). PorTLAND: west slope of Silver Hill, Anderson & Sternberg 3259 (DUKE); Sir John Peak, Shreve s.n. (NY); Morces Gap, Weaver 1827 (A, BM, DUKE, F, MICH, MO, UCWI, Us); east slope of the John Crow Mtns., 1.5-2 mi. sw. of Ecclesdown, Proctor 9990 (13). St. ANDREW: along ridge between Morces Gap and John Crow Peak, Proctor 9517 (13); John Crow Peak, Shreve (Ny). St. THomas: trail from Portland Gap to Blue Mtn. Peak, Weaver 1952 (DUKE); Portland Gap, Weaver 1953 (A, DUKE, UCWI); near summit of Sugar Loaf Peak, Farr (13); summit of High Ij). Peak, Lewis ( Lisianthius latifolius is probably most closely related to L. domingensis, from the island of Hispaniola, but the two should not be confused since they have different distributions and clear cut morphological differences. No other species of Lisianthius is found in the cloud forests and elfin woodland in the Blue Mountains of Jamaica. Of the Jamaican species, L. latifolius most closely resembles L. troyanus Urb. The latter species differs, however, in its styles usually exceeding the corolla lobes, its corolla tubes conspicuously constricted in the basal portions, its stigmas capitate, and the beaks of its capsules less than 1/4 as long as the cap- sules. In Swartz’s herbarium at Stockholm (s) are three specimens of Lisian- thius latifolius, two inscribed “Ind. occ.” and one inscribed “Jamaica.” One of the former specimens and also the latter one include a single, un- attached flower of L. longifolius. Of these specimens the one inscribed “Jamaica” best demonstrates the characters described by Swartz and, excluding the fragment of L. longifolius, is here designated the lectotype of L. latifolius Sw. 20. Lisianthius Db pore eee Urb. Symb. Antill. 3: 333. 1902 (as Lisianthus),/LectotyPr. Haiti: Gonaives, Morne Fourmi, Buch 761 (13!). Spindly subshrubs, glabrous throughout. Stems 0.4-1 m. tall and to 5 mm. in diameter, erect or ascending, terete, brown below but green above. Leaves membranaceous, the principal lateral veins prominent, at least below, subsessile to distinctly petiolate, the blades 1.8—7.5 cm. long and 0.83.1 cm. broad, narrowly ovate to broadly elliptic, abruptly short acuminate to gradually acuminate, the bases more or less rounded to cuneate, or attenuate into the winged petiole; petioles, when present, to 10 mm. long. Inflorescences of reduced dichasia, these 1- or more rarely 2-flowered. Flowers nodding, the pedicels relatively stout, 0.3-2 cm. long; bracteoles 1.5-10 mm. long, narrowly-elliptic to linear-subulate, typically foliaceous but rarely nearly scarious. Calyx 5.5-9.5 mm. long, fused for 2/10-3/10 its length; the lobes 4-7.5 mm. long and 1-2 mm. broad, Jan- ceolate, acuminate, ridged abaxially, with very narrow, scarious margins; the tube 1-3 mm. long. Corolla 2.9-4.4 cm. long, tubular-funnelform, 1972] WEAVER, REVISION OF LISIANTHIUS 281 Maps 9-13. ~~ of Lisianthius species. Map 9, L. laxiflorus. a re - ongifolius. cee L. capitatus (dots), L. umbels (squares). L. latifolius dots), E . troyanus (squares). Map 13, L. adamsii paid en L ’ cordifolius (dots), L. exsertus (half-circles). uniformly bright yellow; the tube 2.3-3.1 cm. long, slightly constricted in the lower third; lobes 0.8-1.6 cm. long and 2.5—4 mm. broad, oblong- lanceolate, acute or short-acuminate, spreading or perhaps recurved, ca. equalling the corolla tube or slightly surpassing it, often equalling the style; anthers oblong, 1.5—2 mm. long and ca. 1 mm. broad. Style 1.6—2.6 cm. long; stigmas peltate. Capsule oblong-ellipsoid, beaked, 11-14 mm. long, not including the 1-3 mm. long beak, and 3.5—5 mm. in diameter; the marcescent corolla sometimes persisting until the seeds are shed. FLOWER- ING: Dec.—Apr DISTRIBUTION: From 900 to 1500 meters in wet, mossy forests in the Cordillera Central of the Dominican Republic and the Massif du Nord of Haiti (Map 8) SPECIMENS EXAMINED Haiti: Gonaives, Morne Fourmi, Buch 761 (cH; isotype); road from Camp #1 to La Barriére Couchant, Nash & Taylor 1096 (Nx); Massif du Nord, Ld. is du Nord, Morne Baron, Ekman 3889 (s); Port Margot, Morne Maleuvre, Ekman 2817 (s). Dominican Republic. Santo Domrnco: Cordillera 282 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Central, Loma la Campana, Ekman 11516 (Ny, s, Us). TRUJILLO VALDEZ: Los Guineos, San Cristébal, Marcano 3658 (us). Lisianthius domingensis, the only species known from the island of Hispaniola, is very poorly collected. It most closely resembles the Ja- maican L, /atifolius, both in morphology and habitat preference. That species differs, however, in its terete inflorescence branches, its usually 3-flowered dichasia, its long-beaked capsules, its broader leaves, and of course, its distribution The type of Lisianthius domingensis, along with most of Urban’s type material, was destroyed during the bombing of Berlin in World War II (H. Scholz, in litt.). Two isotypes (GH, 1J) have been seen. Of these, only the specimen at the Institute of Jamaica includes flowers. Since it is the more complete of the known isotypes and the one which better demonstrates the characters described by Urban, Buch 761 (1J!) is desig- nated the lectotype of L. domingensis Urb. The Haitian plants, including the type collection, differ slightly in leaf morphology from the plants of the Dominican Republic. The Dominican plants are distinguished from the typical element by having leaves usually more rounded at the base and subsessile. These differences are minor and not entirely consistent, and especially in view of the lack of material at hand, represent insufficient grounds for formal recognition of the two groups of populations as distinct taxa. 21. Lisianthius axillaris (Hemsl.) 0. Ktze. Rev. Gen. Pl. 2: 429. 1891. Leianthus axillaris Hemsl. Biol. Centr. Am. Bot. 2: 344. 1882. “Type. British Honduras: without exact locality, Barlee (DUKE, photo.!; x, holo.). Lisianthus francisiae Sprague, Kew Bull. 1929: 8. 1929.“Tvpr. British Hon- duras: Corozal, Francis 1 (DUKE, photo.!; x, holo.). Subshrubs, glabrous throughout. Stems terete, green, or brown near the base, erect or ascending, or the lateral ones spreading. Leaves mem- branaceous, shiny green above and duller below, the blades 1.6—-13.8 cm. long and 0.9-5.1 cm. broad, elliptic (ovate), short-acuminate (acute), the bases shortly attenuate, subsessile or short-petiolate, the petioles then winged and to 8 mm. long. Flowers borne singly in the axils of the foliage leaves, usually opposite, rarely in axillary, short-pedunculate monochasia or dichasia, the peduncles then 0.5-4.5 cm. long. Pedicels 0.9-3.0 cm. long, ascending or horizontal, becoming recurved with age; bracteoles subulate, scarious. Calyx 8.5-13.5 mm, long, fused for 1/5—1/6 its length; lobes 7-11.5 mm. long and 1.5-3 mm. broad, lanceolate, very long- acuminate, scarious margined, keeled abaxially, the keel to 0.5 mm. broad; tube 1.5—2.5 mm. long. Corolla tubular-funnelform, 3.7—5.3 cm. long; the tube 3.1-4.5 cm. long and 5-8 mm. broad at the apex, crimson, narrowly constricted in the basal half; lobes 5—10 mm. long and 3—4.5 mm. broad, greenish-yellow, spreading, narrowly oblong-ovate (ovate-lanceolate or ovate), acuminate or cuspidate-acuminate. Stamens inserted on the 1972] WEAVER, REVISION OF LISIANTHIUS 283 corolla tube about 2/5 of the way from the base to the apex; filaments 1.8-3.1 cm. long equalling or slightly exceeding the corolla tube in mature flowers; anthers 2.5-3 mm. long and 1-2.5 mm. broad, oblong, yellow. Ovary narrowly ovoid; style 3.2-4.6 cm. long, conspicuously exceeding the corolla lobes in mature flowers, exceeding the stamens; stigma peltate, distinctly bilobed. Capsule narrowly ovoid, indistinctly beaked, 11-19 mm. long and 4—6 mm. in diameter, the marcescent corolla persisting until well after the shedding of the seeds. FLOWERING: All year. CHROMOSOME NUMBER: n = 18 DISTRIBUTION: Near sea level to 500 meters along roadsides and in secondary scrub throughout much of British Honduras and into adjacent northeastern Guatemala and northeastern Yucatan, Mexico (Map 6). REPRESENTATIVE COLLECTIONS Mexico. QUINTANA Roo: Chetumal, Miranda 8116 (MEXxU). YUCATAN: Moctezuma, Enriquez 814 (mExvu). British Honduras. BELIZE Dist.: 8 mi. w. of Hattieville on Belize-Cayo road, Weaver & Wilbur 2255 (A, DUKE). Cayo Dist.: between Mileposts 39 & 40 on Belize-Cayo road, Weaver & Wilbur 2251 (A, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, s, UC, US); Mountain Pine Ridge, DUKE, F, MICH, MO, NY, $, UC, US); Blue Hole Section, Humming Bird Highway, Gentle 8945 (¥, LL); CorozAL Dist.: Corozal-Consijo road, Lundell 4857 (R, GH, MICH, MO, NY, US). ORANGE WALK Dist.: Honey Camp, Lundell 538 (BM, Ds, F, GH, MO, NY, S, US). STANN CREEK Dist.: Humming Bird Gap, Humming Bird Highway, Gentle 9246 (F, LL); Freshwater Creek Reserve, Pelly 16 (BM, F). Guatemala. PETEN: 6.5 km. e. of Dos Lagunas, Contreras 1617 (LL); Nictun, Lake Petén, Lundell 3159 (micu, s); Bajo de Santa Fé, ca. 25 km. e. of Tikal, Contreras 1438 (¥, LL). Lisianthius axillaris is one of the more distinctive of the Lisianthius species. With its crimson corolla tube and solitary flowers in the axils of the foliage leaves, it is not easily mistaken for any other species. The only other species with red flowers is L. viscidiflorus (subsect. HERBACE!), of central Guatemala and Chiapas, Mexico. L. viscidiflorus, however, has shorter, obtuse corolla lobes, purple anthers, and both terminal and axil- lary inflorescences of compound dichasia. Bie Lisianthius francisiae Sprague is synonymous with L. axillaris. Sprague, in the description of L. francisiae, mentions that the only differences be- tween the two are in the vegetative portion of the plant, L. francisiae being herbaceous and freely branched, L. axillaris being woody and un- branched; and L. francisiae having larger, narrower leaves more nar- rowed at the base. Leaf shape and size have proved of little taxonomic utility in this genus, both characters varying markedly in different parts of a given plant. All of the species of sect. ListaNTHIUS subsect. Fruticosi are charac- terized by having a variety of growth forms, and L. axillaris is no excep- tion. Freely branched plants and entirely unbranched plants are fre- quently encountered in this group. The populations of L. axillaris which 284 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 I have seen in British Honduras were composed of plants with both types of branching patterns (Weaver & Wilbur 2251 and Weaver & Wilbur 2254). The unbranched plants seemed to be those growing in the more unfavorable spots, locations which were either shady or dry. Lisianthius axillaris is woody only at the base. The type specimen of L. francisiae, of which a photograph has been seen, obviously represents the midportion of a single plant. In the midportions, plants of L. axil- laris are not distinctly woody. Williams (1968, 1969) included Lisianthius francisiae as a synonym of L. axillaris. To my knowledge, no author since Sprague has recognized L. francisiae as a distinct species. Lisianthius petenensis Standl. & Steyerm, was based on Lundell 3153, from Nictun, Lake Petén, Guatemala (holotype us, isotype MIcH). Wil- liams (1968, 1969) considered L. petenensis to be a synonym of L. brevi- dentatus. Lundell 3153, however, differs strikingly from that species in its much broader, red corolla tube, longer corolla lobes, style not exceed- ing either the filaments or the corolla lobes, and subsessile leaves. Careful examination of the type and only known collection of L. petenensis con- vinced me that it is a hybrid between two species widely separated from L. brevidentatus. Lisianthius axillaris and L. saponarioides are the only species presently known from the Guatemalan department of Petén. Both of these species have been collected in the vicinity of Lake Petén, the type locality of L. petenensis. Lundell 3153 is morphologically intermediate between L. axillaris and L. saponarioides and is probably a hybrid between them. The putative hybrid and its presumed parents are compared in TaBLe 1. Lundell 3153 is similar to Lisianthius axillaris in its red corolla tube, distinctly carinate calyx lobes, and stamens of unequal length in a given ower. It resembles L. saponarioides in its inflorescences of frequently sessile, compound dichasia, style not exceeding either the filaments or the corolla lobes, and its indistinctly lobed stigma. In the remaining characters outlined in TaBLe 1, it is intermediate between the two species. In search of further proof of their hybrid origin, pollen was taken from both specimens of Lundell 3153 and stained with cotton blue-lactophenol. As a control, pollen was taken from a specimen of Lisianthius axillaris (Contreras 1438, ¥) and L. saponarioides (Contreras 3766, F), both from the vicinity of Lake Petén, and similarly stained. The percentage of pollen viability for all four specimens is summarized in TABLE 2. 22. Lisianthius seemannii (Griseb.) O. Ktze. Rev. Gen. Pl. 2: 429. 1891 (as Seemannii). Leianthus Seemannii Griseb. in Seemann, Bot. Voy. Herald 170. 1354.“ Type. Panama. Darien: Pifias, Seemann 1064 (mo, photo.!; kK, holo.). Lisianthus corymbosus Perk. Bot. Jahrb. 31: 491. 1902. “No type designated. Cited specimens presumably at (B) and destroyed during World War II. Slender subshrubs, glabrous throughout. Stems 0.5-2 m. tall and to TABLE 1. A morphological comparison of “L. petenensis” and its presumed parents L. SAPONARIOIDES “L. PETENENSIS” L. AXILLARIS INFLORESCENCES COROLLA TUBE COROLLA LOBES INSERTION OF STAMENS/ LENGTH OF COROLLA TUBE FILAMENTS STYLE STIGMA Compound dichasia some nearly sessile Winged 0-1 mm. long 5.5-10 mm. long Ridged Salverform, constricted nearly its entire length, lobes flaring 1.0-1.5 cm. long, green 4-7 mm. long 0.7-0.8 0.3-0.9 cm. long, equal in length, exceeding the style 0.6 1.0 cm. long, not exceeding corolla lobes Indistinctly lobed Compound dichasia, some nearly sessile Narrowly winged 1-3 mm. long 7-10 mm. long Distinctly carinate pape 2/3 ts length, lobes idan 2.3-2.5 cm. long, red 6-6.5 mm. long 0.6 1.4-1.6 cm. long, unequal in length, exceeding the style 1.7—2.0 cm. long, not exceeding corolla lobes Indistinctly lobed Solitary, axillary flowers Terete 9-30 mm. long 8.5-13.5 mm. long Distinctly carinate Tubular-funnelform, constricted 1/2 its length, lobes spreading 3.1-4.5 cm. long, red 5-10 mm. long 0.4 1.8-3.1 cm. long, unequal in length, not exceeding e style 3.2—4.6 cm. long, exceeding corolla lobes Distinctly lobed SOIH.LNVISIT 40 NOISIATY “AIAVAM [27261 S8Z 286 JOURNAL OF THE ARNOLD ARBORETUM [vor. 53 TABLE 2. Pollen viability in “L. petenensis” and its presumed parents TOTAL NO. OF % NOT GRAINS EXAMINED % VIABLE VIABLE L. saponarioides 900 97% 3% “L. petenensis” Lundell 3153 (us) 1040 19% 81% Lundell 3153 (mMIcH) 742 21% 79% L. axillaris 1209 97% 3% 1.5 cm. in diameter, erect or ascending, terete and gray-brown below, 3—-4-angled and green above. Leaves thin-membranaceous, glossy dark green above, paler and dull beneath, the lateral nerves prominent, the blades 1.9-15.7 cm. long and 1.1—5.3 cm. broad, the petioles 0.3—2.5 cm. long or the upper leaves subsessile; principal foliage leaves elliptic, long- acuminate, shortly attenuate into the winged petiole, the leaves on the flowering branches smaller, essentially sessile, lanceolate to ovate, rounded or obtuse at the base. Inflorescences of pedunculate or sessile 1—4-com- pound dichasia, the lateral branches scorpioid; inflorescence branches contracted and the dichasia or groups of dichasia compacted, 4-angled, 2 of the angles frequently expanded into narrow, scarious wings; flowering branches with several to many pairs of reduced leaves below the first division. Flowers horizontal or nodding, sessile or with pedicels to 6 mm. long. Calyx 8.5-14 mm. long, fused for 1/10-1/6 its length; lobes 7.5-12 mm. long and 1-2 mm. broad, narrowly lanceolate, very long- acuminate, with broad scarious margins, ridged or carinate abaxially. Corolla 3.2—5.2 cm. long, narrowly tubular-funnelform, uniformly lemon- yellow outside and cream-colored inside; tube 2.3-3.7 cm. long and 4-7 mm. in diameter at the apex, constricted in the lower third; lobes (9) 10.5- 15 mm. long and 2.5-6 mm. broad, widely spreading and slightly recurved at the tips, narrowly lanceolate, long cuspidate-acuminate, ca. half as long as the tube. Stamens inserted on the corolla tube ca. 1/3 of the distance from the base to the apex; filaments 1.9-3.1 cm. long, not sur- passing the corolla lobes; anthers 1.5—2.5 mm. long and 1—1.5 mm. broad, oblong, acutish, the connective slightly exserted. Style 2.6-3.7 cm. long, at most equalling the corolla lobes, always surpassing the filaments; stigma peltate. Capsules oblong-ellipsoid, very short-beaked, 9-14 mm. long and 3—6 mm. broad; calyx in fruit equalling or slightly exceeding the capsule; the marcescent corolla persisting until the seeds are shed. FLOWERING: All year. CHROMOSOME NUMBER: 7” = 18. DisTRIBUTION: From 200-1200 meters, mostly along roadsides, but also in wet or dry secondary scrub or at the margins of forests from north-central Costa Rica, through the mountainous portion of Panama to northwestern Colombia (Map 7). 1972] WEAVER, REVISION OF LISIANTHIUS 287 REPRESENTATIVE COLLECTIONS 10055 (¥, MO). HeErepIA: Sta. Barbara, au pied du Barba, Pittier 1670 (us; cited collection of L. corymbosus). SAN JOsE: vicinity of El General, Skutch 4109 (GH, MO, s, Us); El Rodeo, Hunnewell 16713 (DUKE, GH). Panama. CHIRIQUT: hetween Hato del Jobo and Cerro Vaca, Pittier 5292 (Ny). COcLE: Bismarck, above Penonomé, Williams 259 (Ny); 3.5 km. se. of El Valle de Anton along road from El Espino to El Valle, Weaver 1671 (a, BM, DS, DUKE, F, IJ, LL, MICH, MO, NY, S, UC, US); 2 mi. s. of Coclé, ee 8248 (DUKE, US). Darrin: vicinity of Pifias, ‘Duke 10aez. (mo). HERRERA: 10 mi. s. of Oct, Tyson, Dwyer & Blum are (mo); ; vicinity of Las Minas, ‘La ‘Pehs, Stern, Eyde & Ayensu 1779 (MIcH, us). Veracuas: Isla de Coiba, Dwyer 1650 (Mo). Colombia. ANTIOQUIA: landelide on Rio Guapa, 8 km. e. of Guapé, Haught 4652 (COL, F, US Lisianthius seemannii is one of the better collected of i Central American species. With its congested inflorescences, its long narrow calyx and corolla lobes, and its corolla paler inside than out, L. Sicesnet should not be confused with any other species of series Longifolii. Its relationships within the group are obscure. It is the only species of series a typically with compound dichasia. It most closely resembles the Jamaican L. longifolius, especially vegetatively, but differs strongly from that species in the inflorescence type. Lisianthius seemannii has the double distinction of being the only species found on a Pacific island (Isla de Coiba), or on the South American con- tinent. Haught 4652 (cot, F, us) from Dept. Antioquia, Colombia, is the only known collection of a ea aust species from South America. Ewan determined this specimen as L. arcuatus Perk. and reported it (Proc. Biol. Soc. Wash. 63: 163-166. 1950) as representing a previously unreported extension of the genus into South America. In the same article Ewan stated that L. seemannii probably represents the same species as L. arcuatus. L. arcuatus and L. seemannii are certainly not synonymous, the former is nat a synonym of L. skinneri. Haught 4652 undoubtedly is L. seeman In the key to the species of "Lisianthius in Perkins’ treatment (1902) of the genus, L. seemannii and L. corymbosus are separated as follows: L. seemannii with ecarinate calyx tobe and the stamens included; L. corymbosus with carinate calyx lobes and the stamens shortly exserted. Judging from her treatment in general, by “stamens exserted” Perkins evidently meant that the stamens exceed the corolla lobes when pressed. A duplicate of one of the specimens cited by Perkins as representative of L. corymbosus, Pittier 1670, is among the collections at the United States National Herbarium. I have seen this specimen and the stamens defi- nitely do not exceed the corolla lobes in any of the flowers. The calyx lobes of L. seemannii are never completely ecarinate, although the keel in some specimens is not particularly distinct. At any rate, Pittier 1670 is so similar in all respects to the type of L. seemannii, of which I have 288 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 seen a photograph, that a more distinctly carinate calyx would hardly serve to separate L. corymbosus as a distinct species. 23. Lisianthius longifolius L. Mant. Pl. 43. 1767 (as Lisianthus). “LECTOTYPE. T. 9, Figure 1 in P. Browne, Civ. Nat. Hist. Jam. 1756. Tachia longifolia (L.) Mart. ex G. Don, Gen. Hist. 4: 197. 1837. Leianthus longifolius (L.) Griseb. Gen. & Sp. Gent. 196. 1839 [1838]. Leianthus longifolius 8 scabridus rng loc. cit. p. 197. Typr. Willdenow 3556, presumably at (B) but not s Lisianthus longifolius var. seabvidas (Griseb.) Perk. Bot. Jahrb. 31: 492. 1902 Spindly subshrubs, minutely puberulous or essentially glabrous. Stems 0.5—2 m. tall and to 2 cm. in diameter, erect, ascending, or sprawling, terete or somewhat quadrate below, green above and gray-brown below, essentially glabrous, or more commonly, moderately to densely short- puberulous above, becoming nearly glabrous below. Leaves firmly mem- branaceous, glossy grass-green or somewhat gray-green above, paler and dull beneath, glabrous to sparsely short-puberulous above and/or on the elliptic to narrowly-elliptic (ovate or lanceolate), short- to rather long- acuminate (acute or rarely obtuse or mucronate), the bases obtuse to cuneate or short-attenuate, the petioles to 2.9 cm. long or the leaves subsessile, glabrous to rather densely short-puberulous; leaves on the flowering branches subsessile or short-petiolate (0.5)1.5—7(8.5) cm. long and (0.3)1-3(4.5) cm. broad, ovate-lanceolate to ovate and lanceolate (elliptic). Inflorescences of dichasia, these often reduced to 1 or 2 flowers; flowering branches with 1-6 pairs of leaves below the first division, sparsely to freely branched, the lower branches usually alternate or bifurcate, each segment diverging from a straight line and appearing almost dichot- omous, the upper branches usually opposite, the axes terete, nearly glabrous to rather densely short-puberulous. Flowers subsessile or more commonly pedicellate, the pedicels then to 16 mm. long, relatively stout, glabrous or short-puberulous; bracteoles linear or subulate, scarious. Calyx (0.7)1-2(2.8) cm. long, fused for 1/8—-1/10 its length, alate or very rarely merely carinate, when alate the wings 0.5—4 mm. broad, gla- brous or sparsely puberulous along the wing or keel; lobes (0.6)0.8-1.8 (2.5) cm. long, lanceolate, long-acuminate, with a scarious or hyaline margin; tube 1-4 mm. long. Corolla (2.9)3.5-6(6.7) cm. long, glutinous, tubular-funnelform; tube (2.3)2.7-3.9 cm. long and 6—-14(16.5) mm. broad at the apex, bright yellow, glossy to rather dull, constricted in the lower third; lobes 0.7—2(2.8) cm. long and 3-9 mm. broad, ovate-lanceo- late or oblong-lanceolate, abruptly . more or less gradually acuminate, yellow-green, strongly recurved or merely spreading and only the tips recurved, 1/4-3/4 as long as the tube. Stamens inserted on the corolla tube ca. 1/4-1/3 of the distance from the base to the apex; filaments 1972] WEAVER, REVISION OF LISIANTHIUS 289 2—3.8 cm. long, never exceeding the corolla lobes; anthers 2—4.5 mm. long and 1-2 mm. broad, oblong to narrowly oblong, obtuse to acutish at the apex, the connective slightly exserted or not. Style 2.3-4.8 cm. long, very rarely exceeding the corolla lobes, but nearly always conspicuously exceeding the filaments; stigma capitate or me peltate. Capsule ellipsoid or oblong-ellipsoid, beaked, 7-17 mm. long, not including the 1-4 mm. long beak, and 3-7 mm. in diameter; calyx in fruit nearly as long to longer than the capsule; the marcescent corolla persisting until long after the seeds are shed. FLOWERING: All year. CHROMOSOME NUM- PER. B= 18, DIsTRIBUTION: From sea level to 1500 meters along roadsides and in rather dry secondary scrub throughout Jamaica except for the western- most parishes (Map 11). REPRESENTATIVE COLLECTIONS amaica. CLARENDON: 1.5 mi. se. of Chapelton on road tg Suttons, Weaver & Weaver 1264 (A, DUKE); vicinity of Bog Hole, 1 mi. ne. of Corner Shop, Weaver & Weaver 1265 (A, DUKE); summit area of Croft's Hill, Proctor 29266 (tJ); 1 mi. due s. of Arthur’s Seat, Proctor 26443 (1J, LL, MICH), PoRTLAND: 3 mi. e. of Port Antonio, Weaver & Weaver 941 (A, DUKE, UCWI); near mouth of Rio Grande, Proctor & Steore 11860 ss Sr. ANDREW: above Gordon i Jamaica A-1 to Hollymount on Mt. Diablo, Weaver & Weaver 894 (A, DUKE, F, MICH, UCWI, Us); between Lime Hall and St. Ann’s Bay, Stearn 718 (A, BM); Roaring River, Stearn 739 (A, BM). St. CATHERINE: halfway between Ewarton ig Worthy Park, Adams 13008 (DUKE, ucw!) ; Bog Walk, Crawford 819 (cu, Y, US); vicinity of Red Ground, so 5.5 mi. n. of Old Harbour, Weaver & Wager 1118 (A, BM, DUKE, F, LL, MO, NY, S, US); 4 mi. sw. of Ewarton on road to Worthy Park, Weaver & Wasaat 1251 (A, BM, DUKE, F, IJ, MICH, MO, UcwWI, Us); St. ELIzABETH: hills behind ete School, Weaver & Weaver 1191 (A, BM, DUKE, MICH, MO, UCWI, US); 1.5 ne. of Malv Ae on road to Leeds, Weaver & Weaver 1197 (A, DUKE, IJ, MICH, s, US). St. JAMES: hills overlooking Ironshore Estate, Proctor 23239 (IJ, LL, MICH, Ny); Glen evon, Montego Bay, Norman s.n. (pm). St. Mary: Gibraltar Beach, near Oracabessa, Prictet 7455 (13). St. THomas: Helen’s Gap, Cinchona, Nichols 64 (Fr, MO, NY, US); 1-2 mi, nnw. of Whitehall, _— 19805 (A, 1); ca. 3 mi. nw. of Trinityville, Proctor 24396 (1J, LL); 2.5 mi. nw. of Richm Vale, Weaver & Weaver 953 (A, DUKE, MICH, US); Arntully, Waawer & eam 976 (A, DUKE, F, US); 1.5 mi. nw. of Bath on Foad fn. macy emgage d road from - Weaver 1200 (A, BM, DUKE, UCWI, US); along syed inl Pee Epo DU LL, MICH, MO, Us); road from Burnt Hill to Barbecue Bottom, Weaver & Weaver 907 (A, DUKE, F, MICH, UCWI, US). According to Stearn (1957), Linnaeus had virtually no first-hand ex- 290 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 perience with Jamaican plants except for those in Patrick Browne’s her- barium, which he purchased in 1758. Savage (1945) does not list a speci- men of Lisianthius longifolius in his catalogue of the extant portion of the Linnaean Herbarium in London. Therefore, it must be assumed either that Patrick Browne’s herbarium did not contain a specimen of this species when it was purchased by Linnaeus, or if a specimen was present, it was lost in the interim. It is entirely possible that Linnaeus based his Lisian- thus (= Lisianthius) longifolius on an illustration rather than on an actual specimen. In the original description of this species, Linnaeus (1767, p. 43) cited two figures, one of Patrick Browne (1756, t. 9, figure 1) and one of Hans Sloane (Voyage 1: ¢. 101, figure 1. 1707). Therefore, a lectotype must be chosen from the illustrations. Linnaeus’s (1767, p. 43) diagnosis of Lisianthus (= Lisianthius) longi- folius consists merely of the phrase, “Lisianthus foliis lanceolatis.” In addition, he included polynomials given by Browne (1756, p. 157) and Sloane (Cat. Pl. Jam. 58. 1696) for the same species. Linnaeus’s diagnosis was obviously made with Browne’s polynomial, “Lisianthus [Lisianthius| erectus, foliis lanceolatis, floribus singularibus terminalibus,’ in mind. In addition, Browne’s figure more accurately portrays L. longifolius than does Sloane’s figure. Therefore, Browne’s figure (Civ. Nat. Hist. Jam. f. 9, figure 1. 1756) is designated the lectotype of L. longifolius L. Lisianthius longifolius, a rather weedy plant, is found most commonly on roadside banks and in the secondary scrub that covers much of Ja- maica. It is frequently encountered and its large, yellow flowers are quite conspicuous. Indeed, L. longifolius has been by far the most frequently collected of the species. Lisianthius longifolius is also the most variable of the species. The characters in which this variation is most pronounced are as follows: (1) pubescence; (2) length of the corolla; (3) length of the calyx; (4) the ratio between the length of corolla lobes and length of corolla tube; and (5) width of the calyx wing. Both the range and the pattern of the variation are most perplexing. These vary independently and abruptly without any obvious correlation with elevation, habitat, or geography. As a result a given population may differ noticeably in certain of these critical characters from another only a few miles away, while closely resembling in these same characters yet another population far removed geographically but differing in others. An analysis of the variation 1s presented in Ficure 8. Eight populations which display the range of variation within the species were scored for each of these critical variables. The degree of pubescence in each population was ranked on a scale from 0 to 4 and represented schematically by means of proportionately filled circles. The empty circles represent the most nearly glabrous type (0); the complete- ly filled circles represent the type which is most densely pubescent (4). The ranges of the remaining characters were numerically evaluated and the values placed proportionately on each of the four axes of a tetragon. The eight populations are: Corolla Length Pubescence Calyx Length Calyx Wing Width Corolla Lobe Length Corolla Tube Length F Ficure 8. An analysis of the variation in Lisianthius longifolius. See text for explanation. [ZL61 SATHINVISIT JO NOISIATY “WAAVAM T62 bo oO bo JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Proctor 23239. St. James: hills overlooking Ironshore Estate. Weaver & Weaver 997. TRELAWNY: road from Burnt Hill to Barbecue Bottom. Adams 13008. Sr. CATHERINE: half-way between Ewarton and Worthy Park Weaver & Weaver 894. St. ANN: road from Jamaica A-1 to Hollymount on Mt. Diablo. Weaver & Weaver 1200. St. Tuomas: road from Bath to Beacon Hill. Weaver & Weaver 976. St. THomas: Arntully. Weaver & Weaver 1118. St, CATHERINE: vicinity of Red Ground. Weaver & Weaver 1191. St. Evizasetu: hills behind Hampton School. ROAR oS ff wp Populations D and F are strikingly different; D is glabrous and F is quite densely pubescent and the other characters do not overlap at all. Population E is similar to Population F in length of corolla, length of calyx, and corolla lobes/corolla tube ratio, but is only slightly pubescent and the width of the calyx wing is much broader, the range of variation of which overlaps that of Population D. Population C is similar to Popu- lation D, but the corolla and calyx are both slightly shorter in the former. Population B is similar to Population C in all respects except that the wing on the calyx is much narrower and the plants are densely pubescent; it is intermediate between Populations C and F. Population A resembles Population B except that it is less densely pubescent and the calyx wing is broader, and it is therefore intermediate between that population and Population C. Population H resembles Population E, from the other end of the island, except that it is slightly larger in all respects. Population G is intermediate between Populations C and H. Both Grisebach (1839, 1845) and Perkins (1902) recognized a variety scabridus of Lisianthius (or Leianthus) longifolius. No authentic material of this has been seen, but ex char. it would seem to be something resem- bling Populations E or F above. With such a striking array of intermedi- ate and overlapping forms, it is impossible to segregate any of these popu- lations, or the groups of populations of which they are typical, as a distinct taxon of any rank. In his Flora of the British West Indian Islands, Grisebach (1862), instead of recognizing var. scabridus as before, described a new variety, gracilis, of Leianthus (= Lisianthius) longifolius. At the same time he described the range of L. longifolius as including Cuba. No authentic material of his var. gracilis has been seen, but the description could apply equally well to some specimens of Lisianthius longifolius as here inter- preted or to the Cuban L. glandulosus. Since the name cannot be cor- rectly applied at present, it is here considered a nomen dubium. 24. Lisianthius cordifolius L. Mant. Pl. 1: 43. 1767 (as Lisianthus). YPE. Specimen no. 213.1 (Savage, 1945) in the Linnaean Her- barium (LINN), probably collected by P. Browne. (Microfiche (IDC 5075. 18: II. 1.) seen.) Tachia cordifolia (L.) Mart. ex G. Don, Gen. Hist. 4: 197. 1837. 1972] WEAVER, REVISION OF LISIANTHIUS 293 Leianthus cordifolius (L.) Griseb. Gen. & Sp. Gent. 197. 1839 [1838]. Leianthus longifolius y cordifolius (L.) Griseb. Fl. Brit. W. Ind, Isl. 424. 1862. Lisianthus longifolius var. cordifolius (L.) Perk. Bot. Jahrb. 31: 492. 1902. Low subshrubs, moderately to densely puberulous. Stems 3-6 dm. tall and to 1 cm. in diameter, erect or ascending, greenish or straw-colored above and gray-brown below, the branches nearly always alternate, mod- erately to densely short-puberulous above, becoming nearly glabrous be- low. Leaves firmly membranaceous, dull-textured, bright green above, paler and dull beneath, the midribs short-puberulous (glabrous), the margins minutely ciliolate or rarely entire; blades of the foliage leaves 1.2—4.9 cm. long and 0.6—2 cm. broad, gradually acuminate, the base atten- uate to obtuse and rounded or subcordate, the petioles densely short- puberulous, winged, 1.5-7 mm. long; leaves on the flowering branches decreasing in size toward the apex, 0.3-2.2 cm. long and 0.2-1.5 cm. broad, the lower ones narrowly ovate, rounded or subcordate at the base and short-petiolate, the upper ones broadly ovate, distinctly cordate at the base and subsessile. Inflorescences of apparently simple dichasia, these reduced to 1 or 2 flowers, or very rarely 3-flowered; flowering branches with several to many pairs of leaves between each division of the primary and secondary axes, often unbranched and terminated by 1 to 2 flowers, or sparsely to freely branched, the branching bifurcate or very rarely opposite, the axes terete, moderately to densely short-puberulous. Flowers nodding, the pedicels relatively slender, short-puberulous, 3-10 mm. long; brac- teoles subulate, scarious. Calyx 7-11 mm. long, fused for 1/8-1/5 its length; lobes 6—9.5 mm. long and 1—2 mm. broad, lanceolate, long-acumi- nate, carinate abaxially, with broad scarious or hyaline margins, the cen- tral green band ca. 1/2 the total width of the lobe; tube 1-2 mm. long, sometimes sparsely puberulous at the base. Corolla tubular-funnelform, 3.0-4.4 cm. long; the tube 2.1-3.2 cm. long and 6-8 mm, in diameter at the apex, bright yellow, constricted in the lower third; lobes 9-13.5 mm. long and 3.5—9.5 mm. broad, 1/3—4/10 as long as the tube, spreading and the tips recurved, yellow-green, narrowly ovate to ovate-lanceolate. Sta- mens inserted on the corolla tube ca. 1/3 of the distance from the base to the apex; filaments 1.9-3.3 cm. long, the longest ones nearly equalling or barely surpassing the corolla lobes; anthers 2—2.5 mm. long and 1-1.5 mm. broad, oblong, obtuse or acutish at the apex, the connective slightly exserted. Style 2.4-3.8 cm. long, exceeding (equalling) the corolla lobes, always exceeding the filaments; stigma capitate. Capsule ellipsoid or oblong-ellipsoid, short-beaked, 9-12 mm. long, not including the 1-2 mm. long beak, and 3-6 mm. in diameter; calyx in fruit 2/3 as long to longer than the capsule; the marcescent corolla persisting long after the seeds are shed. FLOWERING: Nov._June. CHROMOSOME NUMBER: ” = 18 DIsTRIBUTION: From 70 to 400 meters in crevices of rock of steep limestone gorges, or rarely in dry scrubby forest on limestone in the vicinity of Kingston, Jamaica (Map 12). 294 _ JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 REPRESENTATIVE COLLECTIONS Jamaica. St. ANDREW: gorge of the Cane River, below the falls at Newstead, Weaver 1821 (A, DUKE, F, MICH, UCWI, Us); Cane River Valley, Harris 12315 , F, MO, NY, S, US); on canyon wall along Hope River, Yuncker 17374 (r, MICH, Ss); Hope River Gorge, Norman 171 (sm); Long Mountain, back of Kingston, Webster & Wilson 4860 (a, 1J, MICH, s, US). Lisianthius cordifolius is one of the most restricted of the species, known from only two limestone gorges and an adjacent mountain in southeastern Jamaica. It is closely related to L. longifolius, and in fact, has been considered to be a variety of that species by Grisebach (1862) and Perkins (1902). L. cordifolius differs from L. longifolius in its carinate calyx lobes, its branching very rarely opposite, its flowering-branch leaves cordate at the base, and its styles equalling or surpassing the corolla lobes. As explained earlier, L. longifolius is a bewilderingly variable spe- cies, and some specimens approach L. cordifolius in their morphology. However, no specimens have been seen which are not immediately refer- able to one species or the other. There is some indication that Lisianthius cordifolius and L. longifolius may hybridize in nature. L. cordifolius is known from only three popula- tions in an eight square mile area in the vicinity of Kingston, Jamaica. Although apparently not common, L. longifolius has been collected in that same area. It appears that the pollen viability of one of the three popu- lations of L. cordifolius, Webster & Wilson 4860, is somewhat lower than might be expected in a fully differentiated, fertile species, giving rise to the possibility of its being a hybrid. The habitat of this population, how- ever, is quite different from that of the other two known populations of L. cordifolius. Webster & Wilson 4860 was collected in dry, scrubby forest on limestone, while the other two populations grew in rock crevices of steep limestone cliffs. The reduced pollen viability of Webster & Wilson 4860 was discovered near the end of my study when field and experimental work was no longer possible. Thus, whether this reduced pollen viability is due to an unfavorable environment or to the population being a hybrid one, I cannot determine at present. Experimental crosses have been carried out between plants of Lisian- thius cordifolius and L. longifolius grown in the greenhouse of Duke University. The flowers of the seed plants were emasculated before the anthers had dehisced and the plants were isolated in a separate chamber of the greenhouse. All unused flowers were removed before anthesis as 2 preventative against stray pollen. Pollination was accomplished by rub- bing an anther from the pollen parent across the stigmas of the seed parents until the stigmas were covered with pollen. The pollination pro- cedure was repeated every day until the flowers withered. The results are summarized in TABLE 3, The seed set in the hybrid cross (Cross 2) is somewhat reduced from that obtained when the seed parent (L. cordifolius, Cross 3) was selfed. The hybrid seeds were planted and the germination was high. The seed- lings are vigorous but to date have not flowered. Therefore the presence 1972] WEAVER, REVISION OF LISIANTHIUS 295 TABLE 3. Summary of experimental crosses between Lisianthius cordifolius and L. longifolius Cross POLLEN SEED ToTAL NO. TOTAL NO. AV. NO. : PARENT PARENT OF FLOWERS OF SEEDS OF SEEDS 1 L. longifolius L. longifolius 10 2515 2515 (Weaver 2344) (Weaver 2344) 2 ~=L. longifolius L. cordifolius 45 11274 250.4 (Weaver 2344) (Weaver 2343) 3. L. cordifolius L. cordifolius 9 3455 383.9 (Weaver 2343) (Weaver 2343) or absence of an effective barrier to hybridization, resulting in reduced pollen or ovule viability, between these two species cannot now be deter- mined. Extensive field and experimental work must be done before the relation- ship between Lisianthius cordifolius and L. longifolius can be explained. However, the morphological evidence stongly suggests that they are dis- tinct taxa. Since the taxa are rather closely sympatric, with obvious in- termediates absent, I prefer to treat them as distinct species. b. Lisianthius sect. Lisianthius subsect. Fruticosi ser. Exserti Weaver, ser. nov. Frutices vel arbusculae graciles, omnino glabrae. Styli filamenta stami- numque longissimi exserta, segmentis exsertis corolla longior. Pollinis grana exinio laevi vel fere laevi. SPECIES TYPICA: L. exsertus Sw. 500 53F I> 25. Lisianthius exsertus Sw. Prodr. 40. 1788; Fl. Ind. Occ. 1: 346. 1797 (as Lisianthus).“LectotyPE. Jamaica, without exact locality, Swartz (s!). Leianthus exsertus (Sw.) Griseb. Gen. & Sp. Gent. 197. 1839 [1838]. Lisianthus callosus Bertero in Spreng. Syst. 1: 585. 1825 [1824]. _ Tachia swartzii G. Don, Gen. Hist. 4: 197. 1837, nom. illegit. (Art. 63, LC.BN.). Slender shrubs, rarely becoming treelike, glabrous throughout. Stems 1-6 m. tall and to 6 cm. in diameter at the base, erect or ascending, gray-brown at the base and green above, terete, brittle, distinctly woody. Leaves membranaceous, glossy bright green above, paler and dull beneath, the principal lateral veins evident but often not conspicuous, the blades 3-13.4 cm. long and 1—5.5 cm. broad, elliptic or elliptic-oblong, with a ten- dency to be widest slightly above the middle, gradually short acuminate or rarely acute, the bases sometimes unequal, attenuate into the 0.5-2.6 cm. long petiole. Inflorescences of pedunculate, apparently simple dichasia : flowering branches once or twice ternately divided, each division termi- nated by a dichasium; stalks of the flowering branches ascending, flat- 296 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 tened in cross section; dichasial branches terete. Flowers nodding, the pedicels 6-11 mm. long; bracteoles subulate, scarious. Calyx 5-11 mm. long, fused for ca. 1/4 its length; lobes 5—-8.5 mm. long and 1-1.5 mm. broad, lanceolate or subulate-lanceolate, short-acuminate, mostly scarious but with a narrow green band in the center, slightly ridged abaxially; tube 1-2.5 mm. long. Corolla tubular-campanulate, 2—-3.3 cm. long; tube 1—1.6 cm. long and 4-8 mm. in diameter at the widest point, yellow, con- stricted for about half its length; lobes 2.5-5.5 mm. long and 2—4.5 broad, nearly as broad as long, 1/6-1/9 as long as the tube, yellow-green, erect or slightly spreading, broadly triangular-ovate, abruptly short- acuminate. Stamens inserted just below the middle of the corolla tube; filaments 3.5—6.6 cm. long, greatly exserted, exceeding the corolla lobes by 2.5—5 cm.; the longest one usually equalling or surpassing the style; anthers 2.5—3.5 mm. long and ca. 1 mm. broad, narrowly oblong, narrowed at the apex, the connective slightly exserted. Styles 3.4-6.7 cm. long, greatly exserted; stigma peltate. Capsule oblong-ellipsoid, long-beaked, 9-14 mm. long, not including the 2.5—-5 mm. long beak, and 3-5 mm. in diameter; the marcescent corolla not persisting until the seeds are shed. FLOWERING: All year. CHROMOSOME NUMBER: 1 = 18. DIsTRIBUTION: From 300 to 1100 meters in secondary thickets or scrub, or along roadsides, or rarely in elfin woodland, in the eastern and central parishes of Jamaica (Map 12). REPRESENTATIVE COLLECTIONS Jamaica: without locality, Swartz wan, S; isolectotypes). St. ANDREW: vicinity of Cooper’s Hill, Red Hills, Proctor 18262 (a, 13); vicinity of Brandon Hill, Proctor 23499 (13, Be ip Moresham River, west side, 1/4 mi. above Her- of Silver Hill Gap, along road from Section to Silver Hill ‘hs Weaver & Weaver 951 (A, DUKE, Ucw1); Proctor’s Pool, ag, 9151 (DUKE); trail n. & w. of Silver Hill Gap, Hespenheide 773 (DUKE, GH, MICH); east slope of the John Crow Mountains, 1.5-2.5 mi. sw. of Eeeadook. oak 22114 (IJ, MICH). Swartz’s herbarium presently at Stockholm (s), includes three speci- mens of Lisianthius exsertus, two with the inscription “Jamaica” and one with “Ind. occ.’”’ There is also a Swartz specimen of the same species, with the inscription “Jamaica,” among the collections at the British Mu- seum (BM). The more complete of the two inscribed “Jamaica” at Stock- holm best demonstrates the characters described by Swartz and is here designated the lectotype of L. exsertus Sw. Although I have seen neither the type nor any other authentic material of Lisianthus callosus Bertero, the original description suggests very strongly that the plant is the same as L. exsertus. This description is re- 1972] WEAVER, REVISION OF LISIANTHIUS 297 produced below in its entirety (from Bertero in Spreng. Syst. 1: 585. 1825). callosus Berter. 2. L. foliis basi inaequalibus acutis, petiolis calloso annulo ramos amplectentibus, pedunculis corymbosis, staminibus longissimis. Jamaica. (Flores ochroleuci. ) This diagnosis fits very well the description of Lisianthius exsertus and excludes all other known Jamaican species. The only other species with stamens that could possibly be described as “longissimis” is L. capitatus, but that species has subcapitate inflorescences. The primary difference between the diagnosis of L. callosus and that of L. exsertus, immediately above it in Sprengel’s treatment, is the phrase “. . . petiolis calloso annulo ramos amplectentibus ...” The epithet “callosus’” obviously refers to this character, which rather than being diagnostic for any single species, would describe equally well every species in the genus. George Don (1837) cited the earlier Lisianthius exsertus Sw. as a synonym of his Tachia swartzii, the specific epithet of the latter there- fore illegitimate (Art. 63, I.C.B.N.). He included two other species of Lisianthius in Tachia [T. longifolia (L.) Mart. = L. longifolius L. and T. cordifolia (L.) Mart. = L. cordifolius L.| but did not change their epithets. c. Lisianthius sect. Lisianthius subsect. Fruticosi ser. Umbellati Weaver, ser. nov. Frutices vel arbores graciles, omnino glabri. Folia oblanceolata vel anguste obovata. Inflorescentiae axillares, dense compactae, capitula vel umbellas simulantes, paribus duobus bractearum involucratarum. Styli filamentaque longe exserta. Pollinis grana exinio reticulato, SPECIES TyPica: L. umbellatus Sw. KEY TO THE SPECIES OF SERIES UMBELLATI Pedicels in fruit 8-20 mm. long; corolla tubular, the tube 2.2-3.1 cm. long and 4-6 mm. in diameter at the broadest point, the lobes abruptly short-acuminate, 2-3.5 mm. long and 2-3 mm. broad, about as broad as long, 1/8-1/10 as long as the tube; filaments exceeding the corolla lobes by no more than 8 mm.; styles exceeding the corolla lobes by no more than 1.5 cm.; calyx 4.5—-7 mm. long, the lobes 2—4.5 mm. long. ......6.-¢ 666-50 e see: 26. L. umbellatus. Pedicels in fruit to 7.5 mm. long, or the fruits sessile; corolla tubular-campanu- late, the tube 1.5-2.3 cm. long and 7.5-10 mm. in diameter at the broadest point, the lobes acute to obtuse, 4.5-7.5 mm. long and 3.5-6 mm. broad, 1/4-1/3 as long as the tube; filaments exceeding the corolla lobes by 8-20 298 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 53 mm.; style exceeding the ost lobes by 1.1-3.4 cm.; paste 6.5-12.5 m lene. the lobes 4-35 mim, lone? sii. een i fae es pancny 26. Lisianthius umbellatus Sw. Prodr. 40. 1788; Fl. Ind. Occ. 1: 350. 1797 (as Lisianthus). YLECTOTYPE. Jamaica, without exact locality, Swartz (s!). Leianthus umbellatus (Sw.) Griseb. Gen. & Sp. Gent. 199. 1839 [1838]. Glabrous, spindly, unbranched or sparsely branched shrubs, often be- coming treelike. Stems erect or ascending, 1.5—6 m. tall and to 6 cm. in diameter, brittle, gray-brown below and green above, terete. Leaves firmly membranaceous or subcoriaceous, grass-green and glossy above, paler and dull beneath, the blades (7)10-29 cm. long and (2)3—9.5 cm. broad, ob- lanceolate to narrowly obovate (elliptic), abruptly or gradually short- acuminate, the bases long-attenuate into the 1.1—4.1 cm. long petiole. Inflorescences appearing umbellate, the primary ae secondary axes greatly contracted, closely subtended by 2 pairs of opposite, decussate, leafy bracts, the outer pair broadly ovate or suborbicular, rounded or broadly cuneate at the base, 1.8-7.5 cm. long and 1.5—5 cm. broad, the inner pair narrower, elliptic or tending toward oblanceolate, more attenuate at the base, 2—6.2 cm. long and 0.7—2.4 cm. broad; the stalks ascending, flattened in cross section, 7.2-19.4 cm. long. Flowers ascending or horizontal, pedicellate. Calyx pale green, and subscarious, fused for 1/3-1/2 its length; lobes 24.5 mm. long and 1—1.5 mm. broad, lanceolate or subulate, acuminate, ecarinate; tube 2—2.5 mm. long. Corolla tubular, 2.5-3.4 cm. long; tube bright yellow, constricted in the lower half, somewhat inflated and widest below the apex; lobes yellow-green, erect or slightly spreading, orbicular, abruptly short-acuminate. Stamens inserted just below the middle of the corolla tube; filaments 1.7-2.9 cm. long, well exserted; anthers 2—3.5 mm. long and ca. 1 mm. broad, more than twice as long as broad, narrowly oblong, narrowed at the apex, the connective slightly exserted. Style 2.7-3.7 cm. long, well exserted, always exceeding the filaments, exceeding the corolla lobes by up to 1.5 cm.; stigma peltate. Capsule oblong-ellipsoid, long beaked, 1.0—-1.5 cm. long, not including the 1.5—2.5 mm. long beak, and 3-5 mm. in diameter; the marcescent corolla not persisting until the seeds are shed. FLowERING: Feb.—Apr. CHROMO- SOME NUMBER: ” = 18. DIstTRIBUTION: In wet limestone thickets at ca. 600 meters; known from 2 localities, on Dolphin Head Mountain and at the edges of the Cockpit Country in northwestern Jamaica (Map 13). REPRESENTATIVE COLLECTIONS Jamaica: near Troy, Perkins 1321 (cH). HANovER: summit of Dolphin Head ie near the village of Askenish, Weaver 1832 (A, BM, DUKE, F, MO, ucwI, us); slopes of Dolphin Head Mountain, Harris 10292 (#, Ny, US). Swartz did not designate a type specimen of Lisianthius umbellatus. His herbarium, now at Stockholm (s), includes three specimens of this 1972] WEAVER, REVISION OF LISIANTHIUS 299 species, two inscribed “Jamaica” and one inscribed “Ind. occ.” The more complete of the former is designated lectotype of L. umbellatus Sw. Several authors before 1909 included under Lisianthius umbellatus plants now referred to L. capitatus. Grisebach (1862) gave the distribu- tion of L. umbellatus as “in high mountains Portland, Guy’s Hill s. Anns, Dolphin westmoreland.” Similarly, Perkins (1902) cited specimens as typical of L. umbellatus from both eastern and western Jamaica, including Harris 6683, from St. Ann, which was later cited by Urban in the original description of L. capitatus. Although in rather poor condition, the lectotype specimen of Lisianthius umbellatus, with its long pedicels, short calyx, and short, orbicular corolla lobes clearly resembles the plants presently known only from Dolphin Head Mountain in Hanover Parish and from near Troy at the edge of the Cockpit Country in western Jamaica. 27. Lisianthius capitatus Urb. Symb. Antill. 6: 33. 1909 (as Lisian- thus). Lectotype. Jamaica. St. ANN: Hollymount, near Ewarton, Harris 8878 (BM!). Glabrous, spindly, unbranched or sparsely branched shrubs, often be- coming treelike. Stems erect or ascending, 1.5—6 m. tall and to 7 cm. in diameter, gray-brown below and green above, terete, brittle. Leaves firmly membranaceous or subcoriaceous, dark glossy green above and paler beneath, the blades (6) 12-31 cm. long and (1.5)3.5—6.5 cm. broad, oblan- ceolate (elliptic), abruptly or gradually short-acuminate, the bases long- attenuate into the 0.7-3.7 cm. long petiole. Inflorescences appearing capitate or umbellate, primary and secondary axes contracted, closely subtended by 2 pairs of opposite, decussate, foliaceous bracts, these elliptic or ovate, rarely orbicular, broadly attenuate or rarely rounded at the base, the outer pair 1.9-8.3 cm. long and 1.5—7.1 cm. broad, the inner pair smaller and narrower, 1.5—7.1 cm. long and 0.4-3.6 cm. broad; stalks ascending, flattened in cross-section, 7.0—23 cm. long. Flowers erect, ses- sile or short-pedicellate. Calyx pale green and subscarious, fused for 1/3-1/2 its length: lobes 4-8.5 mm. long and 1-2 mm. broad, lanceolate or subulate, long-acuminate, ecarinate; tube 2.5-4 mm. long. Corolla tubular-campanulate, 1.8—2.8 cm. long; tube yellow or yellow-green, con- stricted to slightly above the middle; lobes yellow-green, erect or slightly spreading, broadly ovate or rarely suborbicular, the apices obtuse or more rarely acute. Stamens inserted just above the middle of the corolla tube; filaments 1.9-3.2 cm. long, greatly exserted; anthers 3.5—-4.5 mm. long and ca. 1 mm. broad, more than 3 times longer than broad, narrowly oblong, narrowed at the apex, the connective slightly exserted. Style 2.7— 5 cm. long, greatly exserted, surpassing the corolla lobes by 1.1-3.4 cm., always exceeding the filaments; stigma peltate. Capsule oblong-ellipsoid, long-beaked, 10-17 mm. long, not including the 2-3.5 mm. long beak, and 3-8 mm. in diameter; the marcescent corolla not persisting until the the seeds are shed. FLowERINc: Jan.—Sept. CHROMOSOME NUMBER: n= 18. 300 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 DISTRIBUTION: From sea level to 800 meters, at edges of forests and in moist thickets on limestone, rarely on rocky headlands by the sea, in the eastern and central parishes of Jamaica, principally in the John Crow Mountains, the Mt. Diablo Massif, and the Cockpit Country (Map 13). REPRESENTATIVE COLLECTIONS Jamaica. PorTLAND: vicinity of Blue Hole, Weaver & Weaver 940 (a, DUKE, F, UCWI); John Crow Mountains, east slope, 1.5 mi. w. of Ecclesdown, Webster D Park Road, Weaver & Weaver 1291 (A, DUKE, UCWI, us). St. ANN: prope Moneague ad Guy’s Hill, Alexander (ny, cited collection); Hollymount, near Ewarton, Harris 8878(¥, Ny; isolectotypes); road from Jamaica A-1 to Hollymount on Mt. Diablo, Weaver & Weaver 983 (A, BM, DUKE, F, IJ, LL, MICH, MO, NY, S, UC, Us). St. THomAs: John Crow Mountains, Big Level Area, Webster & Proctor 5524 (a, IJ, MICH, US). TRELAwNy: Cockpit Country, roa from Burnt Hill to pagan Bottom, Weaver & Weaver 1002 (a, BM, DUKE, F, LL, MICH, MO, NY, §, I, US); vicinity of Ramgoat Cave, Cockpit Country, Howard, Proctor & ‘Sfeaed 2 14684 (BM, IJ). Although not described until 1909, Listanthius capitatus has been part of various collections made since the middle of the 19th century that were determined as L. umbellatus. It is by far the more common of the two species in series Umbellati. Judging from the determination of some of the more recent collections, it is apparent that the species are still some- times confused. Vegetatively L. capitatus is very similar to L. umbellatus. However, the flowers of the two species are strikingly different. As pointed out by Urban in the original description of L. capitatus, that species differs from L. umbellatus: in its flowers sessile or short-pedicellate; its longer calyx; its shorter and much broader corolla; its corolla lobes acute or obtuse, longer than broad, longer in absolute length and much longer in relation to the tube; and in its more greatly exserted styles and filaments. Urban, in describing Lisianthius capitatus cited three specimens as representative of his new species. These, which are syntypes, are: Harris pie and Harris 6683, from Hollymount on Mt. Diablo, and Alexander “prope Moneague ad Guy’s Hill.” All of Urban’s original material posses for some Ekman specimens at Stockholm (s)] was destroyed during the bombing of Berlin in World War II (H. Scholz, in litt.). Sev- eral duplicates of the above-mentioned specimens have been seen. Of these, the specimen of Harris 8878 at the British Museum is the best preserved and best exhibits the characters described by Urban. This specimen is entirely typical of those from the Mt. Diablo area and un- doubtedly represents the taxon described by Urban. Therefore, Harris 8878 (BM) is designated the lectotype of L. capitatus Urb. SPECIES NON Satis Nota 1. Lisianthius calciphilus s Atandl. & Steyerm., Fieldiana Bot. 22: 267. 1940 (as Lisianthus).”~Type. Guatemala. Attra VeERAPAz: Finca Los Alpes, Wilson 356 (¥F!). 1972] WEAVER, REVISION OF LISIANTHIUS 301 Low shrubs (?), glabrous throughout. Stems terete below and 4-angled above. Leaves firmly membranaceous, papillose above and on the margins, subsessile or short-petiolate, narrowly ovate, the apices acuminate, the bases cuneate or rounded. Inflorescences of compound dichasia, but the terminal flower aborted and the lateral branches then appearing to be complete dichasia; branches of the inflorescence strongly 4-angled, the angles expanded into narrow, scarious wings. Bracteoles oblanceolate, green at least along the midline, to 8 mm. long. Calyx ca. 7 mm. long; lobes ca. 5 mm. long, lanceolate, long-acuminate, scarious margined, carinate abaxially. Corolla tubular-funnelform, 3.6-3.9 cm. long, yellow; lobes ca. 6 mm. long and 3 mm. broad, 1/5 as long as the tube, spreading, ovate-lanceolate, abruptly short-acuminate. Filaments 2.2-2.6 cm. long, not exceeding the corolla lobes; anthers broadly oblong, obtuse. Style 2.9-3.4 cm. long, exceeding the corolla lobes; stigma peltate. Capsules not seen. Lisianthius calciphilus is definitely referable to the typical section of Lisianthius as here interpreted. Only the type specimen is known, and it cannot be placed in any of the species recognized in this study. There- fore, L. calciphilus is very possibly a good species. The type specimen, Wilson 356 from Dept. Alta Verapaz, Guatemala, consists of two flowering shoots. However, it is impossible to determine whether these shoots repre- sent primary or secondary (lateral) axes. They are determinate, but not monopodial, and the branching is not excurrent. Therefore, if these shoots represent primary axes, the plant would not fit into either subsect. HEr- BACEI or subsect. Fruticosi, but would rather appear to be somewhat intermediate between them. If the shoots represent lateral axes, they would be similar to the flowering branches of certain species, particularly L. longifolius or L. seemannii, of subsect. FRuTICosI, series Longifolit. In the original description of Lisianthius calciphilus, Standley and Steyermark suggest that their new plant is close to L. collinus. (= L. brevidentatus var. collinus, subsect. HERBACEI). From that plant L. cal- ciphilus differs in its longer corolla lobes, broader corolla tube, and much larger bracteoles. Williams (1968, 1969) also associated L. calciphilus with LZ. collinus, reducing both to synonymy under L. brevidentatus. In several respects, e.g. flower form, papillose leaves, and subfoliaceous brac- teoles, Wilson 356 closely resembles L. acuminatus. However if the struc- tures preserved are indeed lateral flowering branches, they are much more greatly proliferated than in that species and would more closely resemble those of L. longifolius or L. seemannii. It seemed possible that Wilson 356 might be a hybrid between L. acumi- natus and a species in subsect. HERBACEI, perhaps L. quichensis or L. brevi- dentatus, thus possibly explaining its peculiar inflorescences. However, its pollen showed nearly 100 percent stainability in cotton blue-lactophenol. Until more material is available, or until I have studied it in the field, I cannot place Lisianthius calciphilus with any degree of confidence in my classification of the species of Lisianthius and therefore must treat it as a species non satis nota. 302 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 EXCLUDED NAMES Lisianthus acuminatus Desr. in Lam. Encycl. 3: 660. 1804. “Type, Guadeloupe: de Badier (Pp) (Microfiche, IDC 6207. 446: II. me = L. frigidus Sw. fide Grisebach (1839) = Calolisianthus frigidus (Sw.) Gilg a acutangulus Mart. Flora 21 (2-Beibl.): 49. 1838. Type: Martius 216 = Calolisianthus acutangulus (Mart.) Gilg in Engler & Prantl, Nat. acres 4(2): 101. 1895. Lisianthus eps ate Ruiz & Pav. Fl. Peruv. 2: 14. ¢. 122, figure a. 1799, TYPE: “Tn atahuarum montibus ad Chinchao runcationes in locis meridionali- bus,” Ruiz & Pavon (MA) = Chelonanthus alatus (Aubl.) Pulle fide Williams (1968). Lisianthus acutilobus Steyerm. Bol. Soc. Venez. Cienc. Nat. 26: 439. figure 9. 1966. Type: Steyermark & Nilsson 732 (vEN); referable to Pagaea Griseb. Lisyanthus alatus Aubl. Hist. Pl. Guiane 1: 204; 3: ¢. 80. 1775. Type: “Habitat : on incultis Caienne & Guiane,”’ Aublet (P) (Microfiche, IDC 6213. 2: 6!) = Chelonanthus alatus (Aubl.) Pulle, Enum. Vasc. Pl. Surinam a6. ve Lisyanthus alatus Willd ex Griseb. Gen. & Sp. Gent. 185. 1839 (nom. nud.) = L. trifidus HBK. — Grisebach (loc. cit.); ex char. referable to Chelonan- thus (Griseb.) Gilg Lisianthus albus Spruce ex Prog. in Mart. Fl. Bras. 6(1): 237. 1866. Type: “In — Barra prov. do Alto Amazonas [Brazil],” Spruce = Chelonanthus pe a eas ex Prog.) Badillo in Pittier et al., Cat. Fl. Venezolana 2: img part weee. Mart. Nov. Gen. & Sp. 2: 93. ¢. 171. 1827 = Helia alpestris ( .) O. Ktze. Rev. Gen. Pl. 2: 428. 1891. This should be transferred to Calolisianthus (Griseb.) Gilg, according to Nilsson (1970) Lisianthus amoenus Miq. Linnaea 19: 139. 1847. Type: “Crescit in Para supe- riore juxta sylvam [Brazil],” Focke (U?) = Helia amoena (Miq.) O. Ktze. ee Gen. Pl. 2: 428. 1891; ex char. referable to Chelonanthus (Griseb.) ilg. ene ie am/plissimus Mart. Nov. Gen. & Sp. 2: 96. ¢. 175. 1827 = Calolisian- mplissimus (Mart.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 101. 1895. mimes angustifolius HBK. Nov. Gen. & Sp. 3: 181. 1818. Type: “In radicibus montis Duidae, prope Esmeraldam [Venezuela],” Humboldt & Bonpland (P) ‘Glas IDC 6209. 72: III. 3!) = Chelonanthus angustifolius (HBK.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. 1895. Lisianthus angustifolius Mart. Nov. Gen. & Sp. 2: 99. t. 178, figure 1. 1827 = . tenuifolius Spreng. Syst. Cur. Post. 339. 1827. According to Nilsson (1970) this should be transferred to Calolisianthus (Griseb.) Gilg. Lisianthus anomalus HBK. Nov. Gen. & Sp. 3: 184. 1818. Type: ‘“Crescit in montosis inter ev et Mariquita Novo-Granatensium,” Humboldt & Bonpland (P) = Symbolanthus anomalus (HBK.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 99. 1895. 1972] WEAVER, REVISION OF LISIANTHIUS 303 wae aphyllus Vell. Fl). Flum. 2: 78. 1825. From the illustration this is a s of Voyria Aubl. voles arboreus Spruce ex Prog. in Mart. Fl. Bras. 6(1): 240. t. 64, figure 1. . Type: Spruce 3551 (K) = Adenolisianthus egg A aes ex Prog.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. estat: auriculatus Benth, Pl. Hartweg. 226. 1846. Type: “ate 1240 (Kk) = auriculata (Benth.) O. Ktze. Rev. Gen. Pl. 2: 428. 1891. iniaeees Baltae (Weberb. & Gilg) Macbride, Fieldiana Bot. am 282. 1959 = Symbolanthus Baltae Weberb. as ie Fedde, Repert. Sp. Nov. 2: 56. 1906. Type: Weberbauer 4737 (B? ems to3 bifidus HBK. Nov. Gen. & * 3: 181. 1818. Type: “Crescit in mon- tibus Novo-Granatensibus, prope fodinam Santanna,” Humboldt & Bon- bland (P) (Microfiche, IDC 6209. 72: III. 4!) = Chelonanthus bifidus (HBK.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. 1895. Lisianthus breviflorus Benth. Hook. Jour. Bot. 2: 45. 1840. Type: Guiana, Sierra Mey, heme adalat (kK). According to Bentham (1826) and Nilsson fi 1970) this should be transferred to Pagaea Griseb. reap vis brevifolius (Cham.) Griseb. Gen. & Sp. Gent. 187. 1839 = Helia brevifolia Cham. Linnaea 8: 11. 1833. TYPE: aes [Brazil],” Seliow pee Brittonii Woodson, FI. Trinidad & Tobago 2: 183. 1947. Fide ae oe Her »S. Nat. Herb. 29: 221. 1948 = Chelonanthus arboreus Britt. Bull. t. Agric. Trinidad & Tobago 19: 230. 1922.* Type: Britton, Hazen & Metehes 1295 (NY) = Macrocarpaea arborea (Britt.) Ewan, loc. cit. “Lisianthus browallivides Ewan, Proc. Biol. Soc. Wash. 64: 1951. TYPE: m 4932 (MO) = Macrocarpaea browallioides (Ewan) ony & Nilsson, Bull. Jard. Bot. Nat. Belg. 40: 13. 1970. Lisyanthus caerulescens Aubl. Hist. Pl. Guiane 1: 207; 3: ¢. 82. 1775. Type: “Habitat in pratis humidis Guianae,” Aublet (P) Mic once IDC 6213. 23: IV. 4!) = Irlbachia coerulescens (Aubl.) Griseb. Gen. & Sp. Gent. 195. 1839. peels ees Ruiz & Pav. Fl. Peruv. 2: 14. ¢. 126. 1799. Type: “In s frigidis Acomayo, Mufia et Pallao, ad Chacahuassi et Achapatu- man pe a ” Ruiz & Pavon (MA) = Symbolanthus calygonus (Ruiz & Pav.) Gilg in Engler & Prantl, Nat. ae: 4(2): 99. 1895. Lisianthus campanulaceus Desr. in Lam. Encycl. 3: 660. 1804. Type: “Guyane,” Richard (p-jv) srongonane “ei 6206. 512: III. 3!) = Hela cam- panulacea (Desr.) O. Ktze. Rev. Gen. Pl. 2: 428. 1891; referable to Chelonanthus (Griseb.) Gilg. Lisianthus campanuloides Spruce ex Benth. Jour. Bot. Kew Misc. 6: 201. . TYPE: gona bank of Rio —o opposite Uananaca [Brazil], Spruce (Kk) = nanthus campanul a oe ex Benth.) Gilg in Engler & Prantl, won Pflanzenfam. rp "98. — cardonae (Gleason) Steyerm. Fieldiana Bot. 28: 498. 1953 = Chelo- nthus cardonae Gleason, Brittonia 3: 189. 1939. Type: Tate 1359 (Ny). Re onteg to Nilsson (1970) this should be transferred to Pagaea Griseb. * Reference not seen. 304 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Lisianthus carinatus Desr. in Lam. Encycl. 3: 658. 1804 = Tachiadenus cari- natus (Desr.) Griseb. Gen. & Sp. Gent. 201. 1839. Lisianthus chelonoides L.f. Suppl. Pl. 134. 1781 oe illegit., Art. 63) = alatus Aub]. = Chelonanthus alatus (Aubl.) Pul Lisianthus chimantensis Steyerm. & Maguire, Mem. N.Y. Bot. Gard. 17(1): 461. 1967. TyPE: Steyermark & Wurdack 374-A (Ny). According to Nils- son (1970) this does not fit into any existing genus. Lisianthus corymbosus Ruiz & Pav. Fl. Peruv. 2: 14. ¢. 124. 1799. Type: “In ruviae nemoribus Pei Acomayo et Chinchao ad Pati praedium,” Ruiz & Pavon (MA) = pyerierarin ramones (Ruiz & Pav.) Ewan, Contr. U.S. Natl. Herb. Pe 242. “ Lisianthus crassicaulis Mart. & Gal. Bull. Acad. Brux. 11(1): 374. 1844. TYPE: Galeotti 7176 (BR) = Chelonanthus alatus (Aubl.) Pulle fide Williams (1968) Lisyanthus daturoides Griseb. Gen. & Sp. Gent. 177. 1839. Type: Mathews 1317 K) = Symbolanthus daturoides (Griseb.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 99. 1895, “Lisianthus densiflorus Benth. Pl, Hartweg. 227. 1846. Type: Hartweg 1241 (K) = Macrocarpaea densiflora (Benth.) Ewan, Contr. U.S. Natl. Herb. 29: 232. 1948. Lisianthus elegans Mart. Nov. Gen. & Sp. 2: 98. #. 177. 1827 = L. peduncula- tus Cham, & Schlecht. = Calolisianthus pedunculatus (Cham. & Schlecht.) Gilg. Lisianthus Elisabethae (Schomb.) Griseb. Linnaea 22: 40. 1849 = Leiothamnus Elisabethae Schomb. Verh. Ver. Beférd. Gartenb. KGnigl. Preuss. Staaten 18: 155. ¢. 1. 1847 = Symbolanthus Elisabethae (Schomb.) Gilg, Bot. Jahrb. 22: 340. 1896. oo Benioe Poir. in Lam. Encycl. Suppl. 3: 514. 1816 = Chelonanthus us (Poir.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. oe Lisianthus frigidus Sw. Prodr. 40. 1788; Fl. Ind. Occ. 352. 1797. Type: “In- colit regionem muscosam montis la Soufriere Insulae Do minicae,” Pon- thieu (s!) = Calolisianthus frigidus (Sw.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 101. 1895, Lisianthus glaber L.f. Suppl. Pl. 134. 1781. Type: Mutis ; pam 213.2 (LINN) Sella 1945); (Microfiche, IDC 5057. 118, II. 3!) = Macro- Se ra (L.f.) Gilg in Engler & Prantl, Nat. at eee 4(2): 94. 13895. Lisianthus iene Jacq. Coll. 1: 64. 1787 = Eustoma exaltatum fide Grise- bach (1845). Lisianthus glaucifolius Nutt. Trans. Am. Phil. Soc. 5: 197. 1837 = Eustoma Russellian um fide Grisebach (1845). Lisyanthus gracilis Griseb. Gen. & Sp. Gent. 182. 1839. Type: St. Vincent, ilding (K) = L. uliginosus Griseb. fide Index Kewensis = Chelonanthus inosus (Griseb.) Gilg 1972] WEAVER, REVISION OF LISIANTHIUS 305 Lisyanthus grandiflorus Aubl. Hist. Pl. Guiane 1: 205; 3: ¢. 81. 1775. TYPE: “Habitat in locis humidis Caienne & Guiane,” Aublet (Pp) (Microfiche, IDC 6213. 23: IV. 5!) = Chelonanthus grandiflorus (Aubl.) Chod. & Hassl. Bull. Herb. Boiss. Ser. 2. 3: 549. 1903. Lisyanthus grandiflorus Willd. ex Griseb. Gen. & Sp. Gent. 182. 1839 (nom. nud.) = L. igen Griseb. fide Grisebach (loc. cit.) = Chelonanthus uliginosus (Griseb. Lisianthus imthurnianus spite Trans. Linn. Soc. Ser. 2. 2: 279. 1887. TYPE 6 (K sue lisianthus imthurnianus (Oliver) Gleason, Bull Torr. Bot. Club 56: 4 9. Lisianthus incarnatus Sessé & Moc. Fl. Mex. ed. 1. 39. 1893. Ex char. not re- ferable to Lisianthius. Lisianthus inflatus Mart. Nov. Gen. & Sp. 2: 95. ¢. 174, 1827 = L. speciosus Cham. & Schlecht. hie Grisebach “ 1839) = Calolisianthus etches (Cham. & Schlecht.) Gilg. Lisianthus Kunthii G. Don, Gen. Hist. 4: 208. 1837 (nom. illegit., Art. 63) = L. angustifolius HBK. = Chelonanthus angustifolius (HBK.) Gilg Lesyantiues leucanthus Lyall ex Griseb. Gen. & Sp. Gent. 202. 1839 (nom. nud.) = Tachiadenus longiflorus Griseb. loc. cit. Lisyanthus longiflorus Boyer ex Griseb. oe p sae, Gent. 202. 1839 (nom. nud.) = Tachiadenus longiflorus Griseb. loc. Lisianthus luteus Raf. Atl. Jour. 146. 1832 = Eustoma Russellianum fide Mer- rill (Index Rafinesquianum 192. 1949.) cag ale macranthus Benth. Pl. Hartweg. 144. 1844.Type: Hartweg ae meg anthus macranthus (Benth.) Moldenke, Phytologia 2::235. ee ere Y HBK. Nov. Gen. & Sp. 3: 183. 1818. Type: “Crescit locis subfrigidus, in declivitate Parami de Almaguer, inter Pansitara a flu- ie Humboldt & Bonpland *) (Microfiche, IDC 6209. 72: III. 6!) = pc "4(2): 94, 1895. Lisianthus madagascariensis Spreng. Syst. 1: 586. 1825 (nom. illegit., Art. 63) = L. trinervis Dest. = Tachiadenus trinervis (Desr.) Griseb. geste Martii Griseb. Gen. & Sp. Gent. 187. 1839 (nom. illegit., Art. 63) = . spathulatus HBK. = Helia spathulata (HBK.) Gilg. . Mathewsii Griseb. Gen. & Sp. Gent. 361. 1839. TyPE: Mathews 6 (Kk) = Symbolanthus Mathewsii (Griseb.) ie Bot. Jahrb. 22: 344. sae Lisianthus nerioides Griseb. Linnaea 22: 39. 1849. TyPE: Moritz 1189 (8, but presumably a = Symbolanthus nerioides (Griseb.) Ewan, Proc. iol. Soc. Wash. 63: 165. 1950 sage oblongifolius (Mart, Griseb. Gen. & Sp. ge 187. 1839 = Helia oblongifolia Mart. Nov. Gen. & Sp. 2: 123. t. 191. 1827. Liye obtusifolius Griseb. Gen. & Sp. Gent. 175. 1839. Type: Brazil, Sierra d’Estrella, Sellow (K) = Macrocarpaca 0 obtusifolia (Griseb.) Gilg in ae & Prantl, Nat. Pflanzenfam. 4(2): 94. 1895. 306 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 “Lisianthus Oerstedtii Griseb. Kjoeb. Vidensk. Meddel. 1853: 54. 1854. Type: oxer mellem Matagalpa og Muimui; Provincem agin ” Oersted = Helia Oerstedtii (Griseb.) O. Ktze. Rev. Gen. Pl. 2: 428. 1891. Lisianthus Ophiorriza Vell. Fl. Flum. 2: 78. 1825 = Metternichia principis Mikan ee). om Sampiro & Peckolt, Arq. Mus. Nac. Rio de Ja- neiro 37: 331-394. reer ovalis Ruiz & Pav. Fl. Peruv. 2: 13. 1799. Type: “In Peruviae us per Chinchao runcationes,” Ruiz & Pavon MA) = Macrocar- aa sith (Ruiz & Pay.) Ewan, Contr. U.S. Natl. Herb. 29: 234. 1948. Lisianthus ovatifolius Vell. Fl. Flum. 4 79. 1825 = L. alpestris Mart. fide Pro- gel in Mart. Fl. Bras. 6(1): 233. 1866. pepe paniculatus Spreng. Syst. Cur. Post. 340, 1827 = Irlbachia elegans t. fide Progel in Mart. Fl. Bras. 6(1): 229. 1866. Lisianthus parvifolius Desr. in Lam. Encycl. 3: 660. 1804. Type: “Guyane,” Richard (p-yu) (Microfiche, IDC 6206. 512: III. 4!) = Irlbachia coeru- lescens (Aubl.) Griseb. fide Grisebach (1839). ase peach (Spruce ex Gilg) Macbride, Fieldiana Bot. 13(5): 289. = lanthus pauciflorus Spruce ex Gilg, Bot. Jahrb. 22: 340. shine hy Spruce 4429 (x), Lisianthus pauciflorus Sessé & Moc. Pl. N. Hispan. ed. 1. 24. 1888. Ex char. not referable to Lisianthius Lisianthus Esa Cham. & Schlecht. Linnaea 1: 199. 1826. Type: “In rasi interioribus regionibus,” Sellow (HAL) = Calolisianthus pedun- ulatus che: & Schlecht.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 101. 1895, Lisianthus pendulus Mart. Nov. Gen. & Sp. 2: 94. ¢. 172. 1827 = Calolisianthus pendulus (Mart.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 101. 1895, Lisianthus pratensis HBK. Nov. Gen. & Sp. 3: 180. 1818. Type: “Crescit locis gramineis prope Esmeraldem, in ripa Oronoci,” Humboldt & . (P) (Microfiche, IDC 6209. 72: III..2!) = Helia pratensis (HBK.) O. Ktze. Rev. Gen. Pl. 2: 428. 1891; ex char. appears to be referable to Chelonanthus (Griseb.) Gilg. Lisianthus princeps Lindl. Jour. Hort. Soc. 4: 261. 1849, Type: Schlim 419 (k) = Lage ~~ princeps (Lindl.) Gilg in Engler & Prantl, Nat. Pflan- zenfam. 4(2): 99. 1895, Lisianthus pulcher Hook. Bot. Mag. 75: ¢. 4424. 1849. Type: Colombia, Monte del Moro, Purdie (xk) = Purdiean oe ee (Hook.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 99. 1 fi ert heposabesiz _ Donn. Sm. in Pittier, Prim. Fl. Costaric. 2: 166. 189 bolan pulcherrimus Gilg; Bot. Jahrb. 22: 314. 1896. TYPE: pe Nic 673. Lisianthus pulcherrimus Mart. Nov. Gen. & Sp. 2: 97. t. 176. 1827 = Calolisian- thus pulcherrimus — Gilg in nore & Prantl, Nat. Pflanzenfam. 4(2): 101. 1895. 1972] WEAVER, REVISION OF LISIANTHIUS 307 ee pumilus Benth. Jour. Bot. Kew Misc. 6: 200. 1854. Type: Brazil, undated woods of the Rio Negro, near San Carlos, Spruce (K) = Pagaea aes (Benth.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 102. 1895. Lisyanthus purpurascens Aubl. Hist. Pl. Guiane 1: 201; 3: #. 79. 1775. TYPE: “Habitat in fissuris spa cn montis Serpent, ” Aublet (p) (Micro- fiche, IDC 6213. 23: IV. 3!) = purpurascens (Aubl.) O. Ktze. Rev. Gen. Pl. 2: 428. 1891; referable 4 priate (Griseb.) Gilg. Lisianthus purpurascens HBK. Nov. Gen. & Sp. 3: 182. 1818. Type: “Crescit locis inundatis, inter pagum San Fernando de Atabapo et cataractam May- purensium,” Humboldt & Boab (P) = L. rane Griseb. fide Grise- bach (1839) = Chelonanthus uliginosus (Griseb.) G Lisianthus ee (Gleason) Steyerm. Fieldiana Bot. es 498. 1953 = Chelo- rmis Gleason, Bull. Torr. Bot. Club 58: 450, 1931. TyPE: Tate 747 aay oe Hg quelchii N.E. Brown, Trans. Linn. Soc. Ser. 2. Bot. 6: 50. ¢. 9, figures 1901. SYNTYPEs: McConnell & Quelch 106, 649 (K) = Macrocarpaca sarees (N.E. Brown) Ewan, Contr. U.S. Natl. Herb. 29: 233. 1948. Lisyanthus racemosus Willd. ex Griseb. Gen. & Sp. Gent. 186. 1839 (nom. nud.) = L. angustifolius HBK. fide Grisebach (loc. cit.) = Chelonanthus angus- tifolius (HBK.) Gilg. Lisianthus ramosissimus Benth. Jour. Bot. Kew Misc. 6: 200. 1854. TyPE: Schomburgk 989 (kK) = Pagaea ramosissima (Benth.) Ewan, Proc. Biol. Soc. Wash. 63: 165. 1950. Lisianthus recurvus Benth. Jour. Bot. Kew Misc. 6: 201. 1854. TyPE: Brazil, nuré, on the Rio mveye — (k) = Pagaea recurva (Ben nth.) Benth. & Hook. Gen. Pl. 2: 814 — revolutus Ruiz & Pav. Fl. Peruv. 2: 14. ¢. 127. 1799. nhc “In ontibus altis frigidis Muna ad Saria,” Ruiz & Pavon (mA) = Macrocarpaea re a (Ruiz & Pav.) Gilg in Engler & Prantl, Nat. St ay 4(2): 94. 1895. Lisianthus revolutus Poepp. ex Griseb. in DC. Prodr. 9: 77. 1845 (nom. nud.) = L. glaber L.f. fide Grisebach (loc. cit.) = Macrocarpaea glabra (L.f.) Gilg. Lisianthus roseus Sessé & Moc. Fl. Mex. ed. 1. 39. 1893. Ex char. this is not referable to Lisianthius. Lisianthus Rusbyanus (Gilg) Macbride, Fieldiana Bot. 13(5): 290. 1959 = Symbolanthus Rusbyanus Gilg, Bot. Jahrb. 22: 344. 1896. TYPE: Rusby 1227. Lisianthus Russellianus Hook. Bot. Mag. 65: ¢. gi 1839 = Eustoma grandi- florum (Raf.) Shinners, Southwest. Nat. 2: 41. Lisianthus scabridulus Steyerm. Fieldiana Bot. 28: 496. 1953. Type: Steyermark 57979 SADR aca to Nilsson (1970) this should be transferred to Pagaea Gri ' Lisianthus Ee Griseb. in DC. Prodr. 9: 75. 1845. TYPE: Schombur gk 298 (K) elonanthus Schomburgkii gene Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. 1895. Lastaneies sempervirens Mill. ex Steud. Nomencl. ed. 1. 486. 1821 (nom. ae um sempervirens (L.) Ait. (Loganiaceae). 308 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Lisianthus spathulatus HBK. Nov. Gen. & Sp. 3: 181. 1818. Type: “Crescit locis calidissimis, humidis Insulae Pararumae, in flumine Oronoci,” Hum- boldt & Bonpland (P) = Helia spathulata (HBK. ) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 101. 1895. pate -aie speciosus Cham. & Schlecht. Linnaea 1: 198. 1826. Type: “Crescit asiliae regionibus interioribus,” Sellow (HAL) = Calolisianthus speciosus pee & Schlecht.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 101. 895. Lisianthus splendens Hook. Lond. Jour. Bot. 6: 264. 1847. Type: “Hills of red clay n i die ( attributed by Ewan (loc. cit.) to Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 101. 1895. Lisianthus Spruceanus Benth. Jour. Bot. Kew Misc. 6: 202. 1854. Type: “Brazil, in grassy inundated places near San Gabriel, Rio Negro,” Spruce (K) = Chelonanthus spruceanus (Benth.) Pilger in Koch-Griinberg, Zwei Jahre unter den Indianern 2: 31. 1910.* Lisianthus subcordatus Benth. Jour. Bot. Kew Misc. 6: 200. 1854, Type: “Brazil, near Panuré, on the Rio Udpes,” — (K) = Pagaea subcordata (Benth.) Benth. & Hook. Gen, Pl. 2: 814 Lisianthus Tatei (Gleason) Steyerm. Fieldiana Bot. 28: 498. 1953 = Caloisian- Tatei Gleason, Bull. Torr. Bot. Club 58: 449. 1931. Type: Tate 743 (ny). According to Nilsson (1970) this should be transferred to Pagaea Griseb. — tenuifolius Spreng. Syst. Cur. Post. 339. 1827. eee ig Nilsson 1970) this should be transferred to Calolisianthus (Griseb.) G Lisianthus tetragono-alatus Steud. Nomencl. ed. 2. 2: 55. 1841 ja aici nis L. acutangulus Mart. fide Progel ae Fl. Bras. 6(1): 234. 1866 = Calolisianthus acutangulus (Mart.) G “Lisyanthus tetragonus Benth. Pl. meh Pes 68. 1840. Type: Hartweg 496 (K) = Chelonanthus alatus (Aubl.) Pulle fide Williams (1968). es thamnoides Griseb. Fl. Brit. W. Ind. Isl. 424. 1862. Type: Jamaica adyen (K) = Macrocarpaea thamnoides (Griseb.) Gilg in Engler & Pest Nat. Pflanzenfam. 4(2): 94. 1895. Lisianthus trifidus HBK. Nov. Gen. & Sp. 3: 182. 1818. Type: “Crescit inter eat et fodinam Santanna, Novo-Granatensium,” Humboldt & Bon- (P) oo IDC 6209. 72: III. 5!); referable to Chelonanthus Came Lisianthus trinervis Desr. in Lam. Encycl. 3: 659. 1804. Type: (p-yu) (Micro- fiche, IDC 6206. 512: III. 1!) = Tachiadenus trinervis (Desr.) Griseb. Gen. & Sp. Gent. 201. 1839. Lniodiy trinervis Sessé & Moc. Fl. Mex. ed. 1. 39, 1893. Ex char. not referable o Lisianthius. * Reference not seen. 1972] WEAVER, REVISION OF LISIANTHIUS 309 Lisianthus tubiflorus Thou. ex Roem. & Schult. Syst. 4: 788. 1819, Type: “In Madagascar,” Willdenow (B) = Tachiadenus tubiflorus (Thou. ex Roem. & Schult.) Griseb. Gen. & Sp. Gent. 202. 1839. Lisyanthus uliginosus Griseb. Gen. & Sp. Gent. 181. 1839 = Chelonanthus uliginosus (Griseb.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. 1895. Lisianthus ro Griseb. Linnaea 22: 37. 1849. SyntypEs: Karsten 27, ritz 28 = Symbolanthus pe (Griseb.) Gilg in Engler & Prantl, Nat. Piscecdiar 4(2): 99. ioe virgatus Boj. ex Griseb. in i Prodr. 9: 82. 1845 (nom. nud.) = chiadenus gracilis Griseb. loc prate virgatus tee) in Mart. Fl. Bras. 6(1): 239. 1866. Type: In pascuis Ega prov. do Alto Amazonas Brazil,” Martius (m) = Adenolisianthus ofl a (Prog.) ne in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. 1895. Lisianthus oe Mart. Nov. Gen. & Sp. 2: 94. ¢. 173. 1827 = Chelonanthus iridiflorus (Mart.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 98. 1895 Lisianthus viscosus Ruiz & Pav. Fl. Peruv. 2: 14. ¢. 125. 1799. Type: “In a locis Mufia ad Tambo,” Ruiz & Pavon (mA) = Macrocarpaea viscosa (Ru & Pav.) Gilg in Engler & Prantl, Nat. Pflanzenfam. 4(2): 94. 1895. Lisianthus pane Salisb. Prodr. 137. 1796 = Gelsemium sempervirens (L.) Ait. (Loganiaceae). Lisianthus sia Spreng. Syst. 1: 586. 1825 = Exacum zeylanicum fide Grisebach (1845). LITERATURE CITED AvBLET, J. C. B. F. 1775. Histoire rs get de la Guiane francaise. London & Paris. 4 vols. [Lisyanthus, 1: 203. BENTHAM, G. 1876. Gentianeae Ey Gentine} In G. BentHAM & J. D. Hooker, Genera Plantarum 2: 820. Browne, P. 1756. The civil -. natura history of Jamaica in three parts. ndon. [Lisianthius, p. 157. Don, G. 1837. A general history eo hg PE plants. J. G. & F. Rivington. ondon. 4 vols [Gentianaceae, 4: 173-214. Ewan, J. 1948a. A revision of Macrocarpaea, a neotropical genus of shrubby stapepioe Contr. U.S. Natl. Herb. 29(5): 209-249. 48b. A review of Purdieanthus and Lehmaniella, two endemic Colom- vets anae of prerraiee and biographical notes on Purdie and Leh- mann. sea 5: A review a the neotropical lisianthoid genus Lagenanthus (Gen- aaa nce Mutisia 4: 1-5. Gite, E. 1895. Gentianaceae. In A. ENcLER & K. PrantL, Die Natiirlichen Pflanzenfamilien 4(2): 50-108. Gora, KrisHna, G., & V. Purt. 1962. Morphology of the flower of some Gentianaceae with special reference to placentation. Bot. Gaz. 124: 42-57. Grisepacu, A. H. R. 1839 [1838]. Genera et species Gentianearum. Stuttgart. 3 64 pp. 310 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 - 1845. Gentianaceae, Jn A. pE CANDOLLE, Prodromus systematis nat- uralis regni vegetabilis 9: 38-141. 18 Gentianaceae. Flora of the British West Indian Islands, 422-426. HemsLey, W. B. 1882. Gentianeae [Gentianaceae]|.. Biologia centrali-ameri- cana: Botany 2: 344-353. JoHANsEN, D. A. 1940. Plant microtechnique. McGraw-Hill, New York & London. Kuntu, C. S. 1818. Lisianthus. In F. H. A. von Humpsotpt, A. J. A. Bon- PLAND, & C. S. KuntH, Nova genera et species plantarum 3: 180-184 Kuntze, C. E. O. 1891. Revisio generum plantarum. Leipzig. 2 vols. [Lisian- thius, 2: 427, 428.] Lanjouw, J. & F. A. Startev. 1964. Index herbariorum. I. The herbaria of the world. Ed. 5. Regnum Vegetabile 31: 1-251. Linpsey, A. A. 1937. The floral anatomy and morphology of the Gentianaceae. Unpublished Ph.D. thesis, Cornell University Library, Ithaca, N.Y. - 1940. Floral anatomy in the Gentianaceae. Am. Jour. Bot, 27: 640-652. Linnaeus, C. 1767. Mantissa plantarum [prima]. Stockholm. [Lisianthus, 43 Linnaeus, C., FILIUS. 1781. Supplementum plantarum. Braunschweig. [Lisian- thus, p. 134 Love, D. 1953. Cytotaxonomical remarks on the Gentianaceae. Hereditas 39: 225-235. Maceripe, J. F. 1959. Gentianaceae. Flora of Peru. Fieldiana, Bot. 13(5): 27 NILsson, S. 1968. Pollen morphology in the genus Macrocarpaea (Gentiana- ceae) and its taxonomical significance. Sv. Bot. Tidskr. 62: 338-364. - 1970. Pollen morphological contributions to the taxonomy of Lisian- thus L. s. lat. (Gentianaceae). Ibid. 64: 1-43. PERKINS, J. 1902. Monographische Ubersicht der Arten der Gattung Lisianthus. Bot. Jahrb. 31: 489-494, Perrot, M. E. 1898. Anatomie comparée des Gentianaceés. Ann. Sci. Nat. Bot. Ser. 8. 7: 105-292, Perry, J. D. 1967. Biosystematic studies in the North American genus Sabatia (Gentianaceae). Unpublished Ph.D. Thesis, Duke University Library, urham, N.C. Port, V. 1952. Placentation in angiosperms. Bot. Rev. 18: 603-651. Rosinson, B. L. 1910. Spermatophytes, new or reclassified, chiefly Rubiaceae and Gentianaceae. Proc. Am. Acad. 45: 394-412. Ropyns, A., & S. Nizsson. 1970. Macrocarpaea browallicides (Ewan) A. Robyns & S. Nilsson comb. nov. (Gentianaceae). Bull. Jard. Bot. Nat. Belg. 40: 13-15. Rork, C. L. 1949. Cytological studies in the Gentianaceae. Am. Jour. Bot. 36: 687-701. SAVAGE, S. 1945. A catalogue of the Linnaean herbarium. London. : SPRENGEL, K. P. J. 1825. Caroli Linnaei... Systema vegetabilium. Editio ecima sexta. Gottingen. 5 vols. [Lisianthus, 1: 585-587. STEYERMARK, J. 1953. Gentianaceae. Contributions to the flora of Venezuela. Fieldiana Bot. 28(3): 496-499 ce de 1972] WEAVER, REVISION OF LISIANTHIUS 311 Swartz, O. 1788. Nova genera & oo plantarum seu prodromus. Stockholm, Uppsala, Abo. [Lisianthus, p. 40.] WEAVER, R. E., Jr. 1969. Cytotaxonomic bn on some neotropical Gentiana- ceae. Ann. Missouri Bot. Gard. 56: 4 1970. A new Jamaican species a aiabae (Gentianaceae). Brit- tonia PY tio 13. Witutams, L. O. 1968. Tropical American Plants, IX. Fieldiana Bot. 31(18): 401-425. [Lisianthus, pp. 406-411. . 1969. Gentianaceae. Jn P. C. Stanptey & L. O. Wixttiams, Flora of Guatemala. Fieldiana Bot. 24(3): 302-344. Woopson, R. E., JR. 1936. Observations on the floral fibers of certain Gen- tianaceae. Ann. Bot. 50: 759-766. ARNOLD ARBORETUM € MASSACHUSETTS 02130 312 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 SEEDS FROM SURFACE SOILS IN A TROPICAL REGION OF VERACRUZ, MEXICO * SERGIO GUEVARA S. AND ArTURO GOMEz-POMPA IN THE DEVELOPMENT of studies on the secondary vegetation of low- land tropical regions of Mexico (Gémez-Pompa, 1971; Gémez-Pompa et al., 1964; Sarukhan, 1964; Sousa, 1964) it has been very clear that addi- tional studies must be carried on if we want to learn more about the process of secondary succession. An important factor in the initiation of a secondary succession is the amount of seeds stored in the soil (“seed crop” or “floristic potential”). This area of ecological research has been developed to some extent in temperate regions ( Champness, 1949; Champ- ness & Morris, 1948; Milton, 1936, 1939, 1943, and 1948; Brenchley, 1918; Brenchley & Adam, 1915; Ovington, 1955; Olmsted & Curtis, 1946), but no records exist for tropical areas. 1965; Rico, 1972). Several hypotheses have been proposed to explain different aspects of succession but it is clear that what is needed is ex- perimental work to obtain data as a basis for some of these hypotheses (Gémez-Pompa, 1967). the problem of the trigger mechanisms in the germination of seeds of re- Fora or Veracruz. Contribution number 6. A joint project of the Instituto de Biologia of the National Universi i ity, to prepare an ecological floristic study of the state . Inst. Biol. Univ. Nal. Autén. México [Ser. Bot.] 1, 2. 1971.) This work was partially supported by a grant from the Fund for Over- i E). The authors are grateful to Dr. Bernice G. Schubert and Dr. Lorin I. Nevling, Jr., for reading the manuscript and making some helpful suggestions. 1972] GUEVARA & GOMEZ-POMPA, SEED GERMINATION 313 primary vegetation type (Oosting & Humphreys, 1940, 1947; Vazquez & Goémez-Pompa, 1971). MATERIALS AND METHODS Four areas were chosen for sampling purposes, two in secondary vege- tation and two in primary vegetation. In each area a quadrat of 100 m? was selected for floristic, environmental, and soil studies. In each of the quadrats sixteen samples of soil were taken at random for seed germination purposes. In order to understand the changes through time these were repeated eight times during the year of the study. So, a total of eight time periods were surveyed. In each quadrat a floristic inventory was compiled to obtain an idea of the species growing in that soil site. Air (75 cm.) and soil (2 cm. deep, FicurE 1) temperature and records of relative humidity 40- Exposed Soi seeeeee S0i1 With Vegetation 35- ott 8 Z 5 Ww ao = < 30- / ° S _ ° ~~ ei \ . 25- » ba _iaMtaisens “e., ° ng gee, "~e—e . —— ee oi $--.9—of te li eteneee oo ee 17 Hrs. 24 12 17 Hrs. HOURS Ficure 1. Soil temperature (24 hrs.). were taken through the year in all the sampling sites, and precipitation data were available from the meteorological station in the Biological Sta- tion located a few hundred meters from the farthest point of the sampling sites (FIcuRE 2). The soil samples were taken with a specially constructed cylindrical metal sampler 477 cm® in volume. Each soil sample was placed in a plastic bag and taken to a “germination house.” There each sample was put on 400 gms. of vermiculite and placed in plastic plates in the “house.” 314 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 LOCALIZATION OF THE SAMPLING SITES o m \ Cerro \ Ven Ay a El Vigia NN VA_ LP [J PRIMARY VEGETATION Ey] SECONDARY VEGETATION (] I-I SAMPLING SITES SC. 1:33 (C] MGERMINATION HOUSE Wy WZ STATION EQUIPMENT FicuRE 2. Localization of the sampling sites. 1972] GUEVARA & GOMEZ-POMPA, SEED GERMINATION 315 Germination was selected for evaluating the soil seed crop instead of a direct observation and count of the seeds in the soil (Barbour & Lange, 1967). The reasons for this are various but the most important one is the difficulty of identification of species by seeds (especially difficult for tropical species). The only objective of the “germination house” was to protect the soil samples from seed contamination. It had a ceiling of thin plastic and walls of mosquito netting to permit free circulation of air. Temperature and humidity data were taken in this house also. The plates were cir- culated during the six weeks they remained in the house in order to have homogeneous conditions. They were watered daily with equal amounts of water (3.5 liters). Four plates with only vermiculite were always in the house as controls for possible contamination. None were contaminated during the study. After six weeks the plates were analyzed for the germinated plants. A count was made of the number of different species and number of indi- viduals of each species. The species that we were not able to identify botanically were transplanted to a “nursery” for continued growth (age) and future identification. As the flora of the Station is known quite well (Gémez-Pompa & Nevling, 1971) the identification is not as unattainable a task as it may appear. Voucher specimens of the seedlings and of the “aged” plants were prepared and deposited in the N ational Herbarium of Mexico (mExu). We want to acknowledge the help of Dr. L. I. Nev- ling, Jr., with the identification of some of the difficult species and that of Dr. R. McVaugh who identified some of the Compositae. RESULTS Description of the region of Los Tuxtlas. The Tropical Biological Station is located in the Sierra de los Tuxtlas in the southern part of the state of Veracruz in the parallel 18° 30’ N and in the meridian 95 Ww. Because of its orography the region is one of the most humid areas in FicuRE 3. Meteorological records from Coyame Station and the Estacion de Biologia Tropical Los Tux’ METEORO- MEAN EXTREME EXTREME RELATIVE LOGICAL TEMP. MAX. MIN. Humipity RAINFALL STATION (°C) Temp.(°C) Temp.(°C) % (mm.) CoyAME * 23.6 38.5 9.5 —_ 4419.8 (May) (January) ESTACION DE BIOL. 25.0 43.3 12.2 81.9 4012.0 Trop, Los TuxTLAs * (June) (February ) 1 The record covers the period from 1921 to 1970. * The record covers one year (1970). * Taken from Soto (1972 Ficure 4. Soil properties * Nitric + HCs- Poras- DEPTH NITROGEN CALCIUM — PHORUS SIUM (cm.) COLOR TEXTURE PH GANIC MATTER Ko./Ha Ke./Ha Ke./Ha_ Ke./HA 0-30 reddish brown clay loam 6.25 4.45 27 3129 14.6 413 30-60 reddish brown clay loam 6.20 1.37 14 1498 e: 90 60-100 reddish brown clay loam 5.80 1.71 14 1735 7.9 124 * Flores, 1971. O1E WOLAYOIAVY GIONUV AHL AO TYVNUNOL €$ *I0A] 1972 | GUEVARA & GOMEZ-POMPA, SEED GERMINATION 317 the Gulf Coast of Mexico and precipitation of more than 4 meters per year has been recorded (Garcia, 1971) from nearby stations. Data avail- able from the meteorological records of the station during the year of the study are recorded in Ficure 3, but cannot be used for evaluation of climate because of lack of comparative data from preceding years. The closest station which can give a better idea of the climate is Coyame (Fic- URE 3). The vegetation of the region is composed of many types produced by variations in temperature, precipitation, and soil properties (G6mez- Pompa, 1972; Sousa, 1968), with a very rich floristic composition. Floristic differences may be seen within short distances. In adjacent areas floristic variations arise chiefly from secondary successional stages (Gé- mez-Pompa, ‘ Description of the Station. The station is located on the NE slope of the Sierra of San Martin, very close to the coast, and is covered mainly by a High Evergreen Selva of an average of 30 meters in height, and a few secondary vegetation sites along the main road crossing the station property (F1icurE 2) on one side. The sampling was done on the lower slopes of a small hill (Cerro El Vigia) belonging to the station (as can be seen in FIGuRE 2). The soil in the station is a brown and yellow acid soil derived from volcanic ash. Some of its characteristics are indicated in FicureE 4. Description of the sampled sites. Site 1. This area is covered with secondary vegetation about five years old. Before that time it was cul- tivated with corn for about one or two years, and before the corn, it was a primary High Evergreen Selva. This quadrat is located on a slope of about 15 to 30° inclination. There is a considerable quantity of volcanic alluvial rock over the surface (40%). The average height of the upper stratum is about 7 to 8 meters. The altitude is 108 meters above sea level. Some records of soil and air temperature during the study are shown in Ficure 5, those of the humidity and dew point in Ficure 6. The species collected are listed in FIGURE 7. SirE 2. This area is covered with a High Evergreen Selva about 30 meters tall. The quadrat is located on a slope of about 15 to 30° inclina- tion. It is also covered with a considerable proportion of volcanic rocks (30%). The altitude is 120 meters above sea level. Soil and air tempera- ture are shown in Ficure 5, records of humidity in Ficure 6. The species collected are listed in FIGURE 7. Site 3, When the study started, the area was covered by secondary vegetation of two months’ duration. The area was covered before that by secondary vegetation for a few months and earlier it was a Tall Ever- green Selva. The quadrat is located on a slope of about 20 to 40° incli- nation. Free of rocks. Observations of soil and air temperature and hu- midity are shown in Ficures 5 and 6. The species collected are listed in FIcureE 7. 318 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 TE W 2 _SITE! Pes _SI El ae L 4 Fotis | Jf ‘i | “ TEMPERATURE °C = ¢ A sista F = OF nie os ae, 2 Fo MN Y = a 3 = DP. a . z DP. e w 504 / -e" esta ee -20 0 a 504 am #9 +20 0 a 7 Pd g 2 ii Z & 4 9 = 2 od Fa 4 SS vSISS SEED SSIES CEES ANUARY MON JANUARY ANUARY JANUARY woo SITE MM ae oom SITE IV eer axes be A pat oe iia i Re \/f 5 eo of ff RH. \* RH é > s E z E s seo Ey g 3 ioontng onee®, 4 5 Pa. ‘ s 5 ey = = DP. ye 2 DP. ue Qe 50 i a 4 +20 0 w fe 209 w 50 A ae rs = F z= 2 & : a& "4 — ad » a = C3 SUNETTSIs TICs Lt JANUAR MO JANUARY ANUARY JANUARY GERMINATION coal 100% ae rl) + Ss = 2 o = “ m 2 vat p 2 mr er, +20 0 2 — Y = & 5 — a c oe Gag Oe a ee JANUARY MONTHS JANUARY Ficure 6. Air humidity: ———— 7:30 A.M.;..... 4:30 P.M. 320 JOURNAL OF THE ARNOLD ARBORETUM FicurRE 7. Flora of the sampling sites [voL. 53 SPECIES wr A ta! SITE SITE Ill SITE IV Aneilema geniculata Anthurium myosuroides Astrocaryum mexicanum Cardiospermum halicacabum Casearia arguta Cecropia obtusa Celtis igua Citkareesilens ‘et Cupania macrophylla Cymbopetalum baillonit ma Robinsoactlo mirandae Solanum sp. Syngonium podophyllum Aegiphila costaricensis Ilex condensata Justicia sp. Lomariopsis sp. Monstera tuberculata Nectandra salicifolia Philodendron sanguineum a mexicana Polypodiu Prondelasades oxyphyllaria Reinhardtia gracilis X FH+tt ttt ttt++++tttt+++eets+si ++ ++ +HP++++t+ttttt+ttte++sest ++ + + ++ + + ++ + + ofo ++ ++ +4 1972] GUEVARA & GOMEZ-POMPA, SEED GERMINATION 321 FIGuRE 7 Continued - SITE SITE SITE SITE SPECIES I II III IV Serjania sp. a aes ecg donnell-smithii +~ Tradescantia sp. + Trichilia et + Acalypha deppeana Acalypha schiedeana Ageratum conyzoides Aspidosperma oe” Brosimum alicastrum Calathea microcepha resi Zan Cestrum Cirsium oe Clematis dioica Clibadium arboreum Elvira biflora Buderblobien cyclocarpum Eupatorium sp. Eupatorium macrophyllum ICUS S Heliocarpus appendiculatus Inga spuria Tresine celosia Jacobinia sp. Lygodium sp. Ocotea dendrodaphne Ompbhalea cardiophylla Paullinia pinnata Piper sp Pothomorphe umbellata xX i a Spigelia anthelmia Spondias mombin Taraxacum o ficinale Verbesina greenmani FHEEEEHEEE HEHEHE EEE HEF HE EE HHH 4444444 S i) Ay | S 3 8 38 oO Ns i — oerstedii errene te Oreopanax ‘diseide Paullinia costaricensis +++++4++ 322 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FIGURE 7 Continued SITE SITE SITE SITE SPECIES I il iil IV Poulsenia armata + Salacia megistophylla Thelypteris ghiesbreghtii + rophis mexicana + Trophis racemosa - SyMBOLS: X Species of which seeds were found in the soil samples. pecimens of each species were deposited at the Herbario Nacional, UNAM (mexv). Site 4. This area is located in a High Evergreen Selva similar to SITE 2, but with a lower inclination of only 8 to 10°; it is almost free of rocks at an altitude of 131 meters above sea level. Records of soil and air temperature and humidity are tabulated in Ficurrs 5 and 6. The species collected are listed in FicurE 7. Site of the germination house. The experiments were carried on in an area that simulated an open soil free of plants and not covered by shade from any plant, The area was near the laboratories of the station in order to be maintained and looked after properly. A few soil samples were brought to Mexico City to be germinated in the greenhouses of the Bo- tanic Garden for comparison and indication of possible differences, but since in all cases there were none, we proceeded with only the analysis of data from our local germination house. Readings of air and soil tempera- ture and humidity in the house are recorded in Ficures 5 and 6. Germination of the seeds. Germination was abundant and started al- most immediately (after a few days). There were a few seedlings ob- served that did not continue normal growth; in some cases these were transplanted and care was taken to look for microenvironments, In some cases we were fortunate that these plants continued to grow, in other cases, however, they died. This problem will be discussed later as a possibility for future studies. More than 80 percent of the species are now identified, (FicuRE 8), leaving only a few still to be determined. The following species were identified in living condition in the soils of the sampling sites: QUADRAT 1 Phytolacca decandra L. Bidens pilosa L. Tresine celosia L. Euphorbia heterophylla L. Neurolaena lobata (L.) R. Br. Eupatorium sp. 1972] Ageratum conyzoides L. Clibadium arboreum Donn. Sm. Clibadium grandifolium Blake Paspalum sp. Dioscorea s Solanum torvum Sw. Verbesina greenmani Urb. GUEVARA & GOMEZ-POMPA, SEED GERMINATION 323 Urera caracasana (Jacq.) Griseb. Amaranthus hybridus Spigelia palmeri Rose Heliocarpus aff. donnell- we Rose Emilia sonchifolia (L. pao alae (Kuntze) Ryd QUADRAT 2 Solanum cervantesii Lag. Lasiacis papillosa Swallen Phytolacca decandra L. Cecropia obtusifolia Bert. Heliocar pus aff. donnell-smithii Rose Bignoniaceae (unidentified) Tur pinia occidentalis (Sw. G. Don Sapium lateriflorum Hems Tresine celosia L. Belotia campbellii Sprague Erechtites hieracifolia (L.) Raf. Pothomorphe umbellata (L.) Miq. Eupatorium macrophyllum L. Neurolaena lobata (L.) R. Br. Ageratum conyzoides L. Clibadium arboreum Donn, Sm. Paspalum sp. Sida acuta Burm Trema micrantha (L.) Blume Robinsonella mirandae G6mez-Pompa Panicum trichoides Sw. Bidens pilosa L. Enterolobium cyclocarpum (Jacq.) Griseb. se Axono pus compressus ne ) Beauv. Amaranthus hybridus L Pleuranthodendron mexicana (Gray) L. Wms. QUADRAT 3 Solanum nigrum Phytolacca dicondra Ls. Neurolaena lobata (L.) R. Br. Heliocarpus ee Turcz. Bidens pilos saL Tresine celosia L. Potlomiorbhe’ umbellata (L.) Miq. Vernonia aff. deppeana Less. Solanum cervantesit Lag. Robinsonella mirandae Gomez-Pompa Verbesina greenmani Urb. Eupatorium sp Ageratum conyzoides L. Physalis pubescens L. Paspalum sp. Mikania micrantha rg Brosimum alicastru Heliocarpus domaell siti Rose Lauraceae (unidentified) QUADRAT 4 Phytolacca decandra L. Solanum cervantesii La Robinsonella mirandae Gémez-Pompa Trema micrantha (L.) Blume Cecropia obtusifolia Bert. Heliocarpus oon am Rose Eupatorium Costus ace (Jacq.) Sw. Ageratum conyzoides Desmodium adscendens (Sw.) DC. FicuRE 8, Species found in the soil samples ze Srre I Srre II see2eans e285 3795 ZAmnInA Mar 19 May 3 Jun 28 Aug 10 Sep 27 Nov 15 Jan 4 Feb 18 SpEcteEs * Acalypha schiedeana Ageratum conyzoides * ee 7 e . Amaranthus hybridus e . ° Axonopus compressus ° Belotia campbellii ee Bidens pilosa ee . ° Brosimum alicastrum Cecropia obtusifolia ee Cirsium mexicanum ° Clibadium arboreum eee * * ° Clibadium grandifolium . Costus spicatus Croton draco Desmodium adscendens Dioscorea sp. e Emilia sonchifolia e Enterolobium cyclocarpum e Erechtites hieracifolia ° Eupatorium ee ee w ee Eupatorium macro phyllum ° Heliocar pus appendiculatus ° WOLAXOdUV GIONUV AHL AO TVNUNOL €¢ “10A] Heliocarpus aff. donnell-smithii Tresine celosia sear Paspalum Deiiranthodendron mexicana Pothomor phe umbellata Physalis pubescens Phytolacca decandra Robinsonella — Sapium lateri trgaa spaotitétolia ppeana Bignonioceae (unidentified) Lauraceae (unidentified) *Specimens of each species were deposited at the Herbario Nacional, UNAM (mexv). NOLLVNINUYAD AAAS ‘VdNOd-ZaAWOD ® VUVAAND [Z7Z61 sz¢ 326 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 DISCUSSION From the results we have up to now, we can distinguish several groups of species in relation to their presence in the soils sampled: Species present in the soil all year in most samples: Ae eos ee b ge - L. Eupator: Tresine c N. intned iia (L.) R. Br. Rbbiuonella mirandae Gémez-Pompa Species present for only a short period of time: Brosimum alicastrum Sw. Mikania micrantha HBK. Cecropia obtusifolia Bert. Belotia campbellii Sprague Vernonia aff. deppeana Less. Species present all year but in few samples: Clibadium arboreum Donn. Sm. Heliocarpus donnell-smithii Rose Bidens pilosa L. Paspalum sp. Amaranthus hybridus L. Solanum torvum Sw. Secondary species present in soil of the primary rain forest: Erechtites hieracifolia (L.) Raf. Trema micrantha (L.) Blum Cecropia obtusifolia Bert. Pleuranthodendron mexicana ’ (Gray) Croton draco Schlecht. L. Wms. Desmodium adscendens alle a: Axonopus compressus (Sw.) Beauv. Costus — (Jacq.) Sw Mirabilis jalapa L Sida ac urm. Eupatorium pycnocephalum Less. Belotia ¢ came Sprague Primary species present in soil of the secondary vegetation: Brosimum alicastrum Sw. Lauraceae (unidentified) Secondary species found exclusively in the secondary vegetation: Lasiacis papillosa Swallen “ah seg grandifolium Blake Solanum nigrum L. Diosc Euphorbia heterophylla L. So boca | torvum Sw Eupatorium aff. pansamalense Rob. Urera caracasana (Téca. ) Griseb. Physalis pubescens L. Verbesina greenmani Urb. Emilia oe (L.) DC. Spigelia palmeri Rose Paspalum Cirsium mexicanum DC. S cistocorha oppositifolia (Kuntze) Acalypha schiedeana Schlecht. Rydb acobinia s A eliocar pus appendiculatus Turcz. Mikania micrantha HBK. Primary species found exclusively in the primary vegetation: Tur pinia occidentalis (Sw.) G. Don Sapium lateriflorum Hemsl. Enterolobium cyclocarpum (Jacq.) Griseb. It is obvious from all the preceding lists that by far the most important floristic element in the soil is composed of secondary species, which are 1972] GUEVARA & GOMEZ-POMPA, SEED GERMINATION 327 also the species having longer periods of dormancy (Pijl, 1969). This means that a primary forest can not be restored if all the trees from adja- cent areas are destroyed. This is another extremely important argument for conserving large and numerous pieces of tropical selvas. It is also evident that floristic potential in the soil is a very important element in the direction of the succession. The time of year is also im- portant from this point of view because different species are available in the soil at different seasons. 8- NUMBER OF SPECIES b | | j ace ee ] ce ram j j i UJ i 12 ab | j pee J j j Wr j 4 j “a. = j IImy ILlmw IrTMW IImW ImImwW ITmy IrImy ] I MARCH 19 MAY 3 JUNE 28 AUGUST 10 SEPTEMBER 27 NOVEMBER 15 JANUARY 4 FEBRUARY 18 DATES AND SITE NUMBERS Ficure 9. Mean of the number of species. It is a striking fact that in the soils of the primary selvas there is a floristic potential of secondary species ready to function after clearance of the forest. The factor keeping the secondary seeds dormant is un- known and worth investigating, because the species represented are not found in the primary forest. Also the trigger mechanisms should be studied in order to clarify the factors involved in the succession. We are well aware that in this work we have only scratched the surface of the problem and that there are many other ways to trigger the ger- mination of soil seeds which under the method used (temperature variation, humidity, light, chemicals, etc.) have not germinated or germinated but did not continue growing, probably because the right combination of con- ditions was not attained. We hope in the future to obtain new data to increase our understanding of this problem. One of the most interesting fields yet to be explored is that of seed and fruit dispersal (Smythe, 1970) because it is perfectly clear that the 328 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 location It is also important to add that there is a great need for floristic studies with a solid ecological approach that can be used to interpret the behavior of different species under different ecological conditions and also to study their evolution in connection with secondary environments (Gomez-Pompa, 1971). floristic potential can not be built if the seed has not arrived at a given 80 ” al < = a 2 a = P : = u 40- Pa ~ - - 5 = a | z +" Ps Mi q i : 4 4 = i = if 4 i IrImw [Lmw ee iguigug TimW Imy [TimW [LmwW L vg MARCH 19 MAY3 SEPTEMBER 27 NOVEMBER 15 JANUARY 4 FEBRUARY 18 UNE 28 AUGUST 10 DATES AND SAMPLING SITES Ficure 10. Mean of the number of individuals. The data on the number of species and of individuals per sampling are difficult to evaluate, because there is much variation through time (FIGURES 9, 10, 11, 12). The only interesting observation we can make is that the greatest number of species in our samples occurs in Sire 1; this may be because it is the oldest secondary site, so, time [is] might be responsible for the greater accumulation of secondary species as it might also be for the increase of the number of species in Strz 3. But this interpretation should be confirmed because there is much variation and the number of samples made by us is not enough to give a completely satisfactory in- terpretation. The temperature data of the four sites and of the germination house showed some differences (Ficure 6) between the secondary and the pri- mary sites, being more variable in the secondary locations. It is of inter- est that there is very little variation in soil temperature throughout the year. Nevertheless the temperature of a naked soil is very different from that of a vegetation-covered one (FIGURE 1), especially in the middle of 1972] GUEVARA & GOMEZ-POMPA, SEED GERMINATION 329 the day. We believe that this difference may explain the response of certain dormant seeds to the clearing of the forest. The temperatures in our germinating house were closer to those of a bare soil and the re- sponse we obtained from the plants was similar to that of an early colonized open soil CONCLUSIONS The germination method is an effective one for evaluating the floristic potential of soils in tropical areas. In order to use it advantageously the composition of the flora and the ecology of the species of the region must be well known. The greatest proportion of the seeds found in the soil of the four sites was comprised of seeds of secondary species, several of them being pres- ent through the year of the study, indicating that dormancy seems to be an important mechanism in the succession. No primary species showed that characteristic. The initial stages of the succession are already imprinted in the soil of the primary selvas and its characteristics may be derived from other factors such as time of the initiation during the year and may also include effects of some chemicals, of temperature, fire, etc There are significant differences in the temperatures obtained in the primary sites compared with those of the secondary ones. But these data are difficult to evaluate in connection with the results obtained in this study. The thermoperiod may play an important role in the initial stages of the succession but this must still be proven. More research should be done on comparable problems to reinforce some of the data obtained by us. Ficure 11. Statistical data on the species found in the soil samples = a SITE I SITET SITE I SITEIV : 2 © 2§ 2}o[2]21s] 2] -[ekse4211elelale]-|2b se 21 eels |el-|2kesed2|ols/2|sle|+|2 sea = $123 Ly a>pc si 5 feos) &) alc] >t al > ‘Oleoles} ©] >} c] o] al > olesles} s| >] c] o] a] =| cl] ofeeea] 8 o|/ SE SPECIES $|3|3|2]3|3|5|8Ba29| S|5|5| <|3| 215/249] 5|5|5| 213|2|8|2 spas |2|3| 21313 |S|@bazae 3 | 2% Yee & i{ {1 2 2 ACALYPHA SCHIEDEANA NUMBER OF INDIVIDUALS St 12 Ss 31 [7s] [4 ly 3 1[ [8 62 8 i 2 42 AGERATU M CONYZOIDES 251 I72l70 ag lst It ) 14 26l 3 29 1 3 ¥ 330 AMARANTHUS HYBRIDUS i 4} 12 ar ‘ i: : 4 13 4 sé 3 7 2\1 3 3 3 ae a { n 4 ~— 4 i 4|4 g 2 BELOTIA CAMPBELLIL 214 3 | B 1|3 4 4 ee 4 2 7 IDENS PILOSA 7A] 1¢ 2 als 3 eS 24 Yji2\4147 § BROSIMUM ALICASTRUM Sl2i4| |& 8 4 {4 2 | | 3 CECROPIA OBTUSIFOLIA 114 2 4 4 3 4 4 4 4 a CirstUM MEXICANUM 2 2 1 1 3 9 CLIBAPIUM ARBOREUM : rc : = i 8 ‘ 29 CLIBADIUM GRANDITOLIUM ' | : 4 | | | 4 1 i Costus SPICATUS Be: 1 4 j a a 4 1 i CROTON DRACO r rt | | y 14 4 1 DESMODIUM ADSCENDENS 2 Droscorea sp. ; { 2 EMILIA SONCHIFOLIA 4 4 1 4 ENTEROLOBIUM CYCLOCARPUM t 4 { ERECHTITES HIERACIFOLIA : 4 4 Evupatoriu : {|10| |éls| [830 413 7 S[2Yi 4] [Zhe 2 UPATORIUM SP rif cal cd et Sle W aatslelal Isol les EUPATORIUM MACROPHYLLUM 4 4 4 4 2 i 4] |5 { 6 : 1 4 EUPATORIUM PANSAMALENSE 5 a ame se a 4 1 I.UPATORIUM P NOCEPHALUM. 2 vy] eh 4 4 EUPHORBIA HETEROPHYLLA 1 4 4 4 HELIOCARPUS APPENDICULATUS 1 4 HELIOCARPUS AFP. 4} {4 2) 1dji 3\9 4}4 2 DONNELL-SMITHII 4 4 Aljtold & ) 20 4)4 2 ore 5112}/019 1413 | 2)6 |e} [6] i113|7|2| 314/339] |9 I7z| v173/4]6 [5 lvol70 3 4 ee 4/\4a\ 4 |33\20|62| |e 7 [24/4/84] 5] Sli] |ttyolsslveilmlavlea|zeliog yas 3 7 1 1 JACOBINIA SP, BHRB Rae a 2 2 FIGURE 11. Statistical data on the species found in the soil samples i . SITET SITE SITE II 22/25 e}ofa]e|S]2]<|2 bse] -[2lols|e| lok steae|-[e]2|s]2|+|2bsea2|o ot FS: the oD > ca, ao > ray ico o> a al> > i= S = SPECIES $13|3]2]3|8| 826359 5] $| 5] 2] 3] 3] §|86369] 5/215 | 213] 2| 5 |e bse3s |S 25 | es LASIACIS PAPILLOSA THE SPECIES WAS FOUND iji 2 NUMBER OF INDIVIDUALS 4 4 2 a MIKANIA MICRANTHA . A} * | a] |4 ala] [az] [29 MirABILis JALAPA t— 4 4 4 Q 2 NEUROLAENA LOBATA EE | 4) jata la 4 | 4 41 d 4 3 } | Brenrr 4} {4 1/4 4] {4 3 PANICUM TRICHOIDES pcs 4 4 2 4 1 5 | al fi af fi 4 PAsPALUM sp. | i411 {ji} i}d 2 7 3 3i1 4 1|2 s 4} 2 3 PLEURANTHODENDRON 4 1 1 MEXICANA 3 S PoOTHOMORPHE UMBELLATA J 3 4 4 Ss 6 6 d i PHYSALIS PUBESCENS d 4 1 i { PHYTOLACCA DECANDRA 6 212171 29 6)3{11)3)4 8 | 528 joj i 1) 4] 1] 9} 2/28 6)/41)4/41/2)4) 3) 3/48 95 8 3/S|aelt6] [68l/v] 51414 2gii¢| |74i}7o] 4} [4 (26) Liuaiyo lass ddl 2i4tilzii{ rie} |3t ROBINSONELLA MIRANDAE | 4\i s S| 12 re 1] 717191) 2) 7/8] 9}50 64 @ Li lile lis|_ lal [sx] [oe tol a lirlarlial 2 142\t51251 _ {0s SAPIUM LATERIFLORUM |_| {4 | Lj4 cA 12 | Lt SCHISTOCARPHA OPPOSITIFOLIA ; : a SIDA ACUTA ; : 2 SOLANUM CERVANTESII ; { : : : 6 SOLANUM NIGRUM: 3 : 5 7 SOLANUM TORVUM ; ; t © SPIGELIA PALMERI t ¥ TREMA MICRANTHA 1 : 2 2 TURPINIA OCCIDENTALIS ; . : URERA CARACASANA ji : 4 VERBESINA GREENMANI ; i { 2 : ‘VERNONIA AFF, DEPPEANA 1 t i i 3 : DBIGNONIACEAE ‘ ; ‘ 2 f HAURACEAE : t 4 : 334 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FIGURE 12. Seed content per m? SEEDS PER & “ % S ~ » a cI 2 > o > e M* 12 cM. % = z S ey z E; DEPTH S = a < n Zz = SITE I = 1982 3879 2937 3275 1982 4051 3189 3706 SITE II = 517 862 344 603 431 431 862 862 Sire III 1120 1293 1439 1637 948 862 1560 2672 SITE IV 603 689 431 258 175 344 431 517 REFERENCES Barsovr, M. G., & R. T. LANcE. oT population in some natural Austra- lian topsoils. Ecology 48: 153- BRENCHLEY, W. E. 1918. Buried Bh seeds. Jour. Agric. Sci. 9: & H. Apam. 1915. ng ae a 5 aaa land ee ath te revert to natural conditions. Jour. Ecol. 3: BupowskI, G. 1965. Distribution of tie ae. rain forest species in the light of nara processes. Turrialba 15(1): 40-43. CHAmMpPNEss, §S. S. . Note on the technique of sampling soil to oe the content of vers viable seeds. Jour. Brit. Grassl. Soc. 4: 115-1 & K. Morris. 1948. The population of buried viable seeds in ae to contrasting pasture and soil types. Jour. Ecol. 36: 149-173. Fiores, J. S. 1971. Estudio de la vegetacién del Cerro del Vigia de la Estacion de Biologia Tropical Los Tuxtlas, Veracruz. Tesis. Fac. de Ciencias U.N.A.M., México. 95 pp Garcia, E. 1971. Los Climas del Estado “2 ee Anal. Inst. Biol. Univ. Nal. Auton. México [Ser. Bot.] 41: G6mez-Pompa, A. be Some problems e Fal plant ecology. Jour. Arnold Arb. 48: 105-1 1971. aL bod de la vegetacién secundaria en la evolucién de la flora tropical. Biotropica 3: 125-135. . 1972. Ecology of the vegetation of Veracruz. Elseviere Publications _ > press I. NEvLINe, JR. 1971. Lista de especies de Veracruz. Listado pe- adie por computadora. De circulacién restringida. Instituto de Biologia. U.N.A.M. ,» J. Vazquez Soro, & J. SaRUKHAN. 1964. Estudios ecolégicos en las zonas célido-himedas de México. Publ. Esp. Inst. Nal. Invest. Forest. México 3: 1-3 , W. ou 936. Buried viable seeds of enclosed and unenclosed hill land. Bull. Welsh. P. Breed. Sta. Ser. 4. No. 14: 58-72 - 1939. The occurrence of buried viable seeds in soils at different eleva- tions and on a salt marsh. Jour. Ecol. 27: 149-159, - 1943. The buried viable seed content of a midland calcareous clay soil. Empire Jour. Exp. Agric. 20: 155-167. 1948. Buried viable seed content of upland soils in Montgomeryshire. Empire Jour. Exp. Agric. 16: 163-177. 1972] GUEVARA & GOMEZ-POMPA, SEED GERMINATION 335 Oxmstep, N. W., & S. D. Curtis. 1946. Seeds of the forest floor. Ecology 28: 9-52 Oostine, H. S., & M. E. Humpureys. 1940. Buried viable seeds in a ireerweg series of old field and forest soils. Bull. Torrey Bot. Club 67: 253-273. & . 1947, Buried viable seeds in the forest floor a a series of stands representing stages of succession from old field climax. Bull. Ecol. Soc. Am. 20: 32. OvincToN, J. D. 1955, Studies on the ht jee of woodland conditions under different trees. III. The ground flora. Jour. Ecol. 43: 1-21 Pry, L. VAN DER. 1969. Principles of dasa? in n higher plants, Springer-Verlag, Berlin. 153 pp. Rico, B., M. 1972. Estudio de la sucesién secundaria en la Estacion de Biologia Tropical Los Tuxtlas, Veracruz. Tesis. Fac. de Ciencias. U.N.A.M. México. ss. SARUKHAN, K. S, 1964. Estudio sucesional de = — talada en Tuxtepec, Oax. Publ. Esp. Inst. Nal. Invest. Forest. 3: 2, SMYTHE, N. 1970. Relationship between rralting seasons or seed dispersal methods in a neotropical forest. Am. Naturalist 104: 23- Soto, M. 1972. Estudio estadistico de algunos fenémenos as no empleados en la clasificacion de Garcia (1971) y su repercusién en las condiciones eco- climaticas en el Estado de Veracruz. Mss. Sousa, M. 1964. Estudio de la vegetacién secundaria “ee la region de Tuxtepec, ax. Publ. Esp. Inst. Nal. Invest. Forest. 3: 91- ee Ecologia de las leguminosas de Los Osi uaa An. Inst. VAZQUEZ Y., C., & A. GomeEz-Pompa. 1971. Estudios wae de la ger- minacion de semillas de algunas especies de la vegetacién secundaria en una zona calido himeda de México. IV Simposio Latinoamericano de Fisiologia Vegetal. Septiembre 20-26. Univ. Nac. Agraria La Molina, Lima, Pert. pp. 79, 80. DEPARTAMENTO DE BOTANICA INSTITUTO DE BIOLOGIA UNIVERSIDAD NACIONAL AUTONOMA DE MEXICO 336 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FLORAL ANATOMY OF MYRTACEAE, II. EUGENIA RUDOLF SCHMID UNTIL FAIRLY RECENTLY, the floral anatomy of the Myrtaceae had been a rather neglected area of botanical research (see Schmid, 1972c, for references to earlier work, especially those cited for Carr & Carr). Recent work on the Eugenia-Syzygium controversy (Schmid, 1971, 1972a-c) based the recognition of both taxa on new floral anatomical as well as on hither- to unemphasized organographic evidence. The bearing of floral anatomy on the taxonomy of the strictly Old World genus Syzygium sensu lato was the concern of a previous report (Schmid, 1972c), whereas the compara- tive floral anatomy of the mainly New World Eugenia sensu stricto (but including, for convenience, the Old World segregate Jossinia) is dealt with here. MATERIALS AND METHODS Reproductive material of three extra-American and 20 American spe- cies (including two varieties) of Eugenia s. s. was examined. TABLE I lists voucher specimens and other pertinent information for the species analyzed. Detailed methodology as well as reviews of the taxonomic and anatomical literature were given in earlier papers (Schmid, 1972b, c). TABLE I: Collection and preparation data for species studied + Sect.3 CLEARED 3 SPECIEs, Locatiry, CoLLECTOR 2 Mrp. UNMTD. Eugenia acapulcensis Steud. Mexico: Chiapas. a 0 0 Breedlove 14490, — —. Mexico: Michoacan. McVaugh 22554. 4 5 0 — —. Nicaracua: Zelaya. Standley 20016. 5 0 0 Eugenia aeruginea DC. BritisH Honpuras. Gentle 8073; 4 0 0 — —. GUATEMALA: Petén. Contreras 1403. 0 3 6 —. GUATEMALA: Petén. Contreras 114. 0 4 0 Eugenia biflora (L.) DC. Brazit: Para. Black 52-15623. 2 4 0 — —. CoLomsia: Vaupés. Schultes & Cabrera 14236. 1 6 0 — —. VENEZUELA: Bolivar. Steyermark & Wurdack 3 4 0 oi, — —. Brazit: Amapé. Pires & Westra 48813. 3 5 0 — —, var. wallenit (Macf.) Krug & Urban. JAMAICA. Hart 1048. 4 3 0 Eugenia capuli (Schlecht. & Cham.) Berg. Mexico: 5 12 4 Chiapas. Breedlove 14407. — —. British Honpuras. Gentle 5044. 6 2 6 Eugenia cartagensis Berg. Costa Rica. Austin Smith 77. 6 5 0 Eugenia cofeifolia DC. Lesser ANTILLEs: Guadeloupe. 9 0 0 L’Herminier s.n. 1972] SCHMID, MYRTACEAE, II. EUGENIA 557 — —. LEssER ANTILLES: St. Lucia. Proctor 17978. 2 7 0 Eugenia confusa DC. Lesser ANTILLEs: St. Lucia. 3 0 0 Proctor 18135. — —. JAMAIcA. Proctor 23563. 0 is 4 Eugenia duchassaingiana Berg. LESSER ANTILLES: Guade- 5 1 2 loupe. L’Herminier s.n. Eugenia flavescens DC. Surinam. Florschiitz & Maas 5 a 0 2653. Eugenia agen, DC. Cotomsra: Amazonas. Schultes & 6 0 0 Cabre — —. veiatere Bolivar. Steyermark & Gibson 0 4 0 95781. Eugenia gregii fag Poir. Lesser ANTILLES: St. Lucia. 5 0 0 Proctor ca — —. LEss e Aneace Ss: Martinique. Hahn 641. 0 4 2 Eugenia sibs otaenis Urban. Jamaica, Proctor 26872. 2 0 1 — —. Jamaica. Proctor 19686. 3 1 2 Eugenia muricata DC. Brazit: Amapa. Pires et al. 7 2 2 50887. Eugenia oerstediana Berg. Mexico: San Luis Potosi. 8 0 King 4405. —_——. ia Honpouras. Gentle 2533. 6 ) 0 ‘ co: Jalisco. McVaugh 23418 = 5 0 Bole pina ir Standl. Mexico: Nayarit. McVaugh 7 7 1 15322; Eugenia salamensis Donn. Sm. var. hiraeifolia (Standl.) 5 0 0 McV. Costa Rica. Pittier 13952. — —, var. salamensis. Mexico: Jalisco. Carter & 3 nt 0 Chisaki 1209. Eugenia tikalana Lundell. GUATEMALA: Petén. Contre- 7 7 0 ras 1116, oe uniflora L. BERMUDA. Taylor 49-1048. 4 5 0 —. Lesser ANTILLES: Nevis. Proctor 19469. 0 3 0 ugenia venezuelensis Berg. GUATEMALA: Petén. Con- 4 5 0 treras 2971. ieee whytei Sprague in Stapf. Liperta. Cooper 452. 1 0 0 Eugenia winzerlingii Standl. GUATEMALA: Petén. Con- ) 0 treras 1619. — —. Mexico: Yucatan. Lundell 851. 0 8 0 — —. MEXICco: ne ee Enriquez 815. 0 2 0 Jossinia aherniana (C. B. Rob.) Merr. PHILIPPINES. 4 1 0 Velasco s.n. (Forestry Bureau 21779). Jossinia palumbis (Merr.) Diels. MartaNa IsLanps. 2 0 0 Hosokawa 7570 * Voucher nia - Jossinia aherniana at us (specimen courtesy Dr. R. H. mes gis others at MI ment Peni paneer (e.g., state, province, department, terri- tory vetted given for larger countries o * Number of buds, flowers, and/or fruits sectioned Foon ened cleared. Col- umns under “Cleared” refer to number of items mounted (Mtd.) on permanent slides and unmounted (Unmtd.). All preparations from herbarium material. 338 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 OBSERVATIONS — GENERAL Descriptions of the general organography, histology, and vasculature of Eugenia s. 1. were presented in Schmid (1972b). Only the distinctive transeptal ovular supply of Eugenia s. s. is detailed below. A variable number (up to about 11 per flower in my material) of bundles (placental strands) originate in several ways and constitute the transeptal ovular supply: (1) exclusively from the vascular cylinder of the floral tube (e.g., Eugenia capuli, Ficure 13; E. confusa, Ficure 1), (2) exclusively from a lateral carpellary system of various types (e.g., E. flavescens, FicureE 18), (3) exclusively from the dorsal carpellary bundles (e.g., sometimes in E. biflora), or, usually (4) from any combi- nation of the above. The mode and level of origin of the placental strands is by no means constant for a species or even an individual. The placental strands enter the septum at different levels in the locular region and fuse variously (cf. Ficures 3, 18, 19), eventually terminating at one or two centrally located placental plexuses. If a single placental plexus occurs, it may be a solid mass of vascular tissue (e.g., Eugenia coffeifolia, E. tikalana, E. uniflora, Ficure 19) or it may be perforated by varying amounts of interfascicular parenchyma (e.g., E. biflora). As might be expected, these features are very variable within a species and in an individual. a complex network of lateral carpellary bundles is interpolated between the dorsals and the placental vascular system, as in most species (e.g., E£. flavescens, FicurE 18). The presence or absence of an anastomosing lateral carpellary system is apparently not related to the size of the flower. Many of these features of the transeptal ovular system are quite vari- able within a species or an individual. N evertheless, it seems the ovarian vasculature of the species of Eugenia s. s. that I studied in detail can be broadly categorized as follows: Dichotomous comparison of main types of ovarian vasculature in Eugenia s. s. (1) Placental strands pendulous from near tops of loculi, the placental system (including strands and plexuses) thus appearing U- or V-shaped. Placental strands on entering septum usually massive and often two, sometimes additional slender placental strands occurring; FLoripa-type (FIc- URE 15): e.g., E. florida, E. biflora, E. coffeifolia, E. gregii, E. oerstediana, E. salamensis. (1) Placental strands entering sides of septum, the strands on entering septum usually slender. 1972] SCHMID, MYRTACEAE, II. EUGENIA 339 (2) Placental strands usually few, with no connections to other carpellary vasculature (dorsal or lateral bundles); dorsal bundles also not con- nected to other carpellary bundles; Caputt-type (Ficure 13): e.g., E. capuli, E. venezuelensis. (2) Sears strands few to many, with connections to other carpellary culature; dorsal bundles also connected to other carpellary bundles. (3). Dorsal bundles obscure; bundles to style shige to often many (about 12); Unrrtora-type (Ficure 19): E. uniflora (3) Dorsal bundles readily definable; bundles to “biyle . to several (to about 6); common (4) Placental strands usually few, with few connections to other carpellary vasculature, and these connections only to dorsal bundles; anastomosing carpellary system scanty or absent; Conrusa-type (Ficures 1, 3): e.g., E. confusa, E. tikalana. Placental strands often many, with few to many connections to dorsal and/or to lateral carpellary bundles, both of which generally form a simple to complex anastomosing system; FLAVESCENS-type (FIGURE 18): e.g., E. flavescens, E. cartagen- sis, E. winzerlingii. oN EN 7 EXPLANATION OF TABLE II: Floral structure of Eugenia s. s. and Jossinia ialag abbreviations at the head of each column correspond to the following Bif A, Eugenia biflora, Pires & Westra 48813 (micu); Bif B, E. biflora, pam 52-15623 (micH); Bif C, E. biflora, Schultes & Cabrera 14236 (micH); mensis; Sal h, E. salamensis var. hiraeifolia; Tik, E. tikalana; Uni, E. uniflora; Ven, E. venezuelensis: Win, E. winzerlingii; Jos, Jossinia aherniana. CHARACTER 1. The maximum size of the bud near anthesis, width (at widest part of perianth) times length (base to tip of bud), was determined from sec- tioned or cleared material. This creates a bias toward smaller buds since these are often selected for easier manipulation and study. Maximum dimensions as indicated in the taxonomic literature are often much greater. CHARACTER 2. The maximum ak asians of the floral tube above the ~aiedd was determined according taxonomic criteria. Such measurements, however, may include some ovarian ae due to the top of the ovary being sale (see Schmid, 1972b). The amount of egpasratie was graded as follows: glabrous, essentially glabrous, little, or mu CHARACTER 6. The vasculature xe other floral parts is also invariably closed. CHARACTER 7. See Schmid (1972b) for definition of terms. CHARACTER 8. Minor bundles in some species are more or less uniformly dis- persed erat a nae of floral tube in the locular region of flower. CHARACTER 10. The major bundles usually branch distally. CHARACTER 11. See text for descriptions of types of ovarian vasculature. CHARACTER 12. This includes all carpellary vasculature (except dorsal bundles) in locular region. TABLE II. Features of floral structure of Eugenia s. s. and Jossinia * CHARACTER BifA BifB BifC Cap Car Cof Con Fla Flo Ocer Sals Salh Tik Uni Ven Win Jos ORGANOGRAPHY (1) Maximum size bud 40% 35% 18% 17% 25X 228 97% 26K 25K 24% 38K 38K 16K 30% 16K 35X 54X (mm. width < length) 6.2 52 2.8 2.2 af 3.0 4.1 4.5 3.6 3.9 5.0 §2 2.0 5.0 1.8 5.0 TS (2) Max FT prolonged (mm.) 4 Ss See as .0 ns 6 4 4 Soars Bee, SS. 0 t+ 2A3? HistoLocy (3) Scleren in flower — — ~ — — — ~_ ss i se as ee hes RS ode <_ — (4) Scleren in pedicel — — oo — dna -- — x _ _ — —_ + dna (5) Amount pubescence fl m m m li g li(m) g g eg(m) eg m m li g(eg) li g m VASCULATURE FT (6) Vase open or closed cl cl cl cl cl cl cl cl cl cl cl cl cl (7) Mono- or zonocyclic mono mono mono mono mono mono mono mono mono mono mono mono mono mono mono mono mono (8) Minor VBs diffused a + a BS aa ss a 4. (9) Amount minor vasc m m m li m li m m m m m li m m (10) # MBs at base 8(9) 8 8? & 810 8 8 8 8 8(9) 8 8(-10?) 8 8 8 8 8 VASCULATURE OVARY (11) Type flo flo a Cap fla’ flo “con fla flo flo flo flo con uni cap fla dna (12) Amount VT by loculi ii vii vvii vii li m li m m m m m vii m vii li vvm (13) Anastomoses in LCV + + - = + + oe + + + + + _ + = + +(vm) (14) DCBs connected to other ca VT + - =e = + + + + + + + + +. doar — + dna VASCULATURE STYLE (15) # VBs at base 2(3) 2 2 2-20) 2: 4208) 2 3-5 3-6 2 2 2 5-12 2 3(24) 3-5 (16) SVBs extension of both DCBs + + + + “+ + + + + +} + + + dna aa + dna (17) SVBs extension of LCBs occ _ _ - occ _ occ _ + bE - — ~ a - u + (18) SVBs branched _ _ _ oce + occ occ ao u -L + — _ + _ + + (19) VT to tip + + + + + + + am + - + Ss + + nearly + + (20) VT at tip expanded ew + > u + St 2s u 7 + i an + > _— + x * See tant page for explanation of characters and abbreviations of names of $ LS AND ABBREVIATIONS: The following are used, without punctua- tion, to ob entries: X, could not be determined from my mat — +, present or yes; —, absent or no; cap, capuli-type; cl, closed; con, confusa- type; DCBs, dorsal carpellary bundles; dna, does not apply; eg, coscniate glabrous; fl, flower; fla, flavescens-type; flo, fl g, glabrous; li, little ; LCBs, lateral carpellary bundles florida-type; FT, floral tube; $ ahd S, major vascular aximum ; bundles; a occasionally ; scleren, para ass SVBs, stylar vascular undles Scalar bamles: VT, vascular tissue usually; uni, uniflora-type; v, very; vasc, vasculature; VBs, 342 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 These types of ovarian vasculature are typological and do not necessarily indicate relationships since several patterns seem of general occurrence in the genus. OBSERVATIONS — INDIVIDUAL SPECIES Much repetitive descriptive information for the species of Eugenia s. Ss. investigated is relegated to Taste II. Detailed descriptions are given for E. biflora, E. capuli, and E. confusa. Reference should be made both to the tables and to the species descriptions for a complete profile of any one species. The account for each species is based on all the specimens examined (see TABLE I) and represents a range of variability for one or more specimens (individuals) of that species. Organographic and histological data were not heavily emphasized since their apparent minor taxonomic usefulness in a small survey such as mine did not justify the labor involved in their compilation. Detailed study of a number of taxa revealed that several features of vasculature exhibit much variability. Hence, a number of vascular characters ex- amined for Syzygium s. l. (see Schmid, 1972c) were not studied in detail for many species of Eugenia s. s. For example, the level and nature of origin of the dorsal carpellary bundles were not determined for many taxa of Eugenia s. s. since these features are very variable and difficult to define, much more so than in Syzygium s. 1. (see Schmid, 1972a, c). The vascular characters given in TABLE II were selected as the most promis- ing for comparative purposes. The arrangement of the 13 American species examined in the greatest detail is essentially according to similarity of ovarian vasculature, based Eugenia confusa DC. For synonymy see Urban (1895). The two plants examined are anatomically very similar. VASCULATURE. See TABLE II. A single, unbranched bundle extends to the tip of each of the two bracteoles subtending the flower. The floral vasculature forms a closed system. Eight discrete major bundles comprise the vascular cylinder in the base of the flower (Ficures 1, 2, 6). Through- out the extent of the floral tube, these eight bundles exhibit a monocyclic arrangement (defined in Schmid, 1972b) and frequently branch, producing mostly small minor strands that anastomose profusely in a very irregular pattern (FicuREs 2, 6, 7, 12). Distally, the eight major bundles, which usually extend into the eight perianth parts, branch extensively in an irreg- Ficures 1-3. Eugenia confusa (Proctor 23563, MicH). Ficures 1, 2, Reconstruction of part and counterpart of a cleared flower cut in plane of septum between the two carpels. Minor vasculature omitted from Ficu URE 1, vasculature of FicuRE 2 com- plete. Numerals at edge of Ficure 1 indicate approximate levels of transections shown in FicurEs 6-12. URE 3, ogc aia of Ficure 1 showing placental system formed by five placental strands — sed central axis bundles, ABBREVIATI tral axis bundle; d, dorsal carpellary bundle; 1, limits of ovarian locule; r trace; p, petal bundle; pp, placental bee: ps, placental strand; s, sepal bundle; sb, stylar bundle; st, stamen Ecce, Unlabeled magnification bar equals 2 VING9NG ‘Il ‘AVADVLUAW ‘GINHOS : {ve 344 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ~ FIGURE 5. FiGcure 5. Transectional reconstruction of flower in FIGURE 4, 1972] SCHMID, MYRTACEAE, II. EUGENIA 345 ular manner and supply narrow traces to the stamens and generally thicker strands to the perianth segments (Ficures 1, 2). The ovarian vasculature exemplifies the ConrusA-type (Ficures 1, 3). It is complex and exceedingly variable, much more so than originally thought (Schmid, 1970). The several placental strands originate from various sources in the vascular cylinder of the floral tube, that is, directly from a major bundle or more typically from the anastomosing strands pro- duced by the major bundles (Ficures 1, 8). In addition, they originate at various levels in the flower, from near its base to within the placental region (FicurE 1). As they ascend and traverse the tissue next to the loculi, the placental strands occasionally produce short branches, but eventually the former enter the middle part of the septum, turn down- ward, and terminate at two placental plexuses (Ficures 1, 3, 5, 8, 9). Each placental plexus is usually formed by the fusion and subsequent proliferation of the placental strands of the same carpel (FicurEs 3, 5, 8). There is generally little contact between the placental strands of adjacent carpels, although these strands do come very close together in the septum (FicurEs 5, 8, 9). Several bundles, designated central axis bundles, ex- tend from the four corners of the placental system and connect with the two dorsal carpellary bundles near the tops of the loculi (Ficures 3, 5, 10-12). Ficure 5, based on an atypically tricarpellate flower, illustrates the variability of vasculature in the placental region. The two dorsal carpellary bundles usually originate below the loculi and occasionally produce short branches that terminate in the ovary (Fic- URES 1, 2, 7). The dorsals, sometimes accompanied by a third strand, vascularize the style (Ficures 1, 4). The stylar bundles generally do not branch, but expand near the tip of the style (Ficures 1, 4). Eugenia capuli (Schlecht. & Cham.) Berg Synonyms include Myrtus capuli Schlecht. & Cham. and Eugenia con- trerasii Lundell (see McVaugh, 1963a, b). The only difference detected between the two collections I studied was that the stylar bundles of one specimen (Breedlove 14407) were distally expanded and occasionally branched whereas those of the other (Gentle 5044) were not. The latter collection had been identified as E. contrerasii Lundell. The fact that the two collections are so very similar anatomically supports McVaugh’s (1963a, b) inclusion of E. contrerasii in the synonymy of E, capuli. from levels indicated in Ficure 1. Only locular portions of flowers are shown in Ficures 8-11 since outer vasculature is similar. FicurE 7 is above level of origin of dorsal carpellary bundles, these omitted from F1curEs 8-11. In FIGURES 6, 7, and 12 the eight largest bundles correspond to the major bundles, four sepal bundles alternating with four petal bundles. ABBREVIATIONS: c, compitum; cb, central axis bundle; d, dorsal carpellary bundle; p, petal bundle; pe, petal; pp, placental plexus; ps, placental strand; s, sepal bundle; sb, stylar bundle; sc, stylar canal; se, sepal; sta, stamen. Magnification bars equal 250 u. 346 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Histotocy. The floral tube and sepals are puberulent (see Schmid, 1972b, Figures 1, 2). VascuLaTurE. See Taste II. A single, unbranched bundle extends to the tip of each of the two nearly contiguous bracteoles. The vasculature of the flower forms a closed system. Eight discrete major bundles occur in the tip of the pedicel and the base of the flower (Ficure 13). In con- trast to Eugenia confusa (FicuRE 2), these eight major bundles produce relatively few minor, anastomosing strands throughout the extent of the floral tube (Ficure 13). Distally, the eight major bundles divide pro- fusely in an irregular manner and supply strands to the four sepals, the four petals, or the many stamens (FicurE 13). The ovarian vasculature exemplifies the CaPuLi-type (Ficure 13). Originating either directly from a major bundle or generally from anas- tomoses produced by the major bundles, a variable number of placental strands arise at various levels in the flower and enter the ovarian tissue (Ficure 13). At the ends of the septum, these placental strands fuse into two compound placental strands that traverse the septum. The placental vasculature usually does not exhibit the arching pattern characteristic of Eugenia confusa. In the center of the ovary, the two compound placental strands each divide and form two placental plexuses. The vascular tissue in the placentae thus displays a diamond-shaped pattern, with a parenchy- matous core (FIGURE 13). There are no vascular connections between the placental vascular system and the dorsal carpellary bundles, unlike the usa the ovary and only occasionally, distally, in the style (Ficure 13). The bundles usually expand near the tip of the style. Eugenia tikalana Lundell McVaugh (1963b, p. 366) remarked that “this species is superficially al- most indistinguishable from E. capult.” Anatomically, the two species ture (TABLE II, characters 11 and 14), which is of the Conrusa-type (Ficures 1, 3). In E. tikalana several placental strands enter the sides Eugenia venezuelensis Berg Synonyms include Eugenia origanoides Berg. The name usually used is E. origanoides (see McVaugh, 1958, 1963a, b, for full description and err Reconstruction of cleared flower. — ated, Ficures 13 and 14. Ficure 13, Eugenia capuli (Breedlove 14407, micH). and part of flower with two major bundles omitted for clarity. Ovarian ane not i Minor vasculature four petal bundles. Ficure 14, Eugenia tikalana. Longisection of flower n plane of septum, showing central placental rue ABBREVIATIONS: c, compitum; d, dorsal carpellary bundle; p, petal beadle: om nace eae s, sepal bundle; sb, stylar bundle; st, stamen trace. Magnification ‘bars equal 250 w and 310 & respectively. 348 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 synonymy). However, McVaugh informs me (1971) that he saw the type of E. venezuelensis and it is the same as that of E. origanoides. Eugenia venezuelensis is distinguished from E. capuli or E. tikalana chiefly by its more abundant and conspicuous pubescence (McVaugh, 1963a, b), but the hairs of the three species are of very similar morphology (as in Schmid, 1972b, Figures 1, 2). Anatomically, the three species are very similar (see Taste II) although the placental vasculature of E. venezuelensis more closely resembles that of E. capuli (FicuRE 13). VASCULATURE. See TaBLE II. At each end of the septum two placental strands fuse into a compound strand that terminates at a single placental plexus. The two stylar bundles, which end as far as 210 microns from the stigma, terminate further from the stigma than do the stylar bundles of other species of Eugenia s. 5. examined. The delicate vascular strand (one or two vessel elements thick) of the “filament expands con- siderably (over 10 times) in the connective and abuts against a terminal secretory cavity that is up to 110 microns in dimension. The staminal vasculature of E. capuli and E. tikalana is very similar. Eugenia florida DC. For Synonymy see McVaugh (1958, 1969). The two collections I ex- amined are anatomically very similar. See also remarks under Eugenia oerstediana. is that in E. florida there are numerous tiny bundles between the loculi and the vascular cylinder of the floral tube. The ovarian vasculature exemplifies the FLortpa-type (FicuRE 15). Near the bottoms of the loculi many minor carpellary bundles bend in- ward from the innermost part of the floral tube, become arranged in a ring with the two dorsal carpellary bundles, and distally anastomose with each other and with the dorsals (Ficure 15), Two (usually) massive placental strands originate from this anastomosing system just above the loculi and descend in opposite ends of the septum to the two placental plexuses (Fic- URE 15). However, unlike other species which have the F LORIDA-type of ovarian vasculature, in Eugenia florida there are no additional slender pla- cental strands entering at lower levels. Above the loculi the two dorsals and a variable number of other bundles (some containing only unlignified ele- ments) often fuse with each other before entering the style (Ficure 15). Eugenia oerstediana Berg Synonyms include Eugenia vincentina Krug & Urban ex Urban, E conzattia Standl., E. cocquericotensis Lundell, E. petenensis Lundell, and 1972] SCHMID, MYRTACEAE, II. EUGENIA 349 ture omitted. ApBreviaTions: d, dorsal carpellary bundle; 0, ovular trace; ps, placental strand; sb, stylar bundle. Magnification bar equals 250 u. E. eutenuipes Lundell (see McVaugh, 1963a, b). It should be noted that Berg’s (1855-56) E. oerstedeana is an orthographic error for E. oerstediana, Of the three collections sampled, one had been identified as Eugenia oerstediana (McVaugh 23418), another as E. conzattii (King 4405), a third as E. cocquericotensis (Gentle 2533). McVaugh (1963a, b) con- sidered the latter two names to be synonyms of the former. Floral anatomy supports this treatment since all three collections are very similar. Any two flowers, either from the same plant or from two different col- lections, may vary in a few very subtle respects, but analysis of a total of 27 flowers for the three collections indicates most such variations merely represent different points on a morphocline. Gentle 2533, how- ever, differs slightly and consistently from the other two collections, chiefly in having (1) the dorsal carpellary bundles only weakly connected to the other carpellary bundles and (2) the two massive placental strands attached usually just below (as opposed to just above) the tops of the loculi. King 4405 differs from the other two collections in having (1) an extremely short floral tube (.1 mm. versus .3—.4 mm.), (2) very many tiny bundles occurring between the loculi and the vascular cylinder of the floral tube, and (3) the minor carpellary bundles occurring further from the loculi than do comparable carpellary bundles of the other col- lections. Gentle 2533 and King 4405 both differ from McVaugh 23418 350 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 in having, respectively, four and six (rather than a variable number) strands at the base of the style. Taxonomically, Eugenia oerstediana appears to be closely allied to E. florida (McVaugh, 1963a, b, 1969) and perhaps derived from it (McVaugh, 1963a). McVaugh (1963a, p. 448) questioned “the propriety of rec- ognizing E. oerstediana as a species distinct from E£. florida” since the two are “virtually indistinguishable” except in ovule number and in characters of the inflorescence and pedicel. Anatomically, my three collections of E. oerstediana are exceptionally similar to my two collections of E. florida (see TABLE II). The only consistent difference I found between the two species is that the placental strands in E. florida are attached at a slightly higher level so that the placental system in side view appears more U- shaped, with nearly vertical placental strands (FicurE 15), whereas that of E. oerstediana appears V-shaped, with sloping placental strands. One collection of E. oerstediana (McVaugh 23418), in particular, is anatomical- ly nearly indistinguishable from E. florida. It seems, then, that E. florida is best regarded as a widely distributed complex comprised of such scarce- ly distinguishable forms as E. oerstediana, E. conzattii, and E. cocqueri- cotensis, to mention just the species for which anatomical evidence is now available. HistoLocy. Only the sepals are pubescent (ciliate, with hairs like those in Schmid, 1972b, Figure 1). VASCULATURE. See Taste II. As explained above, the vasculature is very close to that of Eugenia florida. Eugenia biflora (L.) DC. McVaugh (1958, 1969) and Urban (1895) give extensive synonymy. I examined five collections for Eugenia biflora (see Taste I). Since these are rather dissimilar, there are three entries for this species in Taste II, as explained below. One collection (Black 52—15623), listed separately in Tasie II as Eugenia biflora “B”, differs considerably from the other four collections (see also Taste II), viz.: Although the ovarian vasculature is of the FLoripa-type (Ficure 15), the pendulous placental strands are rather narrow, and there are very few connections between the dorsal bundles and other carpellary bundles. Black 52—15623 also differs from the other collections in some less important respects: (1) Its staining properties are very different. (2) Much of the septum is often very thin, consisting of one to three layers of cells. A second collection (Schultes & Cabrera 14236) is even more differ- ent and thus is also listed separately in Taste II (as Eugenia biflora “C”). Differences from the other collections studied include the following: (1) Only four (as opposed to many) ovules are present per flower. (2) The number of major bundles is difficult to determine since these branch repeatedly throughout the floral tube. The relationship between the vas- culature of the floral tube and that of the perianth parts is therefore ob- 1972] SCHMID, MYRTACEAE, II. EUGENIA 351 scure. (3) Only one strand enters each sepal, but more than one supplies each petal. (4) The ovarian vasculature is quite unlike that of the other collections examined, but I cannot detail how it differs as my prepara- tions are of poor quality. (5) The lateral carpellary system is only weak- ly developed, with apparently no anastomoses or connections to the dorsal bundles. The three other collections (Hart 1048; Pires & Westra 48813; Steyer- mark & Wurdack 31), designated Eugenia biflora “A”, are very similar, particularly in their vasculature. The entry E£. biflora “A” in Tasie II is based on Pires & Westra 48813, which was studied in the greatest de- tail. These three collections differ chiefly as follows: (1) The maximum prolongation of the floral tube above the ovary was 0 mm. and .6 mm. for Hart 1048 and Steyermark & Wurdack 31, respectively. (2) In Pires & Westra 48813 usually only the two dorsal carpellary bundles continue to the tip of the style, but in the other two collections the two dorsals are often accompanied into the style by additional carpellary bundles. Ranging from southern Mexico and the West Indies to Bolivia and northern Brazil (McVaugh, 1963b, 1969), Eugenia biflora is “probably the most widespread, and certainly the most variable, of any native [American] species of the family” (McVaugh, 1969, p. 167; see also Mc- Vaugh, 1958). The four South American collections I examined (see TABLE I) are cited under E. biflora by McVaugh (1969) in his most re- cent work on the Myrtaceae. At my request, McVaugh reéxamined (Au- gust, 1971) the collections designated here as E. biflora “B” and “C”, but he still thinks they are good E. biflora. In view of the notorious vari- ability of this species, it is not surprising that E. biflora “A”, “B”, and “C” differ markedly (see TABLE II, which does not do justice to the dif- ferences between “A”, “B”, and “C”). However, perhaps these differ- ences are no more significant than those evident between E. salamensis varieties salamensis and hiraeifolia (see TABLE II). One collection of E. biflora “A” (Hart 1048) had been determined as var. wallenii (Macf.) Krug & Urban, and it is cited as such in Urban (1895). However, I am unable to see any clearcut anatomical differences between this variety and the other two collections of E. biflora “A” (Pires & Westra 48813; Steyermark & Wurdack 31). In view of the anatomical differences between Eugenia biflora “A”, “B”, and “C”, but in spite of the organographical similarities, this species apparently deserves taxonomic reappraisal. Histotocy. The dense silky or silky-strigose pubescence of simple hairs (see Schmid, 1972b, Figure 12, also like Figure 9) covering the floral tube and sepals is distinctive for this species (McVaugh, 1969). VascuLature. The following description, based primarily on Pires & Westra 48813, pertains mainly to Eugenia biflora “A” and “B”. For variations see above discussion and also Taste II. A single trace (several in Steyermark & Wurdack 31) supplies each bracteole. The vasculature of the floral tube and style is like that of Eu- 352 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 WW Al\ FIGURES 16, 17, Eugenia biflora (Pires & Westra 48813, mic). Ficure 16, Reconstruction of vasculature to and in three perianth complete. Onl j 1972] SCHMID, MYRTACEAE, II. EUGENIA 353 genia confusa (see TABLE II). However, as in several other species of Eugenia s. s. (e.g., E. florida, E. oerstediana, E. uniflora), in E. biflora there is a very regular relationship between the major bundles of the floral tube and the vasculature of the perianth parts (Ficure 17). Eight major bundles, four sepal bundles alternating with four petal bundles, oc- cur throughout the floral tube and produce many small minor, anastomos- ing strands (FicurE 16). Distally, the four sepal bundles continue up the middle of the four sepals (FicurEs 16, 17). In contrast, each of the four petal bundles trichotomizes (Ficures 16, 17), one of the three bundles that results entering each petal, the other two (lateral) bundles functioning as laterals of the two adjacent sepals. The bundle entering each petal branches profusely (FicurE 16). The medial and two lateral bundles in each sepal anastomose extensively (FicuRE 16). A variable number of smaller bundles also enter the sepals (Ficure 16). The ovarian vasculature is of the FLoripa-type (Ficure 15), but ex- hibits considerable variability, even in the same plant. The lateral car- pellary system, which is not very extensive, fuses with the dorsal bundles near the tops of the loculi. Two massive placental strands descend from points just above the loculi and terminate at a single placental plexus perforated by much parenchyma. The placental strands originate in a variety of ways, in some flowers mainly or exclusively from the two dor- sal carpellary bundles. In addition slender placental strands frequently enter from the sides of the septum. The dorsal bundles, which do not branch, originate at various levels, usually in the locular region, but often even near the base of the flower. Eugenia coffeifolia DC. See McVaugh (1969) and Urban (1895) for synonymy. Data were compiled mostly from the plant from St. Lucia since my preparations of the other collection were of poor quality. As far as I could tell, however, the two collections are very similar. OrcANocGRAPHY and HistoLocy. The presence of only two ovules per locule is unusual in Eugenia s. 1. (McVaugh, 1969), although as few as two or three ovules per locule also do occur in E. oerstediana (McVaugh, 1963a, b). The floral tube is strigose, with hairs similar to those in Figure 7 in Schmid (1972b). The style is sparingly pilose. VaAscULATURE: See TABLE II. The ovarian vasculature is of the FLormpa- type (Ficure 15). Near the tops of the loculi a variable number of strands of the anastomosing lateral carpellary bundles, often along with strands derived from the dorsal carpellary bundles, fuse into two large placental strands, which terminate at a single placental plexus containing little or no interfascicular parenchyma. Occasionally, additional, slender placental strands enter the septum via its sides and fuse with the placental plexus. The vasculature of the other floral parts resembles that of Eu- genia capuli (see Taste II). 354 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Eugenia salamensis Donn. Sm. Synonyms include Psidium rensonianum Standl., Eugenia mexiae Standl., and E. hiraeifolia Standl. (see McVaugh, 1963a, b). I examined material of var. salamensis (this labelled E. mexiae) and var. hiraeifolia (Standl.) McV. Unless otherwise noted, the following comments apply to both va- rieties. Although both varieties are “readily distinguished” on the basis of vegetative characters in particular, McVaugh (1963a, p. 457) considered them “‘surely conspecific.” Although it is a matter of taxonomic opinion, there seem sufficient anatomical differences between the two varieties, as noted below, that both could deservedly be retained as separate species, particularly if such very (and even more) similar taxa as Eugenia capuli and E. venezuelensis (see TABLE IT) are regarded as distinct. ORGANOGRAPHY and Histotocy. Dibrachiate trichomes (see Schmid, 1972b, Figure 13) with short to long, equal or unequal, often contorted arms are very abundant on the floral tube and sepals. Variety /iraeifolia was very tanniferous, var. salamensis only slightly so. The floral tube of the former is much more prolonged above the ovary than that of the lat- ter (ca. 1.1 mm. maximum versus 0.2 mm. maximum, respectively). The only floral difference that McVaugh (1963a) noted for the two varieties was the slightly longer styles of var. salamensis. In buds, styles of var. salamensis are therefore doubled back whereas those of var. hiraeifola are erect or nearly so. VASCULATURE. See TABLE II. The ovarian vasculature of var. sala- Eugenia flavescens DC. See McVaugh (1969) for synonymy. Histo.ocy. This is the only American species of Eugenia s. 1. I studied that lacked secretory cavities in the anthers. VASCULATURE. See TaBLE II. The ovarian vasculature exemplifies the FLAVESCENS-type (FIGURE 18). An extensive lateral carpellary system consisting of slender anastomosing bundles girdles the loculi. From few 1972] SCHMID, MYRTACEAE, II. EUGENIA 355 to many (up to about 11 per flower) placental strands depart from this lateral carpellary system, enter the septum, where they frequently anasto- mose, and terminate at the placental plexus (FicurE 18). The dorsal car- pellary bundles are connected to the lateral carpellary system at a number of points. Other features of floral vasculature resemble those of Eugenia confusa (see TABLE II). Eugenia cartagensis Berg HistoLocy. Phloem fibers occur throughout the pedicel, sclereids chief- ly distally. VASCULATURE. See TABLE II. The ovarian vasculature is of the FLAVESCENS-type (FIGURE 18). There are usually few placental strands, and in the septum these often fuse for a considerable extent. Bundles of the lateral carpellary system, which is not very extensive, anastomose chiefly near the tops of the loculi. The lateral and dorsal carpellary bun- dles are interconnected only at the tops of the loculi. Other features of floral vasculature are very much like those of Eugenia confusa or E. fla- vescens (see TABLE II) except that ten major bundles occur in some flow- ers of E. cartagensis. Eugenia winzerlingii Standl. One collection (Lundell 851) was analyzed most thoroughly. The three collections examined, however, are anatomically similar. ORGANOGRAPHY and Hisrotocy. The flowers are borne on filiform pedicels up to 20 mm. long (McVaugh, 1963b). The pedicels contain numerous narrow thin-walled phloem fibers commonly .4 to .5 mm. and often up to about .8 mm. long. One of the 14 flowers examined was tri- carpellate; its placental vasculature was similar to that of the other flow- ers, Unlike other species I studied, the walls of the loculi of Eugenia winzerlingii are lobate. VASCULATURE. See TABLE II. The ovarian vasculature is of the FLAVE- SCENS-type (Ficure 18.) However, the following differences from Eugenia flavescens were noted: The lateral carpellary system is less extensive. Two separate placental strands, as in E. confusa (Ficures 1, 3), are present. Placental strands occasionally originate directly from the dorsal carpel- lary bundles. The vasculature of the floral tube and style resemble, respectively, that of Eugenia confusa (or E. flavescens) and E. florida (see Tape II). A continuous ring of vascular tissue frequently occurs at the tip of the style. Eugenia uniflora L. Synonyms include Myrtus braziliana L., Plinia rubra L., P. peduncula- ta L. f., Eugenia michelii Lam., Stenocalyx michelit (Lam: ) Berg, and Eugenia decidua Merr. [also ‘ ‘Myrtus indica” Tilli (1723); “Eugenia indica” Mich. (1729)] (see Berg, 1855-56; McVaugh, 1963b, 1969; Mer- rill, 1950; Urban, 1895). The two collections studied are similar. 356 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 re (Pen NF 1972] SCHMID, MYRTACEAE, II. EUGENIA 357 Eugenia uniflora L., the type of Eugenia Linnaeus (1753),1 is a culti- gen commonly known as pitanga, Brazil cherry, Cayenne cherry, Florida cherry, or Surinam cherry. This species is one of the very few native American species of Eugenia s. J. to have been studied anatomically (see Petit, 1908, for the histology of seed and fruit). OrcANocRAPHY and Histoxocy. The filiform pedicels (up to 18 mm. long, McVaugh, 1963b) contain long narrow phloem fibers, as in Eugenia winzerlingit. Most flowers of E. uniflora are entirely glabrous, but one collection (Proctor 19469) I examined has a hairy staminal disc (with hairs as in Schmid, 1972b, Figure 6, excluding the hair on extreme left of figure). The floral tube is conspicuously eight-angled, with a large (up to 70 p in radial dimension) major bundle opposite each rib. The loculi are very large in relation to the rather narrow walls of the inferior ovary. Eugenia pbleurocarpa, with a slightly eight-angled floral tube, is the only other species of Eugenia s. 1. (31 studied in all, Schmid, 1971, 1972a—c) to have either of these characters. The fruit of E. uniflora is also prominently eight-costate. ASCULATURE. See TABLE II. The vasculature of the floral tube is similar to that of Eugenia biflora (see TaBLeE II, Ficures 16, 17). How- ever, other features of vasculature of E. uniflora are quite distinctive, par- ticularly the ovarian vasculature, which exemplifies the UNrIFLoRA-type (FicurE 19). Several placental strands originate from various sources and at various levels in the floral tube (F1cure 19), much as in E. con- fusa (Ficure 1). On entering the septum, the placental strands usually fuse into two large compound placental strands. These turn downward as they traverse the septum and then terminate at a single placental plexus usually consisting of a solid mass of vascular tissue (FIGURE 19). Unlike the species of Eugenia s. s. previously described, discrete dorsal carpellary bundles are not readily definable. Usually two large bundles originate in the positions where the dorsals would ordinarily be expected. These bundles, however, each quickly split into several fine, anastomosing strands (FicuRE 19), most of which eventually continue into the style. Occasionally, other carpellary bundles also supply the style. There are frequent connections between the placental vasculature and the strands Supplying the style (FicurE 19). The many stylar bundles (as many as 12 in the base of the style) occasionally branch and often form a con- tinuous ring of vascular tissue at the stylar apex. A comparable situation Occurs in certain species of Alangium (Eyde, 1968), but here, in contrast to E. uniflora, the two groups of fine, anastomosing strands taking the * Authorities and dates of publication of genera follow the Index Nominum Generi- corum, which should be consulted for details. Other vasculature of floral tube and ovary and outlines of ovarian loculi omit- ted. Four sepal bundles alternate with four petal bundles. BREVIATIONS: 0, ovular trace; p, petal bundle; ps, placental strand; s, sepal bundle. Magnifica- tion bars equal 250 xu, 358 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 place of the dorsals each fuse, forming a pair of well defined bundles that traverse the style. OTHER TAXA WITH TRANSEPTAL OVULAR SYSTEMS All the species of Eugenia s. s. listed below have transeptal ovular systems. In addition, all taxa have closed vascular systems, with the usu- ally eight major bundles of the floral tube exhibiting a monocyclic ar- rangement. Vascular tissue extends to the apex of the style. The flowers lack sclerenchyma. However, the precise nature of these and other char- acters was not determined for these species for several reasons, in part be- cause my clearings and many of my sections were of poor quality, and also because most of the buds of these species are very large and have ex- tensive vascular tissue. A few distinctive features, chiefly of pubescence, are noted below. (1) Eugenia acapulcensis Steud. See McVaugh (1963a, b) for synony- my. My material is glabrous, except for a slightly hairy staminal disc (with hairs like those in Schmid, 1972b, Figure 5). (2) Eugenia aeruginea DC. Synonyms include E. fadyenii Krug & Ur- ban ex Urban (see McVaugh, 1963b). The floral tube and sepals have an appressed, dense pubescence of dibrachiate hairs (as in Schmid, 1972b, Figure 11). (3) Eugenia duchassaingiana Berg. The sepals and floral tube are very pubescent, with hairs as in Figure 7 in Schmid (1972b). (4) Eugenia gregii (Sw.) Poir. Synonyms include Myrtus gregii Sw., Greggia aromatica Solander ex Gaertn., and Eugenia sieberiana DC. (see Urban, 1895). Eugenia gregii is the type of the segregate genus Greggia Solander ex J. Gaertner (1788). This species has very distinctive infun- dibuliform hairs (see Schmid, 1972b, Figure 8) that form a dense mat on the floral tube and abaxial surfaces of the perianth parts. Other than these unusual hairs, however, the species certainly does not have any feature that might justify acceptance of the segregate genus Greggia. Super- ficially, the ovarian vasculature is of the Frortpa-type (FicuRE 15). (5) Eugenia mandevillensis Urban. Only the staminal disc and style are hairy (see Schmid, 1972b, Figure 9). (b) Abundance: The secretory cavities of E. mandevillensis are more numerous than those of any other species of Eugenia s. 1., with the pos- sible exception of clove. In many parts of the flower, which is about 8— 1972] SCHMID, MYRTACEAE, II. EUGENIA 359 9 mm. long in bud, the secretory cavities are so numerous that they occur as chains separated only by a few layers of epithelial cells. (c) Dis- tribution: As in clove, the secretory cavities in the floral tube and the top of the ovary of £. mandevillensis frequently occur in two and even three intermixed layers (compare with Figures 48, 49, 55, 56, in Schmid, 1972c). Many cavities occur in the anthers (up to about seven per anther), filaments, and style, again as in clove. Unlike those of any other species of Eugenia s. 1., however, the secretory cavities of E. mande- villensis also occur in the ovarian tissue adjacent to the loculi and below them. (6) Eugenia muricata DC. See McVaugh (1969) for synonymy. Many appressed-ascending, silky hairs (see Schmid, 1972b, Figure 7, also as in Figure 12) cover the floral tube and sepals. (7) Eugenia pleurocarpa Standl. The floral tube and sepals are tomen- tose, covered by a pubescence of dibrachiate to mostly simple hairs of variable morphology (see Schmid, 1972b, Figures 10, 11). The flowers are very tanniferous, much more so than any other species of Eugenia 5. 5. I studied. See also remarks under Eugenia uniflora above. (8) Jossinia aherniana (C. B. Rob.) Merr. Synonyms include Eugenia aherniana C. B. Rob. and E. melastomoides Elm. (see Merrill, 1950). The simple to dibrachiate hairs are figured in Schmid (1972b, Figure 6). For vasculature see TABLE II. The floral vasculature is similar to that of the species of Eugenia s. s. described above except, that the flowers of Jossinia aherniana seem much more heavily vascularized. The placental vasculature of this species, however, is supplied not only transeptally by strands from the periphery of the septum, as in the previously described species of Eugenia s. s., but also by a few bundles entering from the base of the septum, somewhat as in Syzygium s. l. (see Schmid, 1972c). In sharp contrast to Syzygium s. 1., however, the characteristically massive axile strand is lacking. In all, the ovarian vasculature most closely re- sembles the FLAVESCENS-type (FicurE 18). (9-10) Only fruiting material of the following two species, both from the Old World, was available. Both taxa have closed vascular systems and about eight major bundles of the floral tube in a monocyclic arrangement. Considerable vascular tissue occurs in the floral tube. The staminal discs of both species are hairy, more or less as in Figures 6 and 9 in Schmid (1972b). Sclerenchyma occurs only in the seeds of Jossinia palumbis but throughout the fruit of Eugenia whytei, which has both fibers and sclereids. The vascular supply to the ovules seems to be of the transeptal type, but this conclusion is based on the examination of large, mature, one-seeded fruits in which there had been considerable displacement and disruption of the septum. (9) Eugenia whytei Sprague in Stapf. Chattaway (1959) conjectured that this might represent a species of Jossinia rather than Eugenia. 360 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 (10) Jossinia palumbis (Merr.) Diels. Synonyms include Eugenia pa- lumbis Merr. (see Diels, 1921) DISCUSSION AND CONCLUSIONS (Schmid, 1971, 1972b). While there seems to be convincing evidence that Eugenia gregii is the type species of the segregate genus Greggia So- lander ex J. Gaertner (1788). However, only the authors who proposed it seem to have accepted this segregate. The floral anatomy of E. gregii is entirely consistent with its placement in Eugenia s. s. Stenocalyx Berg (1856), won little credibility among use of so vague and variable McVaugh (1968) concurred. characterized by precocious flowers, has also botanists. Bentham (1869) condemned the a feature, and nearly a hundred years later However, Kausel (1957, 1966), the arch splitter of the Myrtaceae, still assesses Stenocalyx as valid. Anatomically, Eugenia uniflora (= Stenocalyx michelii), the type species of Stenocalyx (as well as of Eugenia Linnaeus, 1753), isc To be fair, however, this species is undeniabl genus (see Schmid, 1972b). Actually, in vi genia s. s. exhibits, I am not at all certain that E. uniflora and its relatives *See McVaugh (1968) for an e can Myrtaceae, and also Kausel ( generic synonymy of Eugenia xcellent discussion of the taxonomy of the Ameri- 1957), McVaugh (1968), and Niedenzu (1893) for 5. S$. 1972] SCHMID, MYRTACEAE, II. EUGENIA 361 are even worthy of subgeneric rank, the status to which Niedenzu as- signed them (1893) in his taxon Macrocalyx. Indeed, Bentham (1869) long before pointed out the great polymorphy of calyx structure in Eu- genia, which became the basis for Niedenzu’s characterization. The floral anatomy of the Old World genus Jossinia Commerson ex De Candolle (1828) is very similar to that of the American species of Eugenia 5. S$. (see TABLE II), differing only in the nature of the vascular supply to the ovules. All the evidence from both vegetative and reproductive or- ganography and anatomy now available (see Schmid, 1971) demonstrates that Jossinia is so very similar to the American species of Eugenia s. s. that segregation of Jossinia as a genus seems unwarranted. Jossinia may well represent a residue of Old World species of Eugenia s. s. that, in some of its taxa, exhibits a rather primitive transeptal ovular system, one perhaps transitional between the axile ovular system of the Old World Syzygium s. 1. and the transeptal ovular system of the New World species of Eugenia s. s. (Schmid, 1971). These aspects will be elaborated in a sub- sequent publication in this series (Schmid, in preparation). SUMMARY There seems to be no anatomical basis for the recognition of the fol- lowing generic segregates of Eugenia s. s.: Greggia, Jossinia, and Steno- calyx. Stenocalyx admittedly occupies a somewhat isolated position with- in Eugenia s. s., but the differences of the former merely conform to the great organographical and anatomical diversity of a very kaleidoscopic genus. The organography and anatomy of the Old World Jossinia species are very similar to those of the American species of Eugenia s. s. The following conclusions were drawn about the relationships of various species of Eugenia s. s. (1) Eugenia capuli, E. tikalana, and E, venezuelen- - Cocquericotensis. (3) Eugenia salamensis varieties salamensis and hiraeifolia seem sufficiently distinct so that at least varietal and probably specific status is justified. (4) Eugenia biflora was found to contain a number of anatomically different but organographically similar forms and deserves taxonomic reappraisal. ACKNOWLEDGMENTS This study is based on a portion of a dissertation submitted in partial fulfilment of the requirements for the degree of Ph.D. at The University of Michigan, Ann Arbor. This research was supported in part by a grant from the National Sci- ence Foundation, GB-8212, to N. G. Hairston, The University of Michi- gan, for research in Systematic and Evolutionary Biology. I am greatly 362 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 indebted to Dr. Rogers McVaugh, Dr. Warren H. Wagner, Jr., and my wife Marvin for their critical reading of the manuscript, and to my wife also for preparing the drawings. In addition, I thank Dr. McVaugh, Cura- tor of Phanerogams, University Herbarium, The University of Michigan, for making specimens available for study. LITERATURE CITED BENTHAM, G. 1869. Notes on Myrtaceae. Jour. Linnean Soc. Bot. 10: 101-166. Bere, O. 1855-56. Revisio Myrtacearum Americae hoc usque cognitarum s otzschii “Flora Americae aequinoctialis’ exhibens Myrtaceas. Linnaea 27: 1-472 CHATTAWAY, M. M. 1959. es Mice’ of bark. VII. Species of Eugenia (sens. lat.). Trop. Woods #111: Drs, L. 1921. Die ee ee Bot. Jahrb. 56: 529-534. Eype, R. H. 1968. Flowers, fruits, and phylogeny of Alangiaceae. Jour. Arnold Arb. 49: 167-192. KAUSEL, E. 1957. Myrtaceae. No. 28 im J. Angely, Catalogo e estatistica dos géneros pnd ee Curitiba, Parana, Brazil: Instituto Para- naense de Botanica. . 1966. Lista ie las Mirtaceas y Leptospermaceas Argentinas. Lilloa 32: 323-368. McVaucu, R. 1958. Myrtaceae. In: Flora of Peru. Field Mus. Nat. Hist., Bot. Ser. 13(pt. 4): 567-818. ———. 1963a. Tropical American Myrtaceae, II. Notes on generic concepts and oe of previously unrecognized species. Fieldiana Bot. 29: 391-532 : 1963b. Myrtaceae. Jn: Flora of Guatemala. Fieldiana Bot. 24(7): i-viii, 283-405. . . 1968. Sg genera of American Myrtaceae —an interim report. Taxon 1 oh fg In: B. Macurre et al., The botany ~ the Guayana Berit Part VIII. Mem. N.Y. Bot. Garden 18: 55-28 MERRILL, E. D. 1950. Readjustments in nomenclature of Seine Eugenia species. Philippine Jour. Sci. 79: 351-430. NIEDENZU, F. 1893. Myrtaceae. Pp. Sa im: A. Engler & K. Prantl, Die natiirlichen Pflanzenfamilien, III. Abt. 7. Leipzig: Wilhelm purgags Petit, L.-A. 1908. Recherches sur la structure anatomique du fruit et de la graine des Myrtacées. Doctoral Thesis. Univ. Paris, Ecole ere Phar- macie, Année 1907-1908, #8. SCHMID, R. 1970. Comparative floral etigs of Myrtaceae, with emphasis on Eugenia and its segregates. Am. Jour. Bot. 57: 744, 745, viii. (Abstract.) . 1971. Floral anatomy or fag sensu lato Cdrtaceie’. Ph.D. Thesis, The University of Michigan. . 1972a. Floral bundle fusion and vascular conservatism. Taxon 21: (In press. 197 2b. A resolution of the Seca Sou controversy (Myrtaceae). Am, Jour. Bot. 59: 423-436 1972] SCHMID, MYRTACEAE, II. EUGENIA 363 : nbs Floral anatomy of Myrtaceae. I. Syzygium. Bot. Jahrb. 92: (In URBAN, L 1895, Additamenta ad cognitionem florae Indiae occidentalis. Particu- la II. Myrtaceae. Bot. Jahrb. 19: 562-681 Present address: Address after 1 Sept., 1972: SMITHSONIAN FELLOW DEPARTMENT OF BOTANY DEPARTMENT OF BOTANY UNIVERSITY OF CALIFORNIA SMITHSONIAN INSTITUTION BERKELEY, CALIFORNIA 94720 WASHINGTON, D.C. 20560 364 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 VENATION PATTERNS IN THE LEAVES OF EPHEDRA ADRIANCE S, FOSTER THE VIEW IS WIDELY HELD that the vascular supply of the leaf in Ephedra consists of two traces associated with a unilacunar type of node (Marsden & Steeves 1955). This interpretation has brought Ephedra prominently into recent discussions of the phylogenetic status of the uni- lacunar, two-trace type of vasculature which occurs not only in Ephedra and certain other gymnosperms but which is also characteristic of the cotyledonary and foliage-leaf nodes of a number of dicotyledons (Mars- den & Bailey 1955; Bailey 1956). The considerable attention which has been devoted to the nodal anatomy of Ephedra has not been accompanied by comparably detailed studies on the foliar venation in any of the numerous species of this genus. This overt neglect of venation probably is due in large part to the fact that the very small “reduced’’ leaves, apparently characteristic of the majority of species, have not seemed very promising types of ap- pendages for detailed comparative studies. My survey of the voluminous literature on Ephedra failed to reveal a single drawing or photomicro- graph illustrating the venation pattern as it would appear in cleared, Stained foliage leaves. As far as I could discover, the great majority of previous descriptions of leaf venation in Ephedra were based on recon- structions derived from a study of serial transections of buds or mature nodes. One of the few exceptions which I have encountered occurs in the recent paper by Lehmann-Baerts (1967) who made considerable use of cleared and stained whole mounts in her study of “tracheogenesis” in the paired veins of developing cotyledons and primary leaves of several species of Ephedra. It is also evident that Scheit (1883) employed clearing techniques in his study of venation and transfusion tissue in the leaves of Ephedra altissima and Gnetum gnemon. On the basis of the information provided by serial transections, the great majority of investigators reached the conclusion that the venation of an Ephedra leaf consists of a pair of veins which neither branch nor anasto- mose during their course through the sheath and lamina (Marsden & Steeves 1955, p. 244; Napp Zinn 1966, pp. 173-174). The largely ne- glected investigations of Monoyer (1937; 1938, a, b, c) however, re- vealed for the first time that the leaf trace in certain species of Ephedra may be trifasciculate. In most of the species he studied, only the two lateral members of the trifasciculate trace extend into the leaf while the median trace ends at the nodal girdle as the result of an arrest in its upward development, a process which Monoyer (1938, b, c) termed acrolysis.” But in E. fragilis var. campylopoda, Monoyer (1937) found that each member of a leaf pair consistently has three veins, a median 1972] FOSTER, EPHEDRA 365 “M” and two lateral veins, “L,L.” This unusual type of venation is il- lustrated in his paper by drawings of transections at two different levels through a bud (see Monoyer 1937, plate II, Figs. 1, 2). Monoyer’s stud- ies led him to conclude that the three-veined type of vasculature may represent the “primitive” condition in Ephedra and that the more com- mon two-veined condition is derivative. The occurrence of three-trace leaves, as reported by Monoyer, raises important but as yet unsolved questions concerning nodal evolution and venation in the genus as a whole. In their survey of nodal anatomy in nine species of Ephedra, Marsden & Steeves (1955, Text-figs. 3-5) de- scribed and illustrated, for both decussate and whorled phyllotaxies, the presence of internodal “complementary bundles” comparable in position to Monoyer’s median leaf traces. Marsden & Steeves also noted that the complementary bundles usually extend from one nodal girdle to the next “without in any way joining the leaf traces.” In rare instances (e.g. E. altissima) however, they found that “the complementary bundle ex- tends into the leaf as a third trace.’ In the light of the conflicting interpretations of leaf venation in Ephe- dra, I decided to explore as thoroughly as possible, the venation patterns of a species in which complementary bundles and three-veined leaves oc- cur. A preliminary examination of available cultivated material indicated that E. chilensis Miers would be suitable for study, especially since the nodal and internodal anatomy of this taxon had already been analysed in detail by Marsden & Steeves (1955, Text-fig. 5). The other two species included in the present study, E. foliata Boiss. and E. altissima Desf., were chosen because in both the leaf fluctuates greatly in size, in some cases reaching a length of 1.5 to 2 cm. The conspicuous variation in leaf dimensions allows one to compare the venation patterns of scale-like leaves with those of the well-developed needle-like leaves. MATERIALS AND METHODS Material of the three species of Ephedra included in my study was ob- tained from the following sources: E. foliata (1) shoots from plants grow- ing in the University of California Botanical Garden at Berkeley and in the Botanical Garden, University of Delhi, India, and (2) specimens from the herbarium of the University of California, Berkeley (Barkley 32 Ir 2299; Stewart 14663A); E. altissima, shoots from plant No. 79214 growing at the Plant Introduction Station, Chico, California; E. chilensis, cultivated plant No. 49542 in University of California Botanical Garden (voucher specimens deposited in U. C. Herbarium). It is a great pleasure to express my thanks to Dr. B. M. Johri for the preserved material of E. foliata and to Dr. R. L. Smith for the living specimens of E. altissima. Individual leaves, removed from the nodes by tangential longitudinal cuts, as well as nodal segments with attached leaves, were cleared by prolonged treatment in 5-10% NaOH followed, after thorough washing in water, by chloral hydrate. The leaves and nodal sectors of living ma- 366 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 terial which had been fixed in 50% alcohol or formalin-acetic acid-alco- hol were more refractory than herbarium specimens and it was necessary to accelerate the clearing phase in NaOH by placing the specimens in petri dishes on an electric slide-warmer for long periods of time. Very instructive preparations of E. chilensis, showing the complete venation of each member of a leaf-pair and the basal connections of the respective leaf traces were secured by making careful longitudinal cuts between the leaves and through the node and subjacent internode of cleared segments of the shoot. The separated halves were then dehydrated in the usual alcohol series and stained for several hours in a 1% solu- tion of safranin in equal parts of absolute alcohol-xylene. After destain- ing in alcohol-xylene, the specimens were mounted in “Piccolyte.” Simi- lar procedures were followed in processing and staining detached leaves of all the species included in the present study. Satisfactory photomicrographs of E. chilensis showing clearly the venation pattern and the mode of attachment of the leaf traces proved difficult to obtain by means of transmitted light because of the very numerous fibers adjacent to the veins in the leaf sheath region. The prob- lem was solved by Mr. A. A. Blaker who employed a combination of directional dark field and transmitted light in which the dark field is oriented along the longitudinal axis of the specimen. Under these light- ing conditions, the veins, leaf traces, and internodal vascular strands appear brilliantly refractive and distinct from the fibers and other tissues (see Prates IV, V). I wish to thank Mr, Blaker for his skillful prepara- tion of these illustrations as well as for the other photomicrographs used in this paper. LEAF MORPHOLOGY In the genus Ephedra which comprises about 35 described species (Markgraf 1926), the leaves occur in pairs or whorls at each node and are basally connate, forming a more or less conspicuous sheath. Napp- Zinn (1966, p. 173) uses the term “gamophyll” to designate the fused character of the leaves. Although the classical monographs of Meyer (1849) and Stapf (1889) plainly reveal the considerable fluctuation in the size and proportion of the lamina in Ephedra \eaves, this variability is either minimized or disregarded in many modern descriptions. A few typical examples of misleading generalizations will be given to illustrate is point. Pearson (1929, pp. 1, 2), in his well known monograph on the Gnetales, states that in Ephedra “the assimilating stems are slender and grooved and bear whorls of reduced leaves; they clearly resemble the aerial shoots of Equisetum.” But he adds that in E. altissima, E. fo- liata, and E. gracilis, the leaf blade “is developed only at the middle and the upper nodes.” Another vague description occurs in Chamberlain’s (1935) text on the gymnosperms, in which it is stated that “the leaves are small and rudimentary, and, even when young, are of scarcely any importance in the vegetative economy of the plant.” Zimmermann (1959, 1972] FOSTER, EPHEDRA 367 p. 460) and Sporne (1965, p. 184), respectively, describe the foliage of Ephedra as “decussate scale-leaves” and as “reduced to minute scales.” In contrast to these characterizations, Stapf (1889) noted that although in the majority of species the leaves are reduced to sheaths, a well-devel- oped lamina occurs in some leaves of E. altissima, E. foliata, E. gracilis, and E. americana var. andina, reaching in individual cases a length of 3 cm. and a width of 1 to 1.5 mm. In all these species, the lamina is delicate and ultimately abscisses from the persistent sheath. It seems likely that the persistent sheathing bases of the leaves of herbarium specimens may often be mistaken for complete “scale-leaves.” It will now be appropriate to describe briefly the results of the present investigation on leaf morphology. A well-developed leaf in both E. foliata and E. altissima consists of a basal sheath, joined by a membranous commissure with the other leaf or leaves at the node, and a subulate lamina (TexT-Fic. 3; PLATE I: 1, 3, 4). The total length (i.e. sheath + lamina) of leaves of this type varies considerably; small leaves, with weakly-de- veloped laminae, measure only 8 to 9 mm. in length while the longest leaves studied are 1.5 to 2 cm. long. No examples of leaves 2.5 to 3 cm. long, as found by Stapf (1889) in E. foliata, were encountered in the present study. In both E. altissima and E. foliata, the foliar appendages of the basal nodes of the shoot tend to be relatively small, with laminae measuring only 3 to 4 mm. in length (Trext-Fic. 4). There is obviously however, no sharp demarcation between “well-developed” and scale-like leaves. Very commonly a single shoot during its development may form a graded series of appendages, with the smaller leaves at the base and the longer, better developed leaves restricted to nodes in the middle region. The consistently smaller laminae of the leaves of E. chilensis corre- spond to some of the current descriptions regarding the “reduced” nature of Ephedra leaves. In E. chilensis the phyllotaxis is variable in a single individual: well-developed shoots commonly — although not invariably — form three-leaved whorls whereas the leaf arrangement on the more slender lateral shoots is usually decussate. Regardless of phyllotaxis, the leaf consists of a prominent sheath and an extremely small lamina (TEXxT- FIG. 5; PLATEs IV—VI). In my material, the average length of the leaf is approximately 5 mm., a dimension which falls within the size range of 4.1 to 6.1 mm. noted by Miers (1869, pp. 164, 165) in his description of E. chilensis. VENATION PATTERNS Review of Literature. As I have indicated in the introduction of this Paper, the consensus of opinion in the literature is that the venation of the leaves (including the cotyledons) of Ephedra consists of two un- branched veins which represent the upward extension of a pair of leaf traces (TExt-Fic. 1). In a comprehensive survey of the old as well as the modern literature I have been struck by the vague and extremely brief descriptions of this venation pattern. It would be tedious to review 368 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 chronologically the various statements because of their repetitive charac- ter (see for example, the descriptions given by Thompson 1912, Goebel 1932, Troll 1938, Campbell 1940, Marsden & Steeves 1955, and Sporne 1965). It was particularly surprising to find that in Pearson’s (1929) description of Ephedra, no mention is made of foliage-leaf venation. His account of vasculature in the adult shoot is strictly limited to nodal anatomy and the course of the double leaf traces in the eustele. Many authors, who make a casual reference to venation, describe the venation of an Ephedra leaf as “parallel.” Geyler (1867-68, p. 142) for example, in his classical paper on leaf-trace systems in gymnosperms, stated with reference to E. equisetiformis that the two bundles belonging to the leaf trace “extend together parallel and unbranched in the leaf.” Similar general orientation, however, reference at this point should be made to TEXT-FIGURE 2 which illustrates the three-veined pattern which I have found in certain leaves of E. chilensis. Venation Patterns in Ephedra foliata and E. altissima. The vascu- lature of the leaves of these species consists of two veins which either ter- minate freely near the leaf apex or are connected by transfusion tracheids at variable levels in the lamina (Prates I and II). This fundamental similarity in venation patterns is interesting from a systematic viewpoint because Stapf (1889, p. 35) included both species in the “Tribus Scan- dentes”’ under the section PsEupoBaccaTaE. In his opinion, the ‘Tribus Scandentes” comprises “a very natural group” of species. general literature on Ephedra. On the contrary, the two veins, which may be either widely spaced or closely approximated in the leaf sheath, pro- gressively converge towards the lamina-apex (PLATE I). _ A salient feature of leaf anatomy in Ephedra foliata and E. altissima is the more or less profuse development of “transfusion tracheids.” This term is used to designate the relatively short box-shaped cells, with reticu- 1972] FOSTER, EPHEDRA 369 Ficures 1, 2. Contrasted patterns of open leaf venation in Ephedra. FIGURE 1, two-veined leaf of Ephedra foliata, Ficure 2, three-veined leaf of E. chilen- Sis. The scale near each drawing equals 1 mm. late or reticulate-pitted walls which occur along the borders of the xylem Strands and which in many cases form a compact tissue between the vein tips (PLate II: 5-7). As far as I am aware, Scheit (1883, p. 626) was the first to observe and describe the transfusion tracheids in the leaf of £. altissima. He introduced the term “tracheid-border” (“Tracheid- ensaiime”) to designate the wing-like extensions of tracheids associated with the foliar veins of certain gymnosperms. In E£. altissima, according to Scheit, reticulately thickened tracheids cover the vein endings as a Cap-like mass of cells. The occurrence of transfusion tracheids in this species was also briefly noted by Thompson (1912, p. 1098). But since he relied on leaf transections rather than cleared leaves, he greatly under- estimated the abundance of transfusion tracheids by stating that “at the edges of the bundle, as Strasburger has noted, there is an occasional trans- fusion tracheid which can be distinguished by its bordered pits and thickening bars.” 370 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The general venation pattern of the leaves of E. foliata and E. altissima may be described as either “open” or “anastomosed,’’ depending on the degree of development of transfusion tracheids. If transfusion tracheids are restricted to the borders of the veins, the latter terminate freely at the lamina apex and the venation is ‘‘open.” This condition has been ob- served in well-developed leaves measuring 1 to 1.5 cm. in length as well as in the more rudimentary laminae of scale-like leaves (TEXT-FIc. 1, 3, 4; PuaTE I: 1, Prate II: 5, 6). Anastomosed venation occurs when the ad- jacent borders of transfusion tracheids merge to form a wedge of tissue between the vein-tips (PLATE I: 2-4; Pirate II: 7). There is no sharp demarcation however, between open and anastomosed venation. A num- ber of the leaves studied had definable vein-tips which were joined for only a very short distance by a few intervening transfusion tracheids (Pirate I: 4). Furthermore, it should be emphasized that great variation exists with reference to the level in the lamina at which the two veins first become confluent. In the leaves of cultivated plants of E. foliata, for example, the convergence of the two veins often begins well below the lamina-apex and, in extreme cases, the veins are confluent at such a low level that nearly half the lamina is vascularized by an apparently single massive strand of xylem and transfusion tracheids (PLATE II: 8). Venation Patterns in Ephedra chilensis. The distinctive feature of the foliar vasculature of this species is the interpolation, in certain leaves, of a midvein between the two major lateral veins (PLates IV, V). One of the most puzzling aspects of this type of venation is its sporadic and un- predictable occurrence among the leaves of a single node. TEXT-FIGURE 5 illustrates this point with reference to a three-leaved whorl: one leaf, B, is provided with three freely-terminating veins while each of the other two leaves is vascularized by a pair of veins as shown in A. Comparable variations seem to occur even more frequently in relation to decussate phyllotaxis. Most commonly, both members of a leaf-pair are two-veined but in some of the cleared nodal sectors, one leaf may possess a midvein while the other appendage conforms to the predominant two-veined pat- tern of vasculature (PLATE IV, A, B). Only a single example was found in which both members of a leaf-pair were three-veined. In order to compare the nodal anatomy of two- and three-veined leaves in E. chilensis, selected nodal segments were split longitudinally and the two halves stained in safranin. Preparations of this type clearly reveal the independence between the complementary bundle and the traces which supply a two-veined leaf. As is shown in PxatE III, the complementary bundle terminates abruptly in a mass of transfusion tracheids at the node whereas the two leaf traces are attached to separate parts of the eustele. On this point, my observations fully agree with the reconstruction of nodal and internodal vasculature diagrammatically shown by Marsden & Steeves (1955, Text-fig. 4). The nature of the vascular connections of three veined leaves in E. chilensis, however, was not included in their description and illustrations of the nodal anatomy of this species. During the present 1972] FOSTER, EPHEDRA 371 investigation, a number of successful preparations of divided nodal seg- ments were obtained which show the contrast in nodal anatomy and vena- tion which frequently occurs between members of a leaf-pair. A repre- sentative example of my results is shown in Pate IV. In the three- veined leaf at the left (leaf A), the strong midvein is basally connected with the complementary bundle and the two lateral veins are joined independently to the stele in exactly the same manner as the traces of the two-veined leaf shown at the right (leaf B). Whether the midvein of leaf A represents the “foliar extension” of the complementary bundle or whether the latter simply provides a point of attachment for the midvein, are questions which can only be settled by appropriate ontogenetic studies. An interesting, but unanticipated, result of my studies on E. chilensis was the discovery that in some leaves, the midvein terminates blindly at various levels in the lamina or sheath and hence fails to establish a con- nection with a complementary bundle (Pirate VI). Various types of “transitions” occur between leaves with basally connected midveins and those with “suspended”? midveins. In some leaves, the basal connection between the midvein and the complementary bundle is tenuous and con- sists only of a file of slender tracheary elements (PLaTE V: 11). Ex- amples have also been found of three-veined leaves in which the midvein ends blindly in the nodal region between the points of attachment of the lateral traces (PLATE V: 12). Leaves with these venation patterns clearly seem to provide “connecting links” with the types of venation depicted in Prater VI. Under high magnification, the basal end of a suspended midvein is usually slender and consists of a few elongated, helically- thickened tracheary elements (PLATE VII: 16). This description of venation patterns in Ephedra chilensis may ap- propriately conclude with a few comments on the occurrence of transfusion tracheids. As shown in PLATEs IV and V, transfusion tracheids are abun- dant in both two- and three-veined leaves. Very often, as shown in these illustrations, the vein-tips are “embedded” in a voluminous mass of transfusion tracheids. In the lower portion of the lamina, where the veins are well-separated, each vein is laterally flanked by continuous or inter- rupted files of conspicuous box-shaped transfusion tracheids (Piate VII: 15). Cells of this type may be so abundant in the upper region of the leaf that the boundaries between adjacent veins are difficult or impossible to determine (Pirate IV: A; Piate V: 12). DISCUSSION One of the most important conclusions of the present study is that the venation pattern is variable not only between the leaves of different species but, as in Ephedra chilensis, even between the members of the same whorl or pair of leaves (TEXT-FIc. 5; PLATE IV). This variability has received very little attention in the past (see “Review of Literature,” p. 367). In most descriptions, the venation is said to consist of a pair of 372 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Ficures 3, 4. Variations in size and morphology of leaves in Ephedra foliata (Barkley 32 Ir 2299, uc). Ficure 3, leaves with well-developed laminae and 1972] FOSTER, EPHEDRA 373 free “parallel, ” veins which represent the upward extension of two leaf traces associated with a single gap. It is now clear, however, that the variation in venation pattern in Ephedra includes differences in both the number of veins — two or three — as well as the mode of termination of the veins in the lamina. Ficure 5, A, B. Ephedra chilensis. Variation in venation pattern among leaves he same whorl. One leaf (B) is three-veined while each of a other leaves is two-veined (A). The scale between the drawings equals 1 m In the three species which I have studied, epee venation” i.e. the free termination of the vein- -tips, may occur in “rudimentary” as well as well-developed laminae, and probably represents the basic condition in the genus as a whole (Texr-ric. 1-5: Prare [: 1, Pirate H: 5, 6). On the other hand, when transfusion tracheids, which are restricted to the borders of the xylem strands in open venation, subsequently develop be- tween the veins, anastomoses are formed. As is shown in Pate II, the degree of development of transfusion tracheids varies considerably and determines the fluctuation in the levels at which confluence of the veins takes place. In preparations of Ephedra chilensis, photographed by com- bining dark field illumination and transmitted light, the slender xylem- strands of each vein, in the lower part of the leaf, appear as dark lines weakly-connate sheaths; note open venation of leaf at right. axb, axillary bud, br, lateral branch. Ficure 4, whorl of sealcleayes, ‘detected from shoot and drawn from lower (abaxial) surface. Each leaf is vascularized by a pair veins ane is oe connected, by a a prominent c ommissure, - —— leaves 374 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 flanked by massive borders of refractive transfusion tracheids. At the points of lateral union between the veins, the interpolated files of trans- fusion tracheids are often clearly displayed (PLates IV, V). Quite apart from the anatomical contrasts between open and anasto- mosed venation, there remains the fundamental but difficult question of the significance of the three-veined type of leaf vasculature which oc- curs in Ephedra chilensis. My investigations demonstrate, in the first place, that there is no correlation between phyllotaxis and the number of traces and veins. Pairs or whorls of leaves may be uniformly two-veined, the predominant condition in E. chilensis, or one of the leaves at a node may be three-veined (TExt-Fic. 5; Pirate IV). In my opinion, the rea- son for this apparently erratic fluctuation in vein-number is obscure, al- though Monoyer (1937, 1938, a, b, c) has proposed an ingenious ex- planation for the three-veined pattern in his study of several species of Ephedra. In E. fragilis var. campylopoda, Monoyer (1937) reported that all the leaves are consistently three-veined. The absence of a midvein in the leaves of the other species is the result, according to his view, the leaf. Using the terminology adopted in the present paper, the so- called “complementary bundle,” shown in PLate III, corresponds in posi- tion to Monoyer’s median leaf trace. From a theoretical point of view, Monoyer concluded that the three-veined pattern is “primitive” where- as a leaf with two veins represents a derived condition which results from the “acrolysis” (i.e. the atrophy) of the median trace (Monoyer 1938 b, p. 60). In his own words, “the bundles absent in the leaves are recovered in the stem in the form of acrolysed strands.” (Monoyer 1938 c, p. 64). Although Monoyer’s theory is appealing because of its simplicity, I have found no evidence to support it. If the complementary bundle in the internode represents an incompletely developed median leaf trace, one might expect the occurrence of a “transitional” condition in which the median trace extends for a short distance into the leaf sheath. But I have never observed a venation pattern of this sort. On the contrary, the pre- sumptive “arrest” in the development of the midvein occurs in the leaf itself rather than at the node, and is illustrated by appendages in which a median vascular strand terminates blindly at some level in the lamina or sheath (PLATE V: 12; Pirate VI), The venation pattern depicted in PLATE V: 12, suggests that the formation of tracheary elements in the midvein may occur basipetally and that the patterns shown in Piate VI simply represent various degrees in the extent of such basipetal differ- entiations, This idea, however, because of the absence of ontogenetic data must be considered for the present as entirely speculative. In angiosperms with unilacunar two-trace nodes the midvein of a foliar appendage may be formed by the union of a pair of central strands which arise as branches of the leaf traces (Bailey 1956). This type of transition 1972 | FOSTER, EPHEDRA 375 traces and veins of the leaf (PraTE IV: A). Although the phylogenetic significance of this condition in such a gymnosperm as Ephedra is at present an enigma, it is interesting that in Gnetum Ula, both the scale- leaves and foliage leaves are vascularized by an odd number (5-7) of dis- crete leaf traces (Rodin & Paliwal 1970). Direct comparisons between the nodal anatomy of Gnetum and Ephedra, however, are of questionable value because of the obscure phylogenetic relationship between the two genera (Pearson 1929, p. 143; Eames 1952, p. 97). There is a great need for a more comprehensive study of leaf venation in Ephedra than was possible in the present exploratory survey. T he only known examples of three-veined leaves occur in E. chilensis and E. fragilis var. campylopoda which are assigned by Stapf (1889, p. 35) to “Tribus Antisyphiliticae” and “Tribus Scandentes,” under the Section Psrupopaccatak. If Stapf’s classification is followed, it is evident that the taxonomic and phylogenetic significance of a trifasciculate type of foliar vasculature in Ephedra is an open question pending the results of further comparative studies. Stapf’s scholarly monograph, frequently overlooked by modern investigators, describes in detail the leaf mor- phology and geographical distribution of 31 species of Ephedra and pro- vides the essential basis for future surveys of nodal anatomy and vena- tion in this genus. [Postscript TO ORIGINAL MANUSCRIPT | Professor P. Marten’s (1971) treatise, “Les Gnétophytes,”’ came to my attention after the present paper had been completed. He takes the posi- tion that the morphological reduction of the leaf in Ephedra also in- volves its vascular system. This, he states, consists of two parallel col- lateral endarch veins which neither branch nor reunite in their course through the leaf. Although he notes (p. 21) that Monoyer (1937) had reported the occurrence of three-veined leaves in E. fragilis, he refrains from making any comment on the possible morphological or phylogenetic significance of this “divergent” pattern of venation. LITERATURE CITED Bartey, I. W. 1956. Nodal anatomy in retrospect. Jour. Arnold Arb. 37: 269- BERTRAND, C. E. 1874. Anatomie comparée des tiges et des feuilles chez les Gnetacées et les Coniféres. Ann. Sci. Nat. Bot. Ser. 5. 20: 5-153. CampseLt, D. H. 1940. The evolution of the land plants (Embryophyta). Stanford University Press, Stanford University, California. CHAMBERLAIN . 1935. Gymuosperms, Structure and Evolution. Univer- sity of Chicago Press. Eames, A, J. 1952. Relationships of the Ephedrales. Phytomorphology 2: 79- 100 Grvier, H. T. 1867-68. Ueber den Gefassbiindelverlauf in den Laubblattre- gionen der Coniferen. Jahrb. Wiss. Bot. 6: 55-208. 376 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 GorBEL, K. 1932. Organographie der Pflanzen. Ed. 3. Teil 3. Gustav Fischer, LeHMANN-Barrts, M. 1967. a morphologie du sporophyte dans le genre Ephedra. La Cellule 67: 7-4 MauesHwarl, P., & V. VAsIL. nel Gnetum. Botanical ote No. 1, Council of Sent and Industrial Research, New Delhi, Marxkerar, F., See In Engler & Prantl, Die Snails Pflan- pisy eke g oe ae a3 19. Marspen, M. P. F., & I. W. ts. 1955. A fourth type of nodal anatomy in dicotyledons, illustrated by Clerodendron trichotomum Thunb. Jour. Ar- nold Arb. 36: 1-50. . A. STEEves. 1955. On the primary ito system and nodal anatomy of Ephedra. . Arnold Arb. 36: 241-258. Martens, P. 1971. Les Gnétophytes. Handb. d. Panzeanatomie, Spezieller Teil. Band XII, Teil 2. Gebriider Borntraeger, Ber Meyer, C. A. 1849. Versuch einer Monographie der ae Ephedra, durch Abbildungen Erlaiitet. Mém. Acad. Imp. St. Pétersb. Sér. 6. 5: 225-298. Monoyer, A. 1937. Parcours des faisceaux chez Ephedra fragilis Desf. var. campylopoda (C. A. Meyer) Stapf. Bull. Bot. Liégeois 1: 101-106. Navas a K. 1966. Aueianae des Blattes. I. eae der Gym spermen. Handb. d. Pflanzenanatomie. Band VIII, Teil 1. Gebriider Boas traeger, Berlin-Nikolassee. Pearson, H. H. W. 1929. Gnetales. University Press, Cambridge sor R. J., & G. S. Pattwar. 1970. Nodal anatomy of Gnetum Ula. Phyto- rphology 20: 103-111. Senay M. 1883. Die Tracheidensaiime der Blattbiindel der Coniferen mit vergleichende Ausblicke auf die iibrigen Gefasspflanzen besonders die Cyca- een und Gnetaceen. Jenai. Zeitschr. Naturwiss. 16: 615-636. SPORNE, K. R. 1965. The morphology of gymnosperms. Hutchinson University Library, London. Stapr, O. 1889. Die Arten hee Ephedra. Denkschr. Akad. Wien. Math. -Naturw. Cl. IT. 56: 1 THompson, W. P. 1912. The fami and relationships of the Gnetales. I. The genus Ephedra, Ann. Bot. 26: 1077-1104. TROLL, W. 1938. Vergleichende Morphologie der héheren Pflanzen. Band I. Teil 2. Lieferung 1. Gebriider Borntraeger, Berlin. Vorx, P. 1934, A study of the vegetative phases of Ephedra. Bot. Gaz. 96: 298-313. arasriaones % W. 1959, Die Phylogenie der Pflanzen. Ed. 2. Gustav Fischer, Stuttg; DEPARTMENT OF BOTANY UNIVERSITY OF CALIFORNIA BERKELEY, CALIFORNIA 94720 1972] FOSTER, EPHEDRA 377 EXPLANATION OF PLATES PLATE I Ficures 1-4. Variations in venation patterns of two-veined leaves of Ephedra. Ficure 1. Free termination of veins near apex of leaf of E. foliata (Barkley IG i proximated veins. Ficure 4. E. altissima (cult. USDA Plant Intro. Station, Chico, Calif.) showing confluence of the two veins distal to leaf-apex. All figures, x 11.8. PLATE II FicurEs 5-8. Details of vein structure in aan ; ey foliata. Figure 5. Apex of one member cad pee: 9 ade 4. Note conspicuous transfusion tracheids. Ficu ; ex O ‘eat Cn re PLATE 1: 1, Each vein is bordered by abundant Gansta tracheids. Ficure 7. Ape of leaf shown in Piate I: 2. The vein-tips are united by a central mass of transfusion tracheids. Ficure 8. Apex of leaf shown in Prate I: ve) vein-tips are laterally merged, forming a single massive xylem- strand. Note cluster of prominent hairs developed at apex of lamina. All figures, X 62. PLATE III Ficure 9. Ephedra chilensis. Cleared and stained longitudinal half of a nodal sector Unowie the independent attachment of the two leaf traces (/t), the nodal girdle composed of transfusion tracheids (¢r) and the mode of termination at the node of the complementary bundle (cb). Note strands of fibers between the two veins of the leaf. x 92. PLATE IV Ficure 10, A, B. Ephedra chilensis. A pair of cleared, stained leaves, separated means of a ‘longitudinal cut through the node and pho tographed by sing a combination of dar illumination and transmitted t..0 af (A) is t bundle whereas the other member of the pair (leaf B) is two-veined. Note mas- sive A rhacasvi of transfusion tracheids near the apices of both leaves. x 2 PLATE V Ficures 11, 12. Three-veined leaves of Ephedra — Bri ie Sy with a combination of dark field and transmitted light. Ficu . Lea a show- ing a very attenuated connection between t : fey: ian leaf vein ge the comple- mentary bundle. ae sabi Ma 2 tips me with the tran cg reecheids near the leaf soir FIGURE Leaf ithetratie the blind termination of midvein between the bases of - two lateral traces. Boe auave rivclbeaneat of franitasion. ‘eicheide in upper part of lamina. X 2 PLATE VI Ficures 13, 14. Leaves of Ephedra chilensis with blindly-terminating mid- veins, aod by the combined use of dark field and transmitted light. 378 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 dian vein in the lower part of the leaf sheath. Note transfusion racked e- amina. F ; c Note the refractive nosed e of the stomata in the lower portions of bo Hhnkonichanterhes 26. PLATE VII Ficures 15, 16. Detailed structure of leaf veins of Ephedra — Fic- URE 15. Small sector of lamina of leaf depicted in PLATE IV, A, show glares URE ture of the blind termination of a suspended midvein. Note the slender a cheary elements of the blind on and the abundant transfusion tracheids flanking the lateral veins. x 9 FuaTe I Jour. ARNOLD ArB. VoL. 53 FosTER, EPHEDRA Jour. ARNOLD Ars. VOL. 53 PiaTeE II Foster, EPHEDRA Jour. ArNotp Ars. VoL. 53 PiaTeE III Ke) Foster, EPHEDRA Jour. ARNOLD ArB. VOL. 53 PLATE IV Foster, EPHEDRA Fe Seaton < 5 fe m Pu Q af fs) = if 2] ros) se i. Spee Jour. ARNOLD ARB. PLATE VI a x Aa ey fi = wn iS) oad Jour. ARNOLD ARB. VoL, 53 PiaTE VII Foster, EPHEDRA 386 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 DIOECISM IN CITHAREXYLUM (VERBENACEAE) P. B. ToMLINSON AND PRISCILLA FAWCETT OUTBREEDING MECHANISMS ARE A feature of perennial plants (Steb- bins, 1950 p. 165 et seg.) and their high incidence in trees is well known. This high incidence is related to the evolutionary advantages of outcros- sing in terms of genetic flexibility. In continuing studies of the distinc- tive woody flora of South Florida it is evident that the majority of species have some simple and obvious mechanism for outcrossing such as dioecism, self-incompatibility as suggested by heterostyly, and other kinds of morphological polymorphism. Field study is revealing further examples among species which have hitherto been thought to have perfect flowers. The present article refers to such an example in Citharexylum frutico- sum L. (“Fiddlewood”). Citharexylum L. is a genus of some 100 species in Central and South America. The flowers seem largely to have been described as “‘perfect” (e.g. Sargent, 1894). Citharexylum fruticosum it- self is widely distributed in the Caribbean area and a number of varieties have been described (Moldenke, 1959). In South Florida it is scattered in pinelands and hammocks on the Florida Keys and northward with a limited coastal distribution. The species tends to be a pioneer in recently burnt pineland and persists in more or less closed hammock communities which are the climax communities in many fire-free areas. The species shows distinct weedy tendencies and is not uncommon on disturbed sites. From this ease of spread and existence in natural communities it is an evident member of the native flora. Descriptions of this tree in floras and other manuals dealing with this part of the United States (e.g. Chapman (1897), Sargent (1894), Small (1933), West and Arnold (1946)) or those which refer to it in other parts of its range (e.g. Britton and Millspaugh (1920), Léon and Alain (1957), Little and Wadsworth (1964), Moldenke (1954)) suggest, directly or indirectly, that flowers are perfect. In South Florida, at least, it is dioecious. This observation has been overlooked because the flowers in the two sexes are very similar and casual examination of herbarium speci- mens would not reveal the constant differences shown in field populations. OBSERVATIONS Phenology and Sexuality. Citharexylum begins to flower in early summer in South Florida (May—June). The flowers are on slender pen- dulous spikes which are either terminal on or from the upper leaf axils of current-year shoots (Ficure 1). During the summer months several generations of renewal shoots with flowering branches may be produced so that the flowering period is extended. Flowers, however, do not occur 1972] TOMLINSON & FAWCETT, CITHAREXYLUM 387 agram 10, m 1 10, in longitudinal section; 11, detail of abaxial staminode and two adjacent sta- mens in a male flower. Figs. 1 & 2, * 1/2; figs. 3-5 & 7, 9 & 10, X 3; fig. 11, mS “throughout the year” as has been stated in several publications, all of which seem to have their origin in the description by Sargent (1894). The phenology and general growth pattern of this tree is very distinctive and requires detailed study, a statement which could be repeated for all the woody plants of South Florida. 388 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Towards the end of summer, and subsequently when fruit set begins, the difference between male and female trees becomes obvious. Fruits enlarge on female trees; male trees remain barren and can be distinguished throughout the fruiting season (September—December) by the pendulous but empty inflorescences. This distinction is evident even earlier, since fertilized flowers on female spikes lose only their corollas but retain the calyx and developing pistil (F1cuRE 1, below). Male flowers fall as a unit after anthesis so the base of the inflorescence axis becomes bare (FIGURE 2) Trees identified as male and female and tagged in one year retain their same sexuality in the following year. Flowers. Superficially, flowers in the two sexes are identical (FIGURES 3-10). Each flower has a tubular, 5-toothed calyx which usually splits down one side. The corolla is tubular with 5(—6) more or less equal petal-lobes although the members of the adaxial pair are often somewhat larger than the three abaxial lobes. This tendency to zygomorphy is most pronounced in the stamens. There are 5 stamens or staminodes in each kind of flower but the adaxial one (indicated by the arrow in F1cuRES 6 and 8) is usually distinctly different from the other 4; it is almost always smaller and it may be so reduced as to be represented by a mere vestige (FicurE 11), In the male flower the 4 abaxial stamens are always fertile; dehiscence is introrse and the large pollen grains adhere into a relatively conspicuous mass. The adaxial stamen may contain pollen, even though it is reduced, but often it is vestigial, as in Ficure 11. In the female flower the stamens are always sterile and the microsporangial tissue re- mains brown and never produces pollen. Such staminodes never dehisce. The ovary appears to be equally well developed in both sexes (FIGURES 4 and 6; 8 and 10) with 4 equal ovules, one in each pseudoloculus. The stigma is more or less bilobed but the length of the style is somewhat variable, especially in the male flower. Usually the stigma is at about the same height as the stamens. There is no evidence that the ovules in the male flowers are functional since we have never seen fruits on male trees. The reason for this abortion of seemingly well-formed ovules is worthy of more detailed investigation. Pollination. Flowers are sweet-scented during the day and nectar is secreted at the base of the ovary tube. Since the pollen is sticky it is evident that the flowers are insect pollinated but we have made no ob- servations on the kinds of insect visitors. The mouth of the corolla tube is more or less occluded by a dense outgrowth of uniseriate multicellular hairs (Ficures 3 and 7) and an insect probing for nectar would certainly touch both stigma and pollen mass. Self-fertilization of male flowers would seem to be quite easy. Fruit set on female trees is very abundant but there seems to be a high incidence of fruits with aborted embryos. 1972] TOMLINSON & FAWCETT, CITHAREXYLUM 389 CONCLUSION It seems evident from these morphological observations that Citharexy- lum fruticosum is dioecious and it may seem remarkable that such a simple observation could have escaped the attention of botanists. However, this is a general reflection on how poorly the tree flora of South Florida has been investigated by field workers. LITERATURE CITED Britton, N. L., & C. F. MitispaucH. 1920. The Bahama Flora. Published by the authors. New York. CuHapman, A. W. 1897. Flora of the southern United States. Ed. 3. Cambridge Botanical Supply Co., Cambridge, Massachusetts. Leon, H., & H. ALAIN. 1957. Flora de Cuba. Vol. 4. P. Fernandez, Havana. Lirt.e, E., & F. H. WapswortH. 1964. Common trees of Puerto Rico and the Virgin Islands. U.S.D.A. Handbook No. 249. Forest Service, Washington, BU. MOoLpDENKE, H. N. 1954. Verbenaceae in: Flora of Trinidad and Tobago. vol. II, pt. VI. Government Printing Office, Trinidad . 1959. A résumé of the Verbenaceae, Avicenniaceae, Stilbaceae, Sym- phoremaceae and Eriocaulaceae of the world as to valid taxa, geographical distribution and synonymy. Published privately. SARGENT, C. S. 1894. Citharexylon in: The Silva a North America. vol. VI. 101-103. Pl. CCXCV. Houghton Mifflin & Co. Boston & New York. SMALL, J. K. 1933. Manual of the southeastern Flora. Published by the author. New York. Stepsins, G. L. 1950. Variation and evolution in plants. Columbia University West, E., & L. E, ARNOLD. 1946. The native trees of Florida. University of Florida Press. Gainesville. (Poe. T.) (P. F.) HARVARD UNIVERSITY FAIRCHILD TROPICAL GARDEN HARVARD FOREST 10901 OLtp CuTLER Roap PETERSHAM, Mass. 01366 Mr1aMI, FLORIDA 33156 390 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 NOTES ON WEST INDIAN ORCHIDS, II * LESLIE A. GARAY AND HERMAN R. SWEET During the preparation of the orchid manuscript for the Flora of the Lesser Antilles by Richard A. Howard of Harvard University, we have noted several nomenclatorial changes and have found a Dichaea species yet undescribed. Since the actual floristic presentation is not designed to incorporate new changes or descriptions of new species, our findings are presented here. Elleanthus cephalotus Garay & Sweet, nom. nov. BastonyM: Bletia capitata R. Br. in Ait. Hort. Kew. ed. 2. 5: 206. 1813. SyN.: Elleanthus capitatus (R. Br.) Rchb.f. ex Cogn. in Urb. Symb. Antill. 6: 561. 1910; Stehlé, Fl. Descr. Antill. Fr. 1: 176. 1939. Type: West Indies, without proper locality. Introduced by Sir Joseph Banks in 1795 and flow- ered at Kew Gardens (pm). Evelyna capitata Poepp. & Endl. Nov. Gen. ac Sp.Pl. 1: 32. 1836. Elleanthus capitatus (Poepp. & Endl.) Rchb.f. in Walp. Ann. Bot. Syst. 6: 475. 1862; Schultes in Am. Orchid Soc. Bull. 25: 387. 1956. Type: Peru, near Pampayaco. Poeppig s.n.! (w). It is most unfortunate that the well-known descriptive name of Ellean- thus capitatus has to be changed. This species has been described twice; first as Bletia capitata R. Br. and later as Evelyna capitata Poepp. & Endl. Recognizing that Elleanthus is the correct generic name, Reichenbach transferred Evelyna capitata to that genus. Cogniaux having found that Bletia capitata and Evelyna capitata were one and the same, because of the same specific epithet, he wrongly combined Robert Brown’s epithet with Reichenbach’s transfer. Because the specific epithet ‘capitata’ is preoccupied in Elleanthus, the transfer of Bletia capitata necessitates the new name proposed above. Eltroplectris calcarata (Sw.) Garay & Sweet, comb. nov. BastonyM: Neottia calcarata Sw. Fl. Ind. Occ. 3: 1413. ¢.28. 1804. Type: Dominican Republic, without proper locality. Swartz s.n. (s). Syn.: Pelexia setacea Lindl. Gen. and Sp. Orch. Pl. 482. 1840; Cogn. in Urb. Symb. Antill. 6: 327. 1909, Centrogenium setaceum (Lindl.) Schltr. in Beih. Bot. Centralbl. 37(2): 453. 1920; Schultes in Am. Orchid Soc. Bull. 25: 387. 1956. Typr: Brazil, Rio oce. Prince Maximilian Wiedens s.n.! (M, w). We have examined a large series of specimens from the entire area of distribution, and have found that the margin of the lip may vary from * Continued from Jour. Arnold Arb. 50: 462-468. 1969. 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, II 391 entire to erose-denticulate or to almost fimbriate. Hence, Neottia calcara- ta i and Pelexia setacea merely represent the two extremes in a single variation pattern. Pleurothallis ophioglossoides (Jacq.) Garay & Sweet, comb. nov. Bastonym: Epidendrum ophioglossoides Jacq. Enum. Syst. Pl. 29. 1760, based on Helleborine ophioglosso similis, Plum. Cat. 9. 1703, and Icon. Burm. t. 176, f.3. 1759, not Epidendrum ophioglossoides Jacq. Sel. Stirp. Am. Hist. 225. 1763. Type: Martinique, without proper locality. Plumier s.n. (P). Holotype is Plumier’s original drawing. Syn.: Specklinia floribunda Lindl. Gen. and Sp. Orch. Pl. 9. 1830. Pleuro- thallis floribunda (Lindl.) Lindl. in Bot. Reg. 28: misc. p. 73. 1842, not Poepp. & Endl. 1836; Rolfe in Kew Bull. 277. 1893; Cogn. in Urb, Symb. Antill. 6: 421. 1909; Stehlé, Fl. Descr. Antill. Fr. 1: 205. 1939; Hodge in Lloydia 17: 213. 1954. Type: Martinique, without proper locality. Sieber 206! (kK, w When Jacquin described Epidendrum ophioglossoides in 1760, he based his description on Plumier’s polynomial of 1703 as well as on Plumier’s drawing published subsequently by Burmann in 1759. Three years later, Jacquin published a new description of Epidendrum ophioglossoides with- out any reference to his 1760 paper, which he prepared from one of his own collections and which he assumed to be identical with Plumier’s plates. Fortunately Jacquin’s collection of that particular plant is in the British Museum. ave examined this specimen and it represents Stelis scabrida Lindl. A closer examination of Plumier’s drawing imme- diately brings into focus the error in Jacquin’s judgment, for Plumier’s drawing is a true Pleurothallis, generally known as P. floribunda (Lindl.) Lindl. Swartz in his Observationes Botanicae 332. 1791, already called atten- tion to Jacquin’s misidentification: “Plantae Plumieri & Jacquini certe distinctae. . . . Jacquini vero ad E. trigoniflorum (prodr. p. 125.) per- tinet, affirmante planta Ipsius in Museo Banksiano serata.” Notwith- standing his own observations, when he established the genus Stelis in 1800, Swartz followed Jacquin in proposing the new combination, Stelis ophioglossoides with his own E. trigoniflorum in synonymy. This mistake has been carried through to today for Epidendrum ophioglossoides be- came the conserved type of the genus Stelis. To prevent the transfer of over 1,000 species of Pleurothallis into the genus Stelis, the retypification of the genus Stelis will be published in a separate paper. Octomeria apiculata (Lindl.) Garay & Sweet, comb. nov. Basronym: Pleurothallis apiculata Lindl. Folia Orch: Pleurothallis 17, no. 88. 1859, Type: Venezuela, without proper locality. Fendler 2155! (K-L). An examination of the type specimen of Pleurothallis apiculata has shown it to possess eight pollinia, one of the chief characters of the genus 392 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Octomeria. A new and emended description of this species is given here for convenience. Epiphyte. Primary stem or rhizome scandent, often covered with fi- brous remnants of sheaths. Secondary stems rather slender, obliquely ascending, enclosed by 3 to 4 imbricating tubular scarious sheaths, up to 4.5 cm. long, proportionately (1:1) as long as or somewhat shorter than terminal leaf. Leaves coriaceous, narrowly ovate-elliptic to oblong- elliptic, acute, basally contracted into a short, up to 4 mm. long, petiole, including the petiole up to 6.5 cm. long and 7 mm. wide. Inflorescence fasciculate, 1- to 2-flowered. Flowers rather thin in texture, yellow, fra- grant. Sepals similar, ovate to ovate-lanceolate, acute, 3-5 mm. long an 2 mm. wide. Petals ovate to rhombic-ovate, acute or acuminate, occasional- ly somewhat contracted below the tip, hence appearing apiculate, up to 4.5 mm. long and 1.5 mm. wide. Lip fleshy with distinct claw, 3-lobed; lateral lobes erect, more or less angular, midlobe ovate to oblong-ovate, obtuse or more or less 3-dentate at apex, margins erose-denticulate, disc provided with a pair of obliquely inserted fleshy lamellae at junction of midlobe and lateral lobes; whole lip up to 3 mm. long and 2 mm. wide. Column somewhat arcuate, up to 2 mm. long. Pedicellate ovary up to 4 mm. long. General Distribution: Nicaragua, Costa Rica, Venezuela, Trinidad, To- bago, Grenada, and Guadeloupe. Epidendrum latifolium (Lindl.) Garay & Sweet, stat. nov. FIcurE 1. BastonyM: Epidendrum nocturnum var. latifolium Lindl. in Bot. Reg. 23: t. 1961. 1837; Cogn. in Urb. Symb. Antill. 6: 523. 1910; Stehlé, Fl. Descr. Antill. Fr. 1: 144. 1939. Type: West Indies. Introduced and cultivated by Paxton s.n.! (K-L) The strongly ancipitous stem, large elliptic leaves, sessile, 1-flowered inflorescence, and the flowers twice as large show no intergradation with Epidendrum nocturnum to which E. latifolium was formerly attached as a variety. Epidendrum latifolium is known from the Greater and Lesser Antilles, Trinidad, and Venezuela. Epidendrum secundum Jacq. Enum. Syst. Pl. 29. 1760, based on Hel- leborine purpurea umbellata Plum. Cat. 9. 1703 and Icon. Burm. 179. t. 184, f.1. 1759, not Jacq. Select. Stirp. Am. Hist. 224. ¢. 137. 1763. Syn.: Epidendrum fuscatum Smith, Spicileg. Bot. fasc. 2: 21. ¢. 23. 1792, nom, superf. Type: Martinique, without proper Seicaliiy: Plumier s.n. (2). Holotype is Plumier’s original drawin Epidendrum anceps Jacq. Select. Stirp. Am. Hist. 224. ¢. 138. 1763; en Stirp. Am. Hist., ed. icon. arch. pict. 108. ¢. 208. 1781; Boldingh, F Eustat., Saba and St. Martin 41. 1909; Cogn. in Urb. Sym per Be - 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, II 393 Ficure 1, Epidendrum latifolium (Lindl.) Garay & Sweet. Drawn by Dorothy Allen. 394 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 516. 1910; Stehlé, Fl. Descr. Antill. Fr. 1: 134. 1939; Hodge in Lloydia 12: 205. 1954; Schultes in Am. Orch. Soc. Bull. 25: 387. 1956; Goodring, Love- less & Proctor, Fl. Barbados ae Sune Tyre: Martinique, without proper locality. Jacquin s.n. No specimen is known to us. Holotype is Jacquin’s original illustration of which reproduction are cited above. Epidendrum. viridipurpureum Hook. in Bot. Mag. 65: t. 3666. 1838 Epidendrum anceps var. oo ae (Hook.) Lindl. Folia Orch. Epi- dendrum 66. 1853; Cogn. in Urb. Symb. Antill. 6: 518. 1910. Type: Ja- maica, without proper locality. Imported and cultivated by Horsfall s.n.! (xk). It should be noted that Plumier’s drawing upon which Epidendrum se- cundum is based and the description and drawing of Epidendrum anceps are identical in every respect. The fact that Jacquin described the same species twice within three years stems from his repeated misidentification of his own material when he superficially compared it with Plumier’s drawings. Pinelia Leochilus (Rchb.f.) Garay & Sweet, comb. nov. BastonyM: Epidendrum Leochilus Rchb.f. in Flora 48: 277. 1865. SYN.: Pleurothallis she Bia Wright ex Griseb. Cat. Pl. Cub. 262. 1866, in syn. Type: Cuba, near Monte Verde. Flowers ees tee green or white, a red spot on the anther. Wright 3346! (amEs This exceedingly rare species has been collected recently for the first time in the Dominican Republic: on banks of Mao River, El Aguacate, La Leonor. Liogier 13210! (Ny). Through this new collection we were able to determine the true generic status of this tiny plant. Oncidium altissimum (Jacq.) Sw. in Vet.-akad. Stockholm Nya Handl. 21: 240. 1800, not of other authors BAsIONYM: Epidendrum altissimum aca: Enum. Syst. Pl. 30. 1760, based on Viscum radice bulbosa majus & elatius, delphinii flore ferrugineo guttato, Sloane, Hist. Jamaica 1: 250. ¢. 148, f.1. 1707, not Jacq. Select. Stirp. Am. Hist. 229. ¢. 141. 1763; ed. icon. arch. pict. $10. 2, 215. 1781.. TyPe:. Ja- maica, without proper locality. Sloane s.n.! (BM). Syn.: Epidendrum maculatum Aubl. Hist. Pl. Guiane Fr. 2: 825. 1775, based on Helleborine maculosa, foliis aloés carinatis Plum. Cat. fe 1703 and Icon. Burm. ¢. 178, f.2. 1759, not Oncidium maculatum Lindl. I, priruoes crispum Lam. Encycl. 1: 186. 1783, not ane i: crispum Lodd. poerie maculatum (Aubl.) Urb. in Sahge Rep. 15: 306. 1918, not Onci- dium maculatum Lindl. 1841. Type: St. Vincent, oy proper locality. Plumier s.n. (P). Holotype is Pade, original dra Oncidium luridum Lindl. in Bot. Rees ob 7a. 1823: "Rolfe in Kew Bull. 278. 1893; Duss in Ann. Inst. Col. Marseille 3: 600. 1897; Hodge in Lloydia 17: 221. 1954; eka in Am. Orchid Soc. Bull. 25: 391. 1956. Oncidium guttatum Rchb.f. in Walp. Ann. Bot. Syst. 6: 782. 1863, excl. syn.; Cogn. in Urb. Symb. Antill. 6: 647. 1910; Kraenzlin in Pflanzenr. IV. 50. Heft 80: 33. 1922; Stehlé, Fl. Descr. Antill. Fr. 1: 248. 1939. Type: In- troduced from South America and cultivated by Griffin s.n.! (K-1). 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, II 395 When Jacquin described Epidendrum altissimum in 1760, his descrip- tion was based on a polynomial and an illustration in Sloane’s Hist. Ja- maica 1: ¢. 148, f.1. Three years later Jacquin redescribed and illustrated an entirely different plant as Epidendrum altissimum which he based on one of his own collections. We have examined the Sloane material in the British Museum and it represents the plant commonly known as Oncidium luridum Lindl., a plant very different from Jacquin’s picture. Hence, the binomial Oncidium altissimum (Jacq.) Sw. is the correct name for this widespread plant. Although this same conclusion was reached by Faw- cett and Rendle in 1910, they did not effect the necessary change. On the other hand, the correct name for those plants which were named with the aid of Jacquin’s plate of Epidendrum altissimum of 1763, is On- cidium W ydleri Rchb.f. Dichaea Hookeri Garay & Sweet, sp. nov. FIGURE 2. Type: Dominica, moist forests bordering Pegoua river in the vicinity of Deux Branches, Concorde Valley. Hodge 3481! (AMES). Epiphytica, caespitosa; caulibus erectis vel pendentibus, compressis, fo- liatis, foliis patentibus, oblongo-ligulatis, acutis, mucronulatis, margine sub lente papilloso-ciliolatis; inflorescentiis unifioris: sepalo postico ovato- lanceolato, acuminato; sepalis lateralibus Sbliquis: petalis lanceolato-el- lipticis, acuminatis; labello e cuneata basi anchoriformi, lobis acuminatis, retrorsis, antice rotundato cum apiculo minuto; ovario glabro. Epiphytic, caespitose herbs up to 15 cm. tall. Stems erect to pendent, compressed, completely enclosed by imbricating leaf-bases. Leaves de- ciduous, length of leaves very uneven on stem, articulate, oblong-ligulate to linear-oblong, acute, indistinctly mucronate, the upper edge under lens finely papillose-ciliolate, up to 2.5 cm. long and 6 mm. wide. Inflorescence axillary, 1-flowered. Flower small, ringent, greenish to greenish yellow. Dorsal sepal ovate-lanceolate, acuminate, up to 5 mm. long and 2.5 mm. wide. Lateral sepals oblique, ovate-lanceolate, acuminate, up to 6 mm. long and 3 mm. wide. Petals lanceolate-elliptic, acuminate, up to 4.5 mm. long and 2.5 mm. wide. Lip from a cuneate base anchor-shaped with retrorse acuminate lobes, rounded in front with a small inconspicu- ous apicule in middle, up to 5 mm. long and 5 mm. wide when spread. Column cylindric. Ovary sessile, glabrous. Flowers from March to July. DistrIBuTION. Guadeloupe: Mateliane, Stehlé 456! Bigs aug Questal 4735! (Pp); without proper locality, —— s.n.! (Pp); H s de Goyave, Stehlé 961! (Ames, NY). Dominica: La Chaudiere, heant® river, Hodge 3509! (AMES). St. Vincent: without sows locality, Guilding s.n.! (K). Trini- d: Aripo, Broadway 2330! (ames); Mora forest, R. O. Williams s.n.! ( AMES). Island Margarita: without proper ee J. R. Johnston 242! (AMES). Venezue- la: Guatopo, Dunsterville 339! (AM Specimens referable to this species have been erroneously reported in the literature as Dichaea Cogniauxiana Schltr., D. graminoides (Sw.) 396 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Ficure 2. Dichaea Hookeri Garay & Sweet. Drawn by G. C. K. Dunster- ville. Lindl., D. picta Rchb.f., and D. Rendlei Gleason. The first illustration of D. Hookeri appeared in Hooker, Exotic Flora 3: t. 196. 1826, as Isochilus graminoides (Sw.) Lindl. Hooker’s plate was based on the above cited collection from St. Vincent (not Trinidad as stated) and it is so marked on the herbarium sheet. Gleason’s contention that Hooker’s plate repre- sented D. Rendlei is incorrect since a comparison of the original material from which the plate was prepared with that of the type of D. Rendlei has shown them to be distinct. 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, II 397 EPIDENDRUM PATENS SW. EPIDENDRUM MUTELIANUM COGN. EPIDENDRUM PALLIDIFLORUM HOOK. Ficure 3. Details of flowers of Epidendrum patens Sw., E. "obggeaae Cogn., and £, Senay eas Hook. Drawn by Magdalena Pena de Sou Dichaea Swartzii (C. Schweinf.) Garay & Sweet, stat. nov. Basionym: Dichaea eee var. Swartzii C. Schweinf. in Bot. Mus. Leafl. arvard Univ. 17: 1955. Syn.: Epidendrum Acie Sw. Nov. Gen. & Sp. Pl. Prodr. 124. 1788, nom. illeg. excl. syn. 398 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Dichaea echinocarpa ag Lindl. Gen. and Sp. Orch. Pl. 208. 1833, nom. illeg. excl. syn.; Duss in Ann. Inst. Col. Marseille 3: 596. 1897. "Type: Jamaica, without Sadie beality. Swartz s.n.! (BM, s The original contention of Swartz that Epidendrum echinocarpon Sw. and Epidendrum pendulum Aubl. were one and the same species has been seriously questioned by Schweinfurth. In examining all of the available evidence we have found Schweinfurth’s insight so justified that we rec- ognize the West Indian material as a species separate from D. pendula (Aubl.) Cogn. from Guyana. THE EPIDENDRUM PATENS COMPLEX The three species, Epidendrum patens Sw., E. Mutelianum Cogn., and EB. pallidiflorum Hook. have been thoroughly confused with one another in the literature. With the exception of the original diagnosis, all subsequent descriptions are mixtures and none may be used with certainty for iden- tifications. The original confusion was started by Lindley who misinterpreted Swartz’s diagnosis of E. patens and applied it to one of the South Ameri- can species of Epidendrum, namely E. coronatum Ruiz & Pav. Cogniaux in preparing the orchid treatment for Urban’s Symbolae An- tillanae accepted Lindley’s interpretation of E. patens, renamed the later homonym of E. flavum Mutel to E. Mutelianum Cogn. and included in this latter species all material now referable to the true EZ. patens Sw. Stehlé in his treatment, Les Orchidales, Flore descriptive des Antilles Frangaises, described and illustrated E. Mutelianum under the name of E. pallidiflorum and E. patens under E. Mutelianum. We were fortunate to locate all three type collections. In order to clarify the distinctions among the three species, we are presenting an out- line of their floral parts as a visual supplement to the following key: 1. Peduncle short, terete, up to 5 cm. long, surrounded by 1 to 4 lanceolate, basally imbricate, laterally compressed spathes; disc of lip plurituberculate DA, ce ee ee Mutelianum. 1. Peduncle eo ancipitous, at least 10 cm. long, basally naked; disc 0 lip with a 3- to 5-dactylate plate at base................ 2. 2. Sepals pea petals broadly ovate-rhombic, much wider than sepals; lip 3-lobed; midlobe linear-oblong with prominent diverging apical 1lo- beled: disc with a median keel. E. patens. bo Sepals lanceolate to linear-oblanceolate; petals linear-oblanceolate, much narrower than sepals; lip 3-lobed; midlobe or without apical lobules; disc without a median ke el. Pe ino aes E. pallidiflorum. Orcuip HERBARIUM OF OAKES AMES BoTANICAL MusEuUM A ITY CAMBRIDGE, MASSACHUSETTS 02138 1972] HOWARD, TIBOUCHINA AND CHARIANTHUS 399 NOTES ON TIBOUCHINA AND CHARIANTHUS (MELASTOMATACEAE) IN THE LESSER ANTILLES RicHArp A. HowARD Five species of Tibouchina are commonly recognized in the Lesser Antilles, but four of them are variously cited as to specific epithet and authors. The following notes are an attempt to clarify the nomenclature of the four species. The current typification of the genus Charianthus D. Don is in error, and a new combination is proposed for a second species. Tibouchina ornata (Sw.) Baillon, Adansonia 12: 74. 1876. Melastoma ornata Sw. Prodr. 69. 1788 (‘“Jamaica,” no specimens cited). Rhexia strigosa Rich. Act. Soc. Hist. Nat. Paris 1: 108. 1792. Rhexia inconstans Vahl, Eclog. 1: 37. 1796 (nom. illeg.). Melastoma ciliata Desr. in Lam. Encycl. 4: 49. 1797 (Badier, Guadeloupe). Osbeckia ornata (Sw.) Sw. Fl. Ind. Occ. 2: 647. 1798 (Masson, St. Kitts; du Ponthieu, Guadeloupe. Herb. Banks). Rhexia ornata Rich. in Humboldt & Bonpland, Rhexies 69. ¢. 26. 1823. Hephestionia strigosa Naud. Ann. Sci. Nat. ser. 3. 13: 36. 1849. Chaetogastra strigosa DC. Prodr. 3: 134. 1828. Pleroma strigosum Triana, Trans. Linn. Soc. -; 47.1871. Pleroma ornatum Triana, ibid. t. 3, f. 31 p. 1 This species is generally known as Tibouchina strigosa, and the au- thority for the name has been cited in a great variety of ways. The plant involved was first described by Olaf Swartz in his Prodromus, and cited as from Jamaica, with no collector indicated. In a subsequent treatment in the Flora Indiae Occidentalis, Swartz referred to the Prodomus treat- ment, did not mention Jamaica, and cited collections by Masson on St. Kitts, and by du Ponthieu on Guadeloupe. Such specimens are preserved in the Banks Herbarium at the British Museum (Natural History), and several collections are mounted on one sheet and are numbered. The specimen labeled #1 is selected as the lectotype, being that of du Pon- thieu from Guadeloupe in flowering condition. The Swartz epithet ornata has been cited in synonymy by many authors, but has not been adopted for the species as is necessary. Baillon did make the combination Tibouchina ornata, but used the basionym Plero- ma ornatum Triana. Triana had published the name Pleroma strigosum in the text of his treatment on the Melastomataceae, and cited in synonymy “Osbeckia ornata Tuz. Fl. Ind. Occ. 647,” which is apparently a typo- graphical error for Swartz, Fl. Ind. Occ. 647. In the explanation of the plates in the same work on page 162, Triana refers to Pleroma ornatum Tri. One can assume he was correcting the name used earlier in the text, and was using the Swartz name as a basionym. 400 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The correct citation of name and authority is Tibouchina ornata (Sw.) Baillon. This species has been collected on Montserrat, St. Kitts, Guadeloupe, and Dominica, within the Lesser Antilles. Cogniaux (DC. Monogr. Phan. 7: 267. 1891) also cites under Tibouchina strigosa collections from St. Vincent by Anderson and Masson, and from “Guiana anglica ad Demerara” by Parker. Gleason (Brittonia 1: 156. 1932) has already suggested the Parker specimen might be of a cultivated plant and not native to the Guianas. Alexander Anderson was the director of the botanic garden on St. Vincent (1783-1811) and might have had the species under culti- vation there. Recent explorations of the Soufriére on St. Vincent, the logical location for the natural occurrence of the species, have failed to produce “7. strigosa,” but quantities of Tibouchina cistoides instead. The vegetation of the summit area of the Soufriére was destroyed in the eruption of the volcano in 1902. It is possible that the species might have been eliminated in that eruption except for the fact that it has not been reported from Martinique, where perhaps the same situation pre- vails on the slopes of Mt. Pelée. Masson is known to have collected on St. Kitts where the species occurs today, but not on St. Vincent. Tibouchina chamaecistus (Naud.) Cogn., DC. Monogr. Phan. 7: 267. Hephestionia chamaecistus Naud. Ann. Sci. Nat. ser. 3. 13: 37. 1849. Chaetogastra chamaecistus Griseb. Fl. Brit. W. I. Isl. 267. 1860. The basionym for this combination is usually attributed to Grisebach, who cited the Naudin work at the generic level only. However, Naudin has used the specific epithet in his publication, and it is presumed that Grisebach implied a transfer and a new combination. The type speci- men was collected by Mme. Rivoire on the slopes of Mt. Pelée in Mar- tinique, the only known location for this species. Tibouchina chironioides (Griseb.) Cogn., DC. Monogr. Phan. 7: 267. 1891. Chaetogastra chironioides Griseb. Fl. Brit. W. I. Isl. 267. 1860. however, by also citing Chaetogastra chironioides Griseb. “part” in the synonymy of his Tibouchina strigosa. Karl Domin has already indicated (Acta Bot. Bohemica 9: 35. 1930) that Grisebach cited but a single col- lector, Imray, and although all of the Imray collections I have seen are clearly referable to this species, a specimen in the Gray Herbarium bears an old annotation in an unknown hand as “Chaetogastra strigosa.” Depauperate plants of Tibouchina ornata collected on Dominica led some scholars to associate T. ornata and T. chironioides as a single species, 1972] HOWARD, TIBOUCHINA AND CHARIANTHUS 401 judging from the annotations. The leaves of Tibouchina ornata are tri- nerved and ovate. The Imray collection, which is the type of T. chi- ronioides, and such subsequent collections as Eggers 114,W. 2G BT. Hodge 1860, Wilbur et al. 8244 all have narrowly lanceolate leaves 6—7 mm. long and rarely exceeding 1 mm. in width. Tibouchina chironioides appears to be endemic to Dominica at altitudes of 1900 to 2100 feet. Tibouchina longifolia (Vahl) Baill. ex Cogn., DC. Monogr. Phan. 7: 264. 1891. Rhexia longifolia Vahl, Eclog. 1: 39. 1796. The author of this name is generally considered to be Baillon in Adan- sonia 12: 74. 1876, although other references and authors have been cited. Baillon did not make the combination, but only implied it, when he noted that Pleroma of Don should be included within a variable Tibouchina. For other species, such as Tibouchina ornata (Sw.) Baillon, he did cite a full basionym. Pleroma longifolia Triana (Trans. Linn. Soc. 28: 45. t. 3, f. 31 n. 1871) is based on Rhexia longifolia Vahl. The correct basionym appears to be cited first by Cogniaux as indicated, although he attributed it to Baillon. For accuracy the authority should be cited as Baillon ex Cogniaux. This species is reported to occur from Bolivia to Venezuela, in Central America, the Lesser Antilles, and the Greater Antilles. The type is a col- lection of von Rohr from “America meridionali,” probably the Guianas. The genus Charianthus was established by David Don (Mem. Wer- nerian Soc. 4: 327. 1823), at which time he cited three species, C. coccineus (L. C. Rich.) D. Don, C. purpureus D. Don, and C. tinifolius D. Don. In a monograph of the genus, W. H. Hodge (Contrib. Gray Herb. 135: 115-— 133. 1941) indicated the ‘“‘type species” to be Charianthus corymbosus (L. C. Rich.) Cogniaux. Since this species was not within the original material cited by D. Don, it cannot stand as the lectotype of the genus, and accordingly I select Charianthus purpureus D. Don, which is based on Melastoma coccinea Vahl, Eclog. 1: 48. 1796, not L. C. Richard, Act. Soc. Hist. Nat. Paris 1: 109. 1792. Charianthus alpinus (Sw.) R. A. Howard, comb. nov. Melastoma alpina Sw. Prodr. 71. 1788; Fl. Ind. Occ. 2: 800. 1797. Melastoma coccinea L. C. Rich. Act. Soc. Hist. Nat. Paris 1: 109. 1792. Charianthus coccineus (L. C. Rich.) D. Don, Mem. Wernerian Soc. 4: 328. 1823. David Don listed as the first species in his new genus Charianthus, .. coccineus (L. C. Rich.) D. Don, based on Melastoma coccinea L. C. Rich. However, he cited in synonymy Melastoma alpina Sw. giving as a refer- ence only Flora Indiae Occidentalis 2: 800. Pages 641 to 928 of this work are said to have appeared in 1797 (Stafleu, Taxonomic Literature, Reg. Veg. 52: 462. 1967). However, Swartz did refer to his earlier use 402 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 of the basic reference, his Prodromus 71. 1788. Don and subsequent authors have ignored this earlier valid basionym and a new combination is necessary. The supporting specimen cited by Swartz was collected by du Ponthieu in the alpine mossy forest region of the Soufriére on Guadeloupe, and such a specimen is in the Banks Herbarium (BM). ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 1972] PORTER, ZANTHOXYLUM FROM PANAMA 403 puTRotices: new, TP 77" — THREE NEW SPECIES OF ZANTHOXYLUM (RUTACEAE) FROM DARIEN PROVINCE, PANAMA ys DuNCAN M. PortER ~*~ 0005 ' LIKE MANY WIDESPREAD, species-rich, tropical woody genera, Zanthoxy- lum suffers from a lack of adequate collections in the herbarium, A few species are represented by numerous specimens, but for the vast majority there are few, one, or more often than that none. It is impossible to gain an adequate grasp of such a genus without examining a large suite of specimens from throughout its range; it is also impossible to have any understanding of specific relationships without doing so. Knowledge of Zanthoxylum is such that the relationships of the species described below from eastern Panama are obscure. Unfortunately, each is known only from the type collection. Zanthoxylum eliasii D. M. Porter, sp. nov. Zanthoxylo keller- manio affinis, sed inflorescentiae ramis puberulis et folliculis maturis 3 vel raro 1 vel 4 differt. Ficurrs 1 and 2. vType. “Darién: “Cerro Pirre, alt. 2,500-4,500 ft. Cloud forest and/or mossy forest. Spindly tree 4 inches DBH; fruits red.” 9-10 August 1967, J. A. Duke & T. S. Elias 13851 (mo, holotype; F, GH, iso- types). Spindly tree; branchlets glabrous. Leaves alternate, odd-pinnate; pet- iole and the rachis together 15-32 cm. long, canaliculate above, minute- ly puberulent to glabrate; leaflets 7, opposite or subopposite, ovate to oval, abruptly acuminate apically, inequilateral and more or less rounded basally, the margins entire, subcoriaceous, pellucid-punctate throughout, minutely and sparingly puberulent on the veins below, 10-22 cm. long and 5.5-8.5 cm. wide, the petiolules canaliculate above, to 1 cm. long. Carpellate inflorescences stout, axillary, minutely puberulent, 7.5-16.5 cm. long and to 10 cm. wide in fruit. Carpellate flowers with the pedi- cels stout, minutely puberulent, 4-7 mm. long in fruit; sepals 5, triangular, less than 1 mm. long; carpels 3(—4), connate. Follicles (1—)3(-4), sub- globose, sessile, more or less beaked, connate basally, punctate-glandular, glabrous, red at maturity, ca. 5 mm. in diameter; seeds subglobose, black, shiny, ca. 5 mm. in diameter. This species is named for Dr. Thomas S. Elias, an avid student of the Rutaceae. Zanthoxylum eliasii appears to be related to ‘Z. kellermanii P. Wilson of ‘British Honduras,”Guatemala,“Honduras, and”El Salvador. In Wilson’s key to the genus in North America (N. Am. Fl. 25: 178-180. 404 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 2, close-up of follicles. ’ men D. M. Porter. 1, photo. of holotype spec iasii soa ee ae ‘ a ea \ * Pd we oe wr : as a SS es tes Ficures 1 AND 2, Zanthoxylum el Ficures 3 AND 4, Zanthoxylum jaimei D, M. Porter. 3, photo. of holotype specimen; 4, close-up of follicles. VAVNVd WOW WOTAXOHLNVZ ‘NALYOd [ZL61 SOv 406 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 1911), which includes Panama, Z. eliasii would be determined as Z. keller- mani. However, the latter differs in having even-pinnate leaves, ap- pressed-pubescent branches of the inflorescence, and 2 or 3 mature fol- licles. vZanthoxylum jaimei D. M. Porter, sp. nov. Zanthoxylo arborescenti aemulans, differt foliolis ca. 9-11 et ramulis, rhachidibus, et inflorescentiae ramis puberulis. Ficures 3 and 4. ’TyPE. /Darién: “Cerro Pirre, alt. 2,500-4,500 ft. Cloud forest and/or mossy forest. Apparently an armed tree; branch picked up off the ground.” 9-10 August 1967, J. A. Duke & T. S. Elias 13793 (mo, holo- type). This species is named for Dr. James A. Duke, prodigious collector of Panamanian plants. The epithet dukei is not used because of the prior publication of Fagara duckei de Albu uerque (Publ. Inst. Nac. Pesq. Amazonia Bot. 27: 6. 1968) from the’Brazilian State ofVPara. I agree with Brizicky (Jour. Arnold Arb. 43: 80-83. 1962) and Hartley (op. cit. 47: 171, 172. 1966) that Fagara L. and Zanthoxylum L. are congeneric. Thus the Brazilian species should be regarded as Zanthoxylum duckei (de Albuquerque) D. M. Porter, comb. nov. States of ‘Baja California, Sinaloa, and Jalisco. The latter has 3 to 7 leaflets, and the twigs, leaf rachises, and branches of the inflorescences are densely hispidulous. “Zanthoxylum pucro D. M. Porter, sp. nov. Zanthoxylo bijugo aemulans, differt ramulis et foliis glabris et inflorescentiis scabridiusculis. Ficure 5. vTypr, “Darién: “Forest 1-4 mis. n. of Pucro. Armed tree 5” DBH; PORTER, ZANTHOXYLUM FROM PANAMA Province Darien ‘ ee cere t PARALYA BOTARTS > ant Ce Sa Porter Yorest 1.4 wis. a. of Pucte 3. A, Duke Bo. syc5) (3) dune 22, 1967 Ficure 5, Zanthoxylum pucro D. M. Porter, photo. of holotype specimen. 408 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 flowers green.” 22 June 1967, J. A. Duke 13029 (mo, holotype; GH, iso- type). Tree, armed with prickles; branchlets glabrous. Leaves apparently odd-pinnate; petiole and the rachis together 44-53.5 cm. long, glabrous, more or less canaliculate above; leaflets apparently alternate, oblong, in- equilateral, abruptly acuminate apically, oblique basally, the margins crenate, membranaceous, pellucid-punctate only at the bases of the crena- tions, glabrous, 6.5-17.5 cm. long and 3-6 cm. wide, the petiolules to 5 m ong, winged. Staminate inflorescences congested, paniculate, ter- minal (?), scabridulous. Staminate flowers green in bud, the pedicels to 1 mm. long, scabridulous; sepals 5, broadly triangular, glandular-punctate, the margins hyaline, ciliate, imbricate, the 3 outermost larger, ca. 1 mm. long; petals 5, obovate, glandular-punctate, yellow (?), imbricate; stamens 5, the anthers sagittate; the filaments inserted on the disc; ovary ovoid, sharply 5-lobed and each lobe bilobed, glandular-punctate, glabrous, less that 1 mm. in diameter, star-shaped in cross-section, the styles 3, free. Follicles unknown. Unfortunately, the type specimen was collected under circumstances of high temperature and high humidity, conditions not conducive to good preservation of plant materials. It is overgrown with fungi, and all the leaflets have fallen from the rachises. However, the material is sufficiently well preserved to show that it is unlike any other member of the genus yet described. In Wilson’s key to the genus in North America (N. Am. Fl. 25: 178- 180. 1911), ”“Zanthoxylum pucro would be determined as Z. bijugum (Engler) P. Wilson, from’Oaxaca, Mexico. However, the latter differs most obviously in having the twigs, foliage, and inflorescences densely pilose. Missourt BoTANICAL GARDEN St. Louts, Missourt 63110 ARNOLD ARBORETUM 1672-1072 CENTENNIAL Journal of the Arnold Arboretum Published poe a April, Hy, and October by = Arnold Arboretum, Harvard University. _ | : Subscription Price. 516. 00 per year, beginning January, 1972. a ‘Subseriptions and remittances should be. sent to Ms. Ellen Bernstein, a d Arbor etum, 22 a Avenue, Cambridge, Fe Volum es L- XLV, Saticeadl and some back suber of volumes 46-50 are available from the Kraus Reprint pad pote 16 East 46th peta New scsi New York kee aes G. Sc fubert, E editor oe So CE, Wood, ir neces eee CIRC ee 7 TON JOURNAL OF THE ARNOLD ARBORETUM Vot. 53 OcTOBER 1972 NUMBER 4 THE GENERA OF SAXIFRAGACEAE IN THE SOUTHEASTERN UNITED STATES #4 STEPHEN A. SPONGBERG SAXIFRAGACEAE A. L. de Jussieu, Gen. Pl. 308. 1789, “Saxifragae,’’ nom. cons. (SAXIFRAGE FAMILY) Variously pubescent annual [rarely biennial], or perennial often suf- fruticose or semisucculent herbs, unarmed or armed shrubs [small trees], or woody vines from fibrous roots (some producing rhizomes) ; herbs (sub- fam. Saxifragoideae) often reproducing vegetatively by stolons and bulbils. Leaves alternate, opposite (subfam. Hydrangeoideae), mostly basal, or clustered on short, lateral shoots (subfam. Ribesioideae), petiolate or sessile, stipulate or, more commonly, exstipulate; blades simple or pinnately [or palmately] compound, the leaf or leaflet margins entire or variously +Prepared for a generic flora of the southeastern United States, a ee project of the Arnold Arboretum and the Gray Herbarium of Harvard University m ible through the support of the National Science Foundation (Grant GB6459X, principal investigator, Carroll E. Wood, Jr.). This treatment follows the format establi in the first paper in the series (Jour. Arnold Arb. 39: 296-346. 1958). The area covered includes North and-South Carolina, Georgia, Florida, a Alabama, Miss ississippi, a ad with supplementary information in brackets. References that I have not seen are indicated by an asterisk. The literature listings gi in this p wontons 1 particularly with regard to sibitics Mittal /eeetonbaral references for the ‘genus Ribes am particularly indebted to Professor Wood for his patience, careful review of the manuscript, and help with taxonomic and editorial ora gt to Dr. Elizabeth A. Shaw for translations of certain Russian references; to Dr. Lorin I. Nevling, Jr., for information concerning the upalnvam ioe to Dr. Bernice G. Schubert for her careful ; and to Mrs. editing ancy Dunkly for checking bibliographic citations na retyping much of the manuscript. The illustrations were prepared by a cohort o ric Flora artists and represent the careful work of the late Dorothy rsh, hake A. Materials used in the eeuireyy of the illustrations have been kindly supplied by R. B. Channell, S. Leonard, N. G. Miller, R. E. Primack, the late H. F. L. Rock, D. H. Thomas, ‘K F. Tryon, R. E. Weaver, Jr., and K. A. Wilson, of collected by Dr. Wood and myself. 410 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 toothed, lobed, or cleft, the venation pinnately netted, palmate, or penni- parallel; hydathodes often associated with the marginal teeth (subfams. Saxifragoideae and Ribesioideae). Flowers regular or occasionally irregu- lar, perfect or imperfect (and the plants hermaphroditic, dioecious, poly- gamous, or polygamodioecious), or occasionally neuter (subfam. Hydran- geoideae), arranged in pedicellate, often bracteate racemes, cymes, panicles, or corymbs, or sometimes occurring singly. Calyx usually persistent, regu- lar or occasionally irregular, of 4 or 5 or 7-10(12) sepals, the sepals val- vate or imbricate in bud and almost always connate below, forming a shal- low to deep, sometimes ribbed floral cup with reflexed or ascending calyx lobes, the insertion of corolla and androecium hypogynous to epigynous; a glandular, nectar-secreting disc or annular ring often developed at the base of the floral cup, sometimes extending over the surface of the ovary. Petals usually as many as calyx lobes, occasionally lacking, or variable in number, imbricate or valvate in bud. Stamens as many as or twice as many as the calyx lobes or sometimes numerous [up to 100 or more] (subfam. Hydran- geoideae), occasionally differentiated into staminodia and functional sta- mens (subfam. Parnassioideae) ; filaments [sometimes with lateral appen- dages] tapering to the subcordate to linear, basifixed, 4-sporangiate (2- loculate at anthesis) anthers; pollen 2- or 3-celled when shed. Gynoecium 2- or 3-, S—7-, or 8-10(12)-carpellate, apocarpous or syncarpous, the styles usually as many as the carpels and free, occasionally united, sometimes forming a massive stylar column; stigmas capitate, occasionally with radiate stigmatic surfaces; ovary superior and free to inferior and adnate to the floral cup, 1-, 2-, 4-, 5—7, or 8-12-loculate with numerous anatropous, 1- or 2-integumented, crassinucellar or tenuinucellar ovules on axile or parietal placentae. Fruit a loculicidal, septicidal, or apically dehiscent capsule (or each carpel circumscissile in subfam. Penthoroideae), or a berry (sub- fam. Ribesioideae) ; seeds usually numerous, often winged; embryo small, straight with flat cotyledons, embedded in scant to copious endosperm; embryo sac of the Polygonum type, embryogeny corresponding to the Solanad or Asterad (subfam. Parnassioideae) types, or irregular after the initial stages, corresponding to no designated pattern (subfam. Ribesio- ideae). (Including Grossulariaceae DC., Hydrangeaceae Dumort., Iteaceae J. G. Agardh, Lepuropetalaceae Nakai, Parnassiaceae S. F. Gray, Pen- thoraceae Rydb. ex Britton, and Philadelphaceae D. Don.) TyPE GENUS: Saxifraga L. According to the traditional Englerian interpretation, a large Rosalian family of cosmopolitan distribution comprised of about 1200 woody and herbaceous species in 80 genera; 44 species in 15 genera in the south- eastern United States and about 18 additional genera in North America, primarily in the western Cordillera. An almost impossible family to distinguish or characterize clearly be- cause of its heterogeneity, the Saxifragaceae was divided into 15 sub- families (17 in the 1964 edition of Engler’s Syllabus) and numerous tribes and subtribes by Engler (1928). The subfamilies (and a number of the 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 411 tribes) exhibit a greater geographical, as well as morphological homo- geneity than the family as a whole, and the family has often been divided into segregate families that generally correspond to the subfamilial (or tribal) taxa established or accommodated within the Saxifragaceae by Engler. In current systems of classification, however, there seems to be little agreement about either the circumscription or the taxonomic ranks that Engler’s subfamilial taxa (particularly the woody groups) should be accorded, but, with the exception of Thorne, most investigators agree that division of the Saxifragaceae sensu Engler is both desirable and necessary if a more nearly natural and accessible taxonomy is to result. In view of this lack of agreement, the family is retained here provisionally in the traditional sense. Of the 17 subfamilies recognized in the current Syllabus, seven (sub- fams. Penthoroideae, Saxifragoideae, Vahlioideae, Francodideae, Eremo- synoideae, Lepuropetaloideae, and Parnassioideae) are almost exclusively herbaceous and have vessel elements with mostly simple perforation plates. With the exceptions of subfams. Vahlioideae (Vahlia Thunb., in tropical and subtropical Africa, Madagascar, and India), Eremosynoideae (Eremo- syne Endl., in Australia), and Francodideae (two monotypic genera in South America) these herbaceous subfamilies are, for the most part, dis- tributed in the temperate and cold-temperate regions of the Northern Hemisphere. In contrast, the ten remaining subfamilies (subfams. Ribesio- ideae, Baueroideae Pterostemonoideae, Hydrangeoideae, Tetracarpaeoideae, Iteoideae, Brexioideae, Escallonioideae, Montinioideae, and Phyllonomoi- deae) are predominantly woody taxa with vessel elements with mostly scalariform perforation plates, and, with the exception of the Ribesioideae and the major part of the Hydrangeoideae that have disrupted Arcto- Tertiary distributions in the Northern Hemisphere, these subfamilies are mostly confined to the Southern Hemisphere and tropical and subtropical regions. The greatest number of genera (ca. 30) occurs in North America, where genera in nine subfamilies (two restricted to Mexico: Phyllonomoideae [Phyllonoma Willd.| and Pterostemonoideae | Pterostemon Schauer]|) are represented. The Saxifragoideae and, to a lesser extent, the Hydrangeo- ideae are particularly well represented, especially in the western Cordillera, an area of diversification within subtribe Saxifragineae (ca. 17 genera) and tribe Philadelpheae (five or six genera) of the Hydrangeoideae. Savile (1954, 1961), in studies of the taxonomy of a group of micro- cyclic rusts of the genus Puccinia that occur on species of several genera of the Saxifragineae (including Mitella, Tiarella, Bergenia, Saxifraga, Chrysosplenium, Heuchera, Tolmiea, and Tellima), has gained indirect insights concerning the saxifragaceous hosts. He has concluded on the basis of observed evolutionary trends in the morphology of the rusts that they are “mostly of a single evolutionary group” that probably originated in the Himalayan region and there parasitized an underived saxifragaceous host-complex. The underived saxifragaceous complex probably crossed the Bering land bridge several times in the late Tertiary and Pleistocene times 412 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 and radiated into new niches in the western Cordillera. Migrations prob- ably occurred in the opposite direction after the initial North American radiation, and Savile has suggested that the oldest species of Mitella and Tiarella (both genera with eastern North American, western North American, and Asian species), and the oldest species of Heuchera and Saxifraga are roughly the same age. Differentiation among genera in North America has been primarily in adaptation to differing ecological situations and the characters that define genera include specialized adaptations for pollination and dispersal. Generic limits of a number of genera are open to dispute, but study of floral mechanisms and breeding systems will un- doubtedly aid in their resolution. In the Englerian sense, the family is not only heterogeneous morpho- logically, encompassing annual, biennial, and perennial herbs, shrubs, trees, and woody vines of essentially every ecological situation, but it also ex- hibits considerable anatomical, palynological, chemical, cytological, and embryological diversifications. The full range of this diversity is too great to discuss here, but information — to the genera occurring in our area is included in the generic discussio Except for various species ices are » cultivated i in rock gardens and peren- nial borders and some of the woody genera that are of considerable orna- mental value, the Saxifragaceae are of little economic importance. Cur- rants and gooseberries (Ribes) are cultivated commercially only on a rela- tively small scale. Key TO THE GENERA OF SAXIFRAGACEAE IN THE SOUTHEASTERN UNITED STATES General characters: Sensu ENGLER, a poorly defined family of annual, rarely biennial, or perennial herbs, shrubs, woody climbers, or trees; flowers with the ‘number of stamens mostly definite, carpels same number or he ely fewer than sepals and either free or partially to completely connate; sepals rarely free (insertion hypogynous), usually connate at least at base, forming a floral cup adnate to ovary; insertion of corolla and androecium perigynous to epigynous on Crassulaceae by carpels not subtended by glands and floral parts of a less regular numerical pattern. A. Plants herbaceous. B. Staminodia absent; styles not commissural; fruit a septicidal or circum- scissile capsule or debiacing apically between styles C. Carpels 5-7; fruit a 5-7-beaked capsule, each a circumscis- sile (subfam. Hatt Re Penthorum. C. Carpels 2 or 3; fruit a septicidal capsule or eines between styles (subfam. Sacttonenices e). D. Leaves te rately decompound; flowers imperfect; oy poly- @amodioetioee:: gis is yi oa a . Astilbe. D. Leaves simple; flowers perfect; plants monoecious. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 413 E. Corolla present; flowers in racemes or panicles. F. Petals pinnatifid or fimbriate; flowers in simple ra- cemes; cauline leaves subopposite. .... itella. F. Petals not rant a or fimbriate; cauline leaves (if present) alternat G. gs tialaies placentae parietal. Inflorescence racemose; stamens 10; car- w Oe: TNA es. he, ok os 4. Tiarella. H. Se ors paniculate; stamens 5; car- TES aa ae 5. Heuchera. pels equal. G. Ovary 2- lomlsta (at least above); placentae xile. 1 Leaves palmately lobed; stamens 5. ...... Loew. beac ka, gaye ee nse Boykinia. I. Leaves not palmately lobed; stamens 10. CiGws Oe ho VON so, 7. Saxifraga. E. Corolla absent; flowers solitary in leaf ‘axils or in leafy COMO ae ee . Chrysosplenium. B. Staminodia present: styles commissural; fruit a loculicidal capsule (subfam. Parnassioideae). J. Plant a diminutive winter annual to 2 cm. tall; oo basal and cau- Hine Sea ves rene oe cv ere, 9. Lepuropetalon. J. Plants perennial from rhizomes or caudices; leaves - basal... .. Parnassia. A. Plants shrubs or woody vin Leaves alternate or ve ee on short, alternate spur branches. L. Leaves palmately lobed; ovary inferior; fruit a see ie srg MEIOTIC ia ih 6 a che sie ak ee Ribes. L. Leaves entire; ovary superior; fruit a capsule (subfam. ehenitng BRR ice, iy yee GEL hee ea styles eal ego + - 12. ftea. K. Leaves opposite, sometimes whorled with 3 at a node (subfam. Hy- drangeoideae M. Petals usually 4; stamens 20-40; fruit an unribbed teak cap- sule, dehiscing longitudinally. Piuus eeoy vee 13. ladelphus. M. Petals 5 or 8-12; fruit a strongly ribbed capsule, Anges either apically or between the capsule ribs. N 2 or 3, free; fruit dehiscing Hey between the styles; nonclimbing shrubs. Wh bik os ae a ok es ydrangea. N. Stigmas capitate, with radiating stigmatic ok "fruit dehisc- ing between capsule ribs; woody vines climbing by roots. 15. Decumaria. FAMILY REFERENCES: AGABABIAN, V. S. Materials toward the palynosystematic study of the family Saxifragaceae s. 1. (In Russian.) Izv. Akad. Nauk Armian. SSR Biol. 14(2): 45-61. 1961. . Evolution of pollen in the orders Cunoniales and Saxifragales relative to several problems of their systematics and phylogeny. (In Russian.) Ibid. 17(1): 59-72. 1964. 414 JOURNAL OF THE ARNOLD ARBORETUM [voL. .53 BaILion, H. Saxifragacées. Hist. Pl. 3: 325-464. 1871 Bros gra transl. M. M. Hartoe, The natural history of plants. 3: 323-464. 1874. | BEAN, W. J. Trees and shrubs hardy in the British See ed. 7. Vol. 1. xviii + 703 pp. pls. 1-40. London. 1950 [Decumaria, 605, 606]; Vol. 2. x+ 636 pp. pls. 1-52. London. 1951 [Hydrangea, 104-115. pls. 10, 11; Itea, 146-148, pl. 13; Philadelphus, 410-420, pl. 34]; Vol. 3. viii + 664 pp. pls. 1-41. London. 1951. [Ribes, 165-179.] BenTHAM, G., & J. D. Hooker. Saxifragaceae. Gen. Pl. 1: 629-655. 1865. BILtek, G., & H. Kinpt. On the phenolic contents of the family Saxifragaceae. (In German.) Monatsh. Chem. 93(1): 85-98. 1962.* Botssiev, H. pe. Les Saxifragées du Japon. Bull. Herb. Boiss. 5: 682-695. 1897 CaNDOLLE, A. P. pe. Saxifragaceae. Prodr. 4: 1-54. 1830. [Saxifraga, 17-47, by N. C. SERINGE.] Curist, K. Beitrige zur vergleichenden Anatomie des Laubstengels der Caryo- phyllinen und Saxifrageen. 83 pp. Marburg. 1887.* Ciute, W. N. The meaning of plant names. XXV. Saxifragaceae. Am. Bot. 32: 11-18. 1926. Core, E. L. Ethnobotany of the southern Appalachian —— Econ. Bot. 21: 198-214. 1967. [Lists uses of Saxifraga and Ribes, 2 Cronguist, A. The evolution and classification of flowering _ x +396 pp. : Boston. 1968. [Rosales, 229-236; Hydrangeaceae, Grossulariaceae, Saxi- fragaceae. } DAHLGREN, K. V. O. Die Morphologie des Nuzellus mit bezonderer Beriicksichti- gung der deckzellosen Typen. Jahrb. Wiss. Bot. 67: 347-426. 1927. Zur Embryologie der Saxifragoideen. Sv. Bot. Tidskr. 24: 429-448. 1930. Danpy, J. E. The genera of Saxifragaceae. Pp. 107-118 in J. HuTcHINSON, Contributions towards a phylogenetic classification of flowering plants: VI. Bull. Misc. Inf. Kew 1927: 100-118. 1927. Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. 1966. [Data under several segregate families as well as Saxifragaceae. | Drppet, L. Handbuch der Laubholzkunde. Vol. 3. vii + 752 pp. Berlin. 1893. [Recognizes Ribesiaceae, 280-308; Saxifragaceae, 308-346. Est, P. vAN DER. — zur Kenntnis der Samenanlage der Saxifragaceae. Diss, Utrecht. 1909 ENGLER, A. mee RO aE In: A. ENctrr & K. Pranti, Nat. Pflanzenfam. III. 2a: 41-93. 1890-91; Nachtraége II-IV(1897): 180-182: Erganzungsh. 1(1900): 29; ed. 2. 18a: 74-226. 1928. [Ed. 1 of Pilanzenfari. , pp. 1-48, 1890; pp. 49-96, 1891. . Beitrage zur Entwicklungsgeschichte der Hochgebirgefloren; erlautert an der Verbreitung der Saxifragen, . . 113 pp. 8 col, pls. Berlin. 1916. (Abh. Akad. Wiss. Berlin Phys. 1916. Nr. 1 ErptMaNn, G. Pollen morphology and plant taxonomy. Angiosperms. xii + 539 pp. From tisp. Stockholm; Waltham, Mass. 1952. [Corrected reprint and new addendum. xiv + 552 pp. New York. 1966. Saxifragaceae sensu lato, 399-402. ———. Handbook of palynology: eee taxonomy — ecology. 486 pp. Foldout pollen diagram. Copenhagen. 1969. GAUMANN, E. Studien iiber die Entvicklnggeschict einiger Saxifragales. Rec. Trav. Bot. Néerl. 16: 285-322. 1972 | SPONGBERG, GENERA OF SAXIFRAGACEAE 415 GE.tus, L. Studien zur Entwicklungsgeschichte an Bliiten der ieee agg lato mit besonderer Beriicksichtigung des Androeceums. Bot. Jahrb. 253-303. GorENFLOT, R. Intérét taxonomique et phylogénique der caractéres stomatiques (application 2 a la tribu des Saxifragées). Boissiera 19: 181-192. a. & F. REAU. Types stomatiques et phylogénie des Saxifraginées (Saxifragacées). Compt. Rend. Acad. Sci. Paris 270: 2802-2805. 1970. GRAENICHER, S. Wisconsin flowers and their pollination. II ase tier and Grossulariaceae. Bull. Wis. Nat. Hist. Soc 4-95. 1907. ae A. Diagnostic characters of new species - seamiaties plants, collected n Japan by Charles Wright, Botanist of the United States North Pacific hid Expedition . . . with observations upon the relationships of the Japanese flora to that of North America, and of other parts of the North Temperate Zone. Mem. Am. Acad. Arts Sci. II. 6: 377-452. 1858. [Saxi- fragaceae, 388-390. | Grecuss, P. Bestimmung der mitteleuropaischen Laubhélzer und Straucher auf xylotomischer Grundlage. 183 pp. 250 pls., 4 special pls., 2 charts. Budapest. 1945, GuicnarD, L. Recherches sur le sac embryonnaire des phanerogames angio- spermes. Ann. Sci. Nat. Bot. VI. 13: 136-199. 1882. HameL, J. L. Notes preliminaires a |’étude caryologique des Saxifragacées. II. Les chromosomes somatiques de trois représentants du genre Bergenia Moench. Bull. Mus. Hist. Nat. Paris II. 20: 558-560. 1948; III. Les chromo- somes somatiques des trois représentants du genre Escallonia Mutis. Ibid. 21: 749-751. 1949; IV. Les chromosomes de six espéces appartenant a la tribu des Astilbinées. /bid. 752-756; VI. Les chromosomes somatiques des Kirengeshoma palmata Yatabe, Deinanthe coerulea Stapf et Schizophragma sain (Franch.) Oliv. Ibid. 23: 651-654. 1951. Lemersemne a tribution a l’étude cyto-taxinomique des Saxifragacées. These Fac. Sci. aay Paris Sér. A, no. 2632, no. d’ordre: 3504. pp. 113-313. 1953. [Also publ. in Revue Cytol. Cytophysiol. Vég. 14: 113-313. 1953.] Hecetmater, F. Zur Entwicklungsgeschichte endospermatischer Gewebek6rper. Bot. Zeit. 44: 529-539, 545-555, 561-578, 585-596. 1886 Hottie, G. Beitrage zur Anatomie der Saxifragaceen und deren systematische Verwerthung. Bot. Centralbl. 53: 1-9, 33-41, 65-70, 97-102, 129-136, 161-169, 209-222. HRYNIEWIECKI, B. Ein neuer Typus der Spaltéffnungen bei den Saxifragaceen. Bull. Acad. Sci. Cracov. Sci. Math. Nat. B. 1912: 52-73. pls. 3-6. 1913. Huser, H. Rosiflorae. Ju: G. Hect, Illus. Fl. Mittel-Europa. ed. 2. 4: 21-448. 1961-1966. [Recognizes Philadelphaceae, 37-42, Grossulariaceae, 43-61, and Iteaceae in the Cunoniales and Penthoraceae, Saxifragaceae, 126-224, and Parnassiaceae, 225-230, in the Saxifragales. | . Die Verwandtschaftsverhaltnisse der Rosifloren. Mitt. Bot. Staatssam. Miinchen 5: 1-48. 1963. Hutcuinson, J. Contributions towards a phylogenetic classification of flowering plants. VI. Hydrangeaceae and Saxifragaceae. A. The genera of Hydran- geaceae. Bull. Misc. Inf. Kew 1927: 100-107. 1927. [Includes the Phila- delpheae as a subfamily of the Hydrangeaceae. | . The families of flowering plants. Vol. 1. Dicotyledons. ed. 2. xi + 510 pp. Oxford. 1959. [Recognizes Philadelphaceae, Hydrangeaceae, and Gros- 416 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 the Saxifragales, 456-466 . The genera of flowering plants. Vol. 2. xi+ 659 pp. Oxford. 1967. ‘ Evolution and phylogeny of flowering plants. gicretpee facts and xXvi + 717 pp. 2 charts. London & New York. 9, IKUSE, M. Pollen grains of Japan. xi-+ 303 pp. pls. ee Frontisp. Tokyo. 1956. [Descriptions of pollen types for Japanese species of Jtea, Ribes, Tiarella, Saxifraga, Mitella, Astilbe, Boykinia, Chrysosplenium, Parnassia, and Hydrangea. | Jay, M. Recherches chimiotaxinomiques sur les plantes vasculaires. Distri- bution des flavonoides chez les Saxifragacées. Compt. Rend. Acad. Sci. Paris. D. 264: 1754-1756. 1967 ———. Quelques problémes taxonomiques et phylogénétiques des Saxifragacées vus a la lumiére de la biochemie flavonique. Bull. Mus. Nat. Hist. Paris. II 42: 754-775. 1970 [1971]. Jonsson B. Om embryosickens untveckling hos Angiospermerna. Acta Univ. Lund. 16(5): 1-86. pls. 1-8. 1879-80. [Saxifraga and Chrysosplenium, p. 45 and pl. VI, figs. 23—30.] Kaneuira, R. Anatomical characters and ey ee of Formosan woods with critical remarks from the climatic po of view. 317 pp. pls. 1-50. Taihoku. 1921. [Saxifragaceae, 102-105. Knurtu, P. Handbook of flower pollination. (English transl. J. R. A. Davis) Vol. 2. vili+ 703 pp. Oxford. 1908. [Saxifragaceae sensu lato, 395—-422.] Kucter, H. Zum Problem der Dipterenblumen. Osterr. Bot. Zeitschr. 102: 529-541. 1955. [Considers Saxifraga, Chrysosplenium, and Parnassia.] . Blitendkologie. ed. 2. xi + 345 pp. Stuttgart. 1970. [Includes notes on Saxifraga, Chrysosplenium, Parnassia, and Ribes. Lepecur, A. Recherches embryogeniques sur quelques Dicotyledones dialy- petales. Ann. Sci. Nat. Bot. XI. 13: 1-160. 1952. Lr, H. Floristic relationships between eastern Asia and eastern North America. Trans. Am. Philos. Soc. 42: 371-429. 1952. [Reprint ed.: Morris Arb. Monogr. ii + pp. 371-429 as above. 1971. Includes distribution maps for Penthorum, Decumaria, Astilbe, and Ttea.] Love, A., & D. Love. Studies on the origin of the Icelandic flora. II. Saxi- fragacene. Sv. Bot. Tidskr. 45: 368-399. 1951. Lussock, J. A contribution to our knowledge of seedlings. Vol. 1. viii + 608 pp. London. 1892. [Saxifragaceae, 502-512.] Martin, A. C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513-660. 1946, or 572, pl. 30; Saxifragaceae, 584, pl. 37; Penthoraceae, 589, pl. 4 Mavrrrzon, ae Studien iiber ie Embryologie der Familien Crassulaceae und Saxifragaceae. 152 pp. Lund. 1933. Contributions to the embryology of the uae Rosales and Myrtales. Lunds Univ. Arsskr. IT. Sect. 2. 35(2): 1-121. Metsner, C. F. Plantarum vascularium genera. 2 yas ‘Eepae. 1837-1843. Metcatre, C. R., & L. CHarx. Anat. Dicot. 1: 553-557, 565-568, 572-577. 1950. [Data under Saxifragaceae, Escalloniaceae, Grossulariaceae, and Hyd- rangeaceae. | Moore, D. M. A check list of the ligneous flora of Arkansas. Proc. Arkansas Acad. Sci. 1: 41-55. 1941. [Lists species of Hydrangea, Philadelphus, Ribes, and Jtea.]} sulariaceae in the Cunoniales, 156-167; Saxifragaceae and Parnassiaceae in 6.] 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 417 Moreau, F. Apport des caractéres stomatiques 4 la taxinomie et 4 la phylogénie des Saxifragées. Bull. Soc. Bot. France 118: 381-427. 1971. Morr, E. Ndjalii chesdaiovehutuainitie Untersuchungen am Gynoeceum der Saxifragaceen. Ber. Schweiz. Bot. Ges. 60: 516-590. 1950 MorvittEz, F. L’appareil conducteur des feuilles des Saxifragacées. Compt. Rend. Acad. Sci. Paris 167: 555-558. 1918. MU ter, H. The fertilization of flowers. (English transl. D. W. Thompson, with a preface by Charles Darwin.) xii + 669 pp. London. 1883. Norpat, A., & D. O1setH. On the occurrence of sedoheptulose in certain species and erie of the plant family Saxifragaceae. Acta Chem. Scand. 6: 446, 447. 1952. PALMATIER, E. ‘a Some studies on the floral anatomy and ee of the Sadifrapaceat: Cornell Univ. Abstr. Theses 1943: 345, 346. 194 Raprorp, A. E., H. E. Antes, & C. R. Bett. Manual of the vascular flora of the Carolinas. Ixi + 1183 - eign Hill. 1968. [Saxifragaceae, 518-529.] Racuavan, T. S., & V. K. Srinivasan. A contribution to the life history of Vahlia viscosa Roxb. and V. permet Proc. Indiana Acad. Sci. B. 15: 83-105. REcorp, S. J. The wood of Escallonia tortuosa. Trop. Woods 26: 12, 13. 1931. —. Classifications of various anatomical features of dicotyledonous woods. Trop. Woods 47: 12-27. 1936. Reuper, A. Manual of cultivated trees and shrubs. ed. 2. xxx-+ 996 pp. New York. 1940. [Saxifragaceae, 263-311.] . Bibliogr. Cult. Trees Shrubs. xl + 825 pp. Jamaica Plain. 1949. [Saxi- fragaceae, 190-218. ] ReENbLE, A. B. The classification of flowering plants. Vol. 2. xix + 640 pp. Cambridge. 1963. [Saxifragaceae, 319-326. ] Ricxett, H. W. Wildflowers of the United States. Vol. 2. The Southeastern States. Part. 1. New York. 1966 Rosertson, C. Flowers and insects: lists of visitors of four hundred and fifty- three flowers. 221 pp. Carlinville, Illinois. 1938. [Saxifragaceae, 186, 187.] Rosenpaut, C. O. Die nordamerikanischen Saxifraginae und ihre Verwandt- schafts-Verhiltnisse in Beziehung zu ihrer geographischen Verbreitung. Bot. Jahrb. 37(Beibl. 83): 1-87. pls. 4, 5. 190 A revision of the genus Sullivantia. Res. Publ. Univ. Minn. Stud. Biol. Sci. 6: 401-427. ye SAMUELsson, G. Studien iiber die Entwicklungsgeschichte der Bliiten einiger Bicornes-Typen: ein Beitrag zur Kenntnis der systematischen Stellung der Diapensiaceen und Empetraceen. Sv. Bot. Tidskr. 7: 97-188. 1913. [Saxifragaceae, 131.] Savitz, D. B. O. Taxonomy, phylogeny, host relationship, and phytogeography of the microcyclic rusts of Saxifragaceae. Canad. Jour. Bot. 32: 400-425. 1954 . Limited penetration of barriers as a factor in evolution. Evolution 13: 333-343. 1959. [Saxifragaceae, 336-338. ] . Evolution of Saxifragaceae from a mycologist’s viewpoint. Pp. 169- 172 in Recent advances in botany, vol. I. (From lectures and symposia presented to the IX Int. Bot. Congr., Montreal, 1959.) xvi+ 947 pp. Toronto. . Co-ordina ted studies of parasitic fungi and flowering plants. Nat. Canad. 98: 535-552. 1971. [Saxifragaceae, 548.] 418 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Sax, K. Chromosome nora in the ligneous Saxifragaceae. Jour. Arnold Arb. 12: 198-206. pl. 3 SCHLECHTER, R. Die pensation’ Papuasiens. Bot. Jahrb. 52: 118-138. 1914. (LAUTERBACH, Q., Beitrage zur Flora von Papuasien IV: 32 SCHOENNAGEL, E. Chromosomenzahl und Phylogenie der Saxifragaceen. Bot. Jahrb. 64: 266-308. 1931. ScHULzE-MeEnz, G. K. Rosales. Jn: H. Metcuior, ed., Engler’s eertpene: der Pflanzenfamilien. ed. 12. 2: 193-242. 1964. [Savifrasdeeats 201-206. ] Sewet, C. A. J. Beitrage zur Anatomie der Saxifrageen. 51 pp. Kiel. 1890. Sxovstep, A. Cytological studies in the tribe Saxifrageae. Dansk Bot. Ark. 8(5): 1-52. 1934. SMALL, J. K. Two new genera of Saxifragaceae. Bull. Torrey Bot. Club 23: 18-20. pls. 256, 257. 1896. [Jepsonia and Saxifragopsis. | P. A. Rypperc. Saxifragaceae. N. Am. FI. 22: 81-158. 1905. Sovutces, R. Les relations embryogeniques des Crassulacées, Saxifragacées et Hypericacées. Bull. Soc. Bot. France 83: 317-329. 1936. Sterner, M., & H. Lorrier. Stickstoffbasen in Eiweissabbau héherer Pflanzen II. Histochemische Studien iiber Verbreitung, Verteilung und Wandel des Ammoniaks und der fliichtigen Amine. Jahrb. Wiss. Bot. 71: 463-532. 1929. SUBRAMANYAM, K. Embryology in relation to systematic botany with particular reference to the Crassulaceae. Pp. 94-112 in Plant Embryology: a sym- posium. CSIR, New Delhi. 1962. [Embryological comparisons of the Cras- sulaceae with the Saxifragaceae and other nage | TakutajAn, A. Die Evolution der Angiospermen. viii + 344 pp. Jena 1959. ———. Flowering plants: origin and aL "(PransL. c Jerrrey.) x + 310 pp. Edinburgh. 1969. THorNE, R. F. Synopsis of a putatively phylogenetic classification of the Boweerine plants. Aliso 6: 57-66. 1968. THovuvENIN, M. Recherches sur 2 structure des Saxifragacées. Ann. Sci. Nat. Bot. VIL. 12: 1-174. pls. 1-22. 1890. TrecHeM, P. van. Structure de sabia ovules et parti qu’on eu peut tirer pour améliorer la classification. Jour. Bot. Morot 12: 197-220. 1898. [Saxi- fragaceae, 205-208, 216, 217.] Tippo, O. Comparative anatomy of the Moraceae and their presumed allies. Bot. Gaz. 100: 1-99. 1938. [Considers among others, the Grossulariaceae, Hy- drangeaceae, Cunoniaceae, and Escalloniaceae. ] Torrey, J., & A. Gray. A flora of North America. Vol. I. xiv-+ 711 pp. New York, 1838-1840. [Saxifragaceae, 562-595; Grossulariaceae, 544-553; Penthorum, 561, 562, included in Crassulaceae; these treatments published in 1840. Vines, R. A. Trees, shrubs and woody vines of the Southwest. xii + 1104 pp. Austin, Texas. 1960. [Saxifragaceae, 302-322 «| WakapayAsHI, M. On the affinity in Saxifragaceae s. lato with special reference to the pollen morphology. (In Japanese; English summary.) Acta Phyto- tax. Geobot. 24: 128-145. 1970. Warmine, E. ma i, Morphology and biology. Medd. Grgnl. 36: 169-236. 1909. We cu, M. B. Notes on the principal indigenous timbers of the natural order Saxifrageae. Jour. Proc. Roy. Soc. New S. Wales 59: 276-292. pls. 9-13. 1926. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 419 WHEELOCK, W. E. A list of species of the smaller herbaceous genera of North American Saxifragaceae. Bull. Torrey Bot. Club 23: 67-78. 1896. Woop, C. E., Jr. Some floristic ae alin the eapeaniy Appalachians and western North America. Pp. 331-404 im P. C. H ed., The distribu- tional history of the biota of the sepals rr Part II. Flora. Research Div. Monogr. 2. Va. Polytech. Inst. State Univ. Blacksburg. 1971. [Includes numerous distribution patterns in the Saxifragaceae. | Subfam. PENTHOROIDEAE Engler Penthorum Linnaeus, Sp. Pl. 1: 432. 1753; Gen. Pl. ed. 5. 197. 1754. a . Erect, fibrous-rooted, often semiaquatic perennial herbs up to 10 dm. tall, multiplying, spreading, and perennating by horizontal stolons or rhizomes. Stems terete, glabrous, and often decumbent below, angled, finely pube- scent, and usually widely branched above. Leaves bright green and shining, alternate, and short petiolate; blades thin, elliptic with an attenuate or nar- rowly cuneate base, an acuminate apex, and serrulate margins; venation pinnate with a prominent midvein. Flowers perfect, arranged in secund, scorpioid [or corymb-like], terminal and axillary cymes. Calyx of 5 (6-7) sepals, united below, forming a shallow, saucer-shaped floral cup; calyx lobes free above, ovate with irregularly toothed margins and acute apices, erect during anthesis, becoming reflexed in fruit, persistent. Corolla absent or comprised of 1—7 greenish [or white], lanceolate, slightly clawed petals inserted on the rim of the floral cup, usually shorter than the calyx lobes. Stamens 10, inserted on the edge of the floral cup; filaments terete, taper- ing only slightly to the basifixed, 2-loculate, longitudinally dehiscent, oblong anthers; pollen subprolate, 3-colporoidate, the sexine as thick as the nexine. Gynoecium 5-7 (typically 5 or 6) carpellate, syncarpous and adnate to the floral cup below the placental area; styles erect during anthe- sis, terminated by slightly rounded, capitate stigmas; each carpel with many crassinucellar, bitegmic, anatropous ovules on a well developed, pendulous, marginal placenta on the adaxial suture above the syncarpous region. Carpels becoming obliquely oriented in fruit, producing a 5—7 beaked, reddish capsule, the beak of each carpel circumscissile above the syncarpous region of the gynoecium. Seeds many, small, ellipsoid to obovoid, the surface tuberculate to echinate; embryo large, straight; endo- sperm of the ab initio Cellular type, scanty. Base chromosome numbers 8,9. Type spectes: P. sedoides L. (Name from Greek, pente, five, and horos, a mark, from the characteristically 5-merous plan of the flower.) — DITCH-STONECROP. Two species, Penthorum sedoides, 2n = 18, in eastern North America, and P. chinense Pursh, 2n = 16, in eastern Asia, ranging from eastern Siberia and Manchuria through Korea, Japan, and China. Penthorum sedoides grows in the wet, muddy soils of river flood plains, in swamps and low woodlands, along ditches, and in fallow fields from New Brunswick, Maine, and Quebec southward through Florida and westward to southern Ontario, Minnesota, Nebraska, Kansas, Oklahoma, and eastern Texas. 420 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 carpel above the syncarpous region, X 8; d, e, fruit showing circumscissile stylar beaks of s 50; g, perennating horizontal rhizome, X 3/4. fruit, beginning to dehisce, X 4; everal carpels, x 5; f, seed, X Although less frequently encountered in the mountains of our region, P. sedoides occurs in every physiographic province and probably in every Coastal Plain county. Originally considered conspecific with P. sedoides, P. chinense differs from the American species in its lower chromosome number, its narrower leaves, shorter and _ thicker styles, and its fewer-flowered, corymb-like inflorescences. Additional species, P. intermedium Turcz. and P. humile Regel, attributed to the genus from Asia are probably based on plants that are variants of P. chinense. Vegetative and floral variation is considerable in both species of Pen- thorum, and insect damage to the thin, shining leaves gives the plants a weedy aspect. Greatest vegetative differences seem to occur in populations that are subject to periodic inundation and alternating periods of drought; during wet seasons the plants may be semi-aquatic. Differences in the number of carpels, sepals, and petals, however, do not appear to be cor- related with habitat, and great variability is often evident within the flowers of a single inflorescence. Flowering occurs more or less continuously throughout the summer months and until frost, and a great number of small seeds are produced 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 421 within each carpel of the flower. Vegetative reproduction and perennation of the plants are accomplished by horizontal stolons or rhizomes that are produced from the rootstock towards the end of the growing season (see Fic. 1, g). Rhizome initiation appears to be triggered by shortening day length. Morphologically well marked from other saxifragaceous genera by its reddish capsules with five to seven circumscissile carpels, Penthorum has not always been included within the Saxifragaceae. Alternatively placed in its own monogeneric family, the Penthoraceae Rydb. ex Britton, or in- cluded in the Crassulaceae, the floral structure of Penthorum seems to be intermediate between the Crassulaceae and Saxifragaceae. The union of the sepals into a shallow floral cup and the basally syncarpous gynoecium (not subtended by a gland) ally the genus to the Saxifragaceae, whereas the number of floral parts suggests affinities within the Crassulaceae. The ob- viously crassulaceous genera Diamorpha Nutt. and Triactinia Hook. f. & Thoms. exhibit shallow floral cups and syncarpous gynoecia, and De Candolle (1828) placed Penthorum and Diamorpha in the tribe Cras- sulaceae Anomalae. Torrey & Gray (1840) also grouped these two genera as the tribe Diamorpheae in the Crassulaceae, but Baillon (1871) re- moved Penthorum to a monotypic tribe in the Saxifragaceae, Berger (1928) omitted Penthorum from his treatment of the Crassulaceae, and Engler (1928) included the genus in a monotypic subfamily in the Saxifragaceae that he considered to be the least derived in the family. On the basis of floral morphology, Quimby (1971) has given ire support to the inclusion of Penthorum in the Saxifragaceae, inasmuc he concluded that the Crassulaceae is a natural group if Penthorum is ex- cluded. From embryological evidence, Mauritzon (1933, 1939) claimed strong reinforcement for placement of Penthorum in the Saxifragaceae, and Jay (1971) concluded from chemical data that Penthorum is significantly different from the Crassulaceae but very similar to Bergenia and Pelto- boykinia in the Saxifragoideae. Cronquist, while admitting that the place- ment of Penthorum is largely a matter of opinion, has included the genus in the Saxifragaceae, while Takhtajan (1959) has chosen to recognize the Penthoraceae and has noted the exclusively scalariform perforation plates of the vessels in these herbs as a diagnostic family character. Penthorum is occasionally a noxious weed that may compete with crop plants or foul irrigation ditches. Local medicinal usages of Penthorum in- clude its use as a demulcent, astringent, and laxative. REFERENCES: Under family references see BAILLON; CRONQUIST; ENGLER, 1928; ERDTMAN, 1952; Hutcuinson, 1969; Jay, 1971; Lt; MartTIN; Mauvritzon, 1933, 1939; RIcKErr, pl. 93; SCHOENNAGEL; SkovsTED: TAKHTAJAN, 1959; and TorREY & Gray. BaLpwin, J. T., Jr., & B. M. Speese. Penthorum: its chromosomes. Rhodora 53: 89-91. 1951. 422 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Bercer, A. Crassulaceae. Nat. Pflanzenfam. ed. 2. 18a: 352-483. 1928. CANDOLLE, A. P. pe. Penthorum. Prodr. 3: 414. 1830 Hara, H. Contributions to the study of variations in the Japanese plants closely related to those of Europe or North America. Part 1. Jour. Fac. Sci. Univ. Tokyo Bot. 6: 29-96. 1952. [Penthorum, 62.] Quimpy, M. W. The floral morphology of the Crassulaceae. 42 pp. pls. 1-11. Published by the author, University, Mississippi, 1971. [Ph.D. thesis, Cornell Univ. Ithaca. 1939.] Rocen, T. Beitrag zur Embryologie der Crassulaceen. Sv. Bot. Tidskr. 22: 368- 376. 1928. [Chromosome number of = 8 reported for P. sedoides should be for P. chinense. Rypperc, P. A. Penthoraceae. N. Am. Fl. 22: 75. 1905. TiecHeM, P. van. Sur le genre Penthore considéré comme type d'une famille nouvelle les Penthoracées. Jour. Bot. Morot 12: 150-154. 1898. [Also pub- lished in Ann. Sci. Nat. VIII. 9: 371-376. 1899.] Subfamily SAXIFRAGOIDEAE A. Br. Tribe SAXIFRAGEAE DC. Subtribe Astilbinae Engler 2. Astilbe Buch.-Ham. in D, Don, Prodr. Fl. Nepal. 210. 1825. sessile, ovate to elliptic in outline and often irregularly 1-3(5)-cleft or lobed; leaflet margins coarsely and irregularly double serrate or crenate, the teeth and lobes with mucronate tips; bases of leaflets oblique, truncate, or cordate, the apices acuminate; venation pinnately netted. Inflorescence a widely spreading, many-branched, plume-like [or congested and _ spike- like], terminal bracteate panicle, the flowers short pedicellate, becoming re- flexed in fruit, perfect or imperfect (in ours, the flowers either perfect, appearing perfect but apparently functionally staminate, or carpellate and the plants appearing to be hermaphroditic, polygamous, or functionally polygamodioecious). Calyx of 5(7-10) deltoid sepals, connate at the base, forming a shallow campanulate to obconic floral cup with 5(7-10) short, erect, imbricate and persistent calyx lobes. Corolla absent [or consistently present] or variously of 1-5 erect, white, yellowish [pink, red, or purplish], spatulate to subulate petals inserted on the inner surface of the floral cup. Androecium abortive in imperfect flowers; when developed, obdiplostemon- ous, the [5, 8, or] 10 stamens erect, about twice as long as and inserted at the base of the floral cup, the filaments tapering gradually to the basifixed, cordate-orbicular, tetrasporangiate anthers; anthers white or yellowish; pollen grains 3-colporoidate. Gynoecium 2(rarely 3)-carpellate, the carpels connate below, tapering into two free styles and terminated by capitate stigmas; ovary superior with a poorly differentiated disc at the 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 423 base, 2-loculate, with numerous anatropous, crassinucellar, bitegmic ovules on axile placentae in the connate region. Fruit a 2 -beake d capsule com- posed of the two follicle-like carpels, each dehiscing longitudinally in the stylar region. Seeds numerous, oblong-cylindrical, the lustrous, reddish- brown seed coat winged and often twisted at both ends; embryo small, straight, embedded in endosperm of the ab initio Cellular type; embryo sac of the Polygonum type. Base chromosome number 7. (Hoteia Morr. & Decne.) TYPE species: A. rivularis Buch.-Ham. (Name from Greek, a, without or lacking, and stilbe, brilliance or sheen, in reference to the foliage of A. rivularis in contrast to that of Aruncus (Rosaceae), which it resem- bles.) — FALSE GOATSBEARD, FEATHER-TREES. A well-defined genus with one or perhaps two species endemic to the southern Appalachians in North America and about 23 in eastern Asia from Japan through China into the Himalayan region and extending southward into New Guinea. Astilbe biternata (Vent.) Britt. (Tiarella biternata Vent., A. decandra D. Don), 2m = 28, an upright, sometimes branched herb up to one or two meters high with glandular trichomes and large, compound leaves is restricted to the Appalachian Mountains from Mingo County, West Virginia southward through Virginia and Kentucky, to the Carolinas, Tennessee, and northwestern Georgia. A second species, A. crenatiloba (Britt.) Small (A. decandra var. crenatiloba Britt., A. biternata var. crenatiloba (Britt.) Wheel.), has been recognized from the slopes of Roan Mountain on the North Carolina-Tennessee boundary, but this taxon is probably only a local variant of A. biternata with crenately lobed leaflets. Growing in rich woods and on moist, shaded or sometimes open seepage slopes, Astilbe biternata is often confused with Aruncus dioicus (Walt.) Fern. (Rosaceae), and this species pair, as well as Asiatic species of both genera, are striking examples of convergent evolution. Occupying the same or similar habitats as Astilbe biternata, Aruncus dioicus is more widely distributed in eastern North America, occurring on the Piedmont and in the mountains from Pennsylvania southward through the Carolinas and into northern Georgia and Alabama and westward into Arkansas, Missouri, and Iowa. Aruncus dioicus can easily be separated from Astilbe biternata, however, by its simple trichomes, 15 or more stamens in staminate flowers, and/or its three or four carpels in carpellate flowers. In carpellate flowers of Astilbe biternata the androecium is abortive, and plants with apparently perfect flowers are probably functionally staminate inasmuch as the carpels are relatively undeveloped and the stigmas appear to be undifferentiated and unreceptive. Although experiments have not been conducted and despite occasional fruit set in inflorescences of perfect flowers, A. biternata seems to be largely dioecious. Anthesis occurs from May into August, and pollination is undoubtedly by insects. The fruits mature during the late summer and fall, and the numerous small seeds are probably scattered from the dehisced capsules by wind and passing animals. Engler (1928) divided the genus into sect. AsTILBE (sect. Compositae 424 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Engl.), which is composed of species with compound leaves, and sect. SIMPLICIFOLIAE, which includes only one or two species with simple leaves of the Himalayan region. Within sect. Aste, A. biternata appears to be intermediate between a group of apetalous species and other species with flowers that consistently have well developed corollas. Flowers of some plants of A. biternata are apetalous, while in other plants the flowers have corollas, although there is considerable variation in the number, size, and shape of the petals. Hamel (1953) hypothesized that the variable corolla trait in Astilbe biternata might have a chromosomal and genetic basis that reflects the A number of the Asian species are used as garden ornamentals, The numerous garden hybrids (involving Astilbe japonica (Morr. & Decne.) A. Gray, A. astilboides Lemoine, A. Thunbergii (Sieb. & Zucc.) Miq., and A. Davidii (Franchet) Henry) were mostly developed by Lemoine, in France, and by Arends, in Germany. In the Philippine Islands the Ingorots dried leaves of A. philippensis Henry for smoking, but other uses of Astilbe have not been reported. REFERENCES: Under family references see DALHGREN, 1930; DANDy; Dr Botssteu; ENGLER, 1890, 1928; HAMEL, 1949, 1953; IkusE; Jay, 1971; Li; Morr; RicKert, fl. 93; SCHOENNAGEL; SKovsTep; SMALL & RYDBERG; THOUVENIN; and Woop. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 425 Conrapr, R. Astilbe. (In Norwegian.) Norsk Hagetid. 76: 192, 193. 1960.* Hamet, J. L. Notes preliminaires a l'étude caryologique des Saxifragacées. — IV. Les chromosomes de six espéces appartenant a la tribu des Astilbinées. Bull. Mus. Hist. Nat. Paris II. 21: 752-756. 1949. Henry, A. The genus Astilbe. Gard. Chron. III. 32: 95; 154-156; 171. 1902. Junce, H. Die neuen Astilbe Arendsi-Hybriden, die Arends’schen Astilbe Thun- bergi major-und Astilbe chinensis-Hybriden. Miller’s Deutsch. Gart. Zeit. 22: 544-546. 1907. Knott, F. Beitrige zur Kenntniss der Astilbe-Arten Ostasiens. Bull. Herb. Boiss. II. 7: 127-135. 1907. _ Studien zur Artabgrenzung in der Gattung Astilbe. Akad. Wiss. Wien Sitz-ber. 118: 45-88. pls. 1-4. 1909. Naxal, T. Notulae ad plantas Japoniae et Coreae XX. Bot. Mag. Tokyo 33: 41-61. 1919. [Astilbe, pp. 54, 55; also see paper XXVIII in this series, Ibid. 36: 117-128. 1922.] Reuper, A. Webb’s “Study of Spiraea.” Bot. Gaz. 34: 246. 1902. [Points out misidentifications in Webb’s study (see below); Webb studied Astibe japon- ica, not Spiraea japonica, and Aruncus astilboides, not S. astilboides. Spiraea salicifolia was correctly identified. ] Scuaeprr, H. Untersuchungen iiber den Habitus von Aruncus, Astilbe und einiger ahnlicher Pflanzen. Beitr. Biol. Pflanzen 46: 371-387. 1970. Wess, J. E. A morphological study of the flower and embryo of Spiraea. Bot. Gaz, 33: 451-460. 1902. [See A. REHDER above. ] Wuson, E. H. Astilbe grandis. Gard. Chron. III. 38: 426. 1905. Subtribe Saxifraginae Engler 3. Mitella Linnaeus, Sp. Pl. 1: 406. 1753; Gen. Pl. ed. 5. 190. 1754. Low, pubescent perennial herbs from scaly, creeping rhizomes [often producing leafy stolons toward the end of the growing season]. Basal leaves alternate from the rhizome and long petiolate; leaf blades 3—5 (—7)-lobed [or entire; orbicular, reniform, or], broadly ovate in outline with a cordate ase; margins crenate with mucronate lobes; venation palmate; petioles slender, dilated at the stipulate base, the stipules inconspicuous, basally adnate to the petiole, membranaceous above with acute apices. Cauline leaves 2, 3 [or 0], subopposite to opposite or alternate, subsessile or dis- tinctly petiolate. Flowers perfect, short [to long] pedicellate and well spaced in elongate [several to] many-flowered simple racemes that become + secund in fruit; rachis of raceme and pedicels finely glandular pube- scent, each pedicel subtended by a small, entire [pinnatifid, or fimbriate] membranaceous bract. Sepals 5, persistent and united below, forming an oblong patelliform [campanulate, or turbinate] floral cup, the calyx lobes erect or slightly recurved at the acute apices [or the sepals connate at the base and forming a strongly flattened floral cup with the calyx lobes strongly reflexed]; floral cup with a ring of glandular tissue on the inner surface below the insertion of the anthers. Corolla white | greenish-yellow, or purple], of 5 [or 0], early deciduous petals inserted on [or just below] the rim of the floral cup, sessile or slightly clawed, the blade pinnately 426 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Ficure 2. Mitella. a—j, M. diphylla: a, flowering plant, showing simple race- mose inflorescences and paired, opposite cauline leaves, X 1/2; b, detail of raceme with bracts subtending flowers, X 3; c, flower showing collar of glandular issue inner surface of floral cup beneath insertion of anthers, X 10; 4d, longitudinal section through flower, showing glandular collar and parietal placenta, 12; e, detail of dehisced anther, 25; f, immature capsule after corolla disarticulation, < 6; g, top view of immature capsule showing line of dehiscence between styles, X 6; h, transverse section through floral cup and capsule, x 6; i, erect “splash-cup” capsule after dehiscence, x 6; J, seed, >< 12; ? several [or numerous] crassinucellar, bitegmic ovules oriented on 2 par- 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 427 several [to numerous], narrowly ovoid, black and shining, the surface pustulate to tuberculate [or smooth], endosperm usually copious, of the Helobial [or ab initio Cellular] type. Base chromosome number 7. (In- cluding Mitellastra (Torrey & Gray) Howell, Ozomelis Raf., Pectiantia Raf., Drummondia DC. non W. J. Hooker, and Mitellopsis Meisner.) Lecrorypr SPECIES: M. diphylia L.; see P. A. Rydberg, N. Am. Fl. 22: 91. 1905. (Name a diminutive of Latin, mitra, a cap or turban, an allusion to the miter-like appearance of the flowers and undehisced capsules.) — MITREWORT, BISHOP’S-CAP. A vegetatively homogeneous genus (leaves stipulate, basal) of about 20 species with centers of diversity in the Cordillera of western North America and eastern Asia; Mitella diphylla L. in our region, seven or eight species elsewhere in North America, eleven in Japan, one in Taiwan, and one widespread species ranging from Labrador to Central Asia. Species of Mitella with the calyx forming a campanulate to oblong- patelliform floral cup, oblong anthers, capitate or crescent-shaped stigmas, and a white or purple-tinged corolla of pinnately-cleft petals belong to sect. MirEL1La (subgen. Eumitella Torrey & Gray pro parte, sect. Eumitella (Torrey & Gray) Rosendahl, subgen. Spuriomitella Boissieu, subgen. Pectinella Boissieu pro parte). Primarily North American, this section in- cludes M. diphylla L., 2n = 14, a species of the eastern deciduous forest; the closely allied M. stauropetala Piper, M. trifida Graham, and M. diversi- folia Green, of the western Cordillera; and M. integripetala Boissieu and the apetalous M. doiana Ohwi, both of Japan. In our area, M. diphylla, found on moist wooded bluffs, rocky slopes, and stream banks and usually colonial, is readily distinguished by its elongate racemes of small, incon- spicuous and delicate flowers and its stratified foliage. The two suboppo- site cauline leaves are usually held horizontally about halfway up the flowering stem above the basal leaves, giving the plants a distinctive ap- pearance. Recognized variants from the typical form include f. triphylla Rosendahl (three alternate cauline leaves) and f. oppositifolia Rosendahl (paired leaves opposite). Restricted to the mountains of northwestern Georgia, western North Carolina, and eastern Tennessee in our area, M. diphylla ranges northward into New Hampshire, Vermont, and Quebec, and westward to Ontario, central Minnesota, and eastern Missouri. Section Miretzastra (Torrey & Gray) Walpers (subgen. Mitellastra (Meisner) Torrey & Gray, subgen. Mitellina Torrey & Gray, subgen. Pectinella H. Boiss. pro parte) includes those species with the calyx form- ing a strongly flattened, saucer-shaped floral cup, cordate to reniform an- thers, a greenish-yellow or purple corolla of pectinate-pinnatifid petals with the divisions spreading nearly at right angles, and pointed or strongly 2—4-lobed stigmas. Mitella caulescens Nutt., 2n = 14, M. pentandra Hook., 2n.= 14, M. Breweri A. Gray, 2n = 14, and M. ovalis Greene are endemic to the Cordillera of western North America, while ten additional species of the section are indigenous to Japan. Mitella japonica (Sieb. & Zucc.) Maxim., 2m = 14, a variable species with several varieties, is also 428 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 known from Taiwan. Mitella nuda L., 2n = 28, a wide-ranging species of the Boreal belt, is known from Central Asia to Labrador. Its amphi-Pacific distribution links the southern cordilleran centers of diversity in Japan and North America and overlaps the range of M. diphylla in northeastern North America. A species of cool, coniferous forests and mossy swamps and bogs, M. nuda is usually ecologically isolated from M. diphyilla, although puta- tive hybrids from Wisconsin have been recognized as M. Xx intermedia Bruhin ex Rydb. pro sp. (M. diphylla {. intermedia (Bruhin) Rosend.). Vegetative differences between species of Mitella consist primarily of slight differences in leaf morphology; the presence or absence, number, and positioning of cauline leaves; and the numbers of flowers comprising the simple racemes. Specific and subgeneric categories in Mitella have been based primarily on floral characters, and historically the species have been placed in several genera on the basis of stamen number and position as well as differences of ovary-floral-cup relationships. The trend toward a broad circumscription of the genus was initiated by Hooker, and Torrey & Gray (1840) united the several previously recognized segregate genera under four subgenera. De Boissieu (1899) recognized two subgenera and three sections, whereas Rydberg (1905), reverting to the narrow generic and specific concepts of Rafinesque, Meisner, and Howell, placed 18 North American species in four genera. Rosendahl (1914), however, redefined subgeneric categories and regrouped the species in two sections under Mitella. Despite differences in stamen number and position between species within a section, Rosendahl’s division appears to be a natural one, reflecting two lines of floral evolution in a genus adapted to moist humid forests. Intergeneric hybrids involving Mitella diphylla, M. nuda, and Tiarella cordifolia L., 2n = 14, have also been detected on grounds of intermediate morphology. Mitella x tiarelloides Engler, from New Brunswick, is a postulated hybrid between M. nuda and T. cordifolia, while M. « Grayana Engler combines characters of M. diphyila and T. cordifolia. The putative hybrids are apparently sterile, and at least in hybrids involving M. nuda, sterility may be due to triploidy. Although reported to be characterized by proterogyny, flowers of Mitella diphylla appear to be slightly proterandrous. Exhibiting no stamen movements, the flowers are adapted to pollination by small, probing insects. Syrphid flies and short-tongued bees are the most frequent visitors, and as they move over the flowers in search of nectar (produced by a collar of glandular tissue on the inner surface of the floral cup) they collect pollen on their mouth parts. Nectar secretion apparently is ephemeral, since once the stigmas become receptive and the ovary begins to enlarge, the glandular collar becomes shriveled and obscured. During anthesis, flower orientation in the raceme is rather variable, but as the fruits mature the racemes become decidedly secund, and the capsules become erect on their short pedicels. Dehiscence of the capsules along lines between the styles but above the two parietal placentae exposes the several to numerous seeds and creates a shallow, cup-shaped fruit with a flared, 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 429 two-beaked rim. As in Chrysosplenium, the seeds are bounced from the “splash-cup” when drops of water land in the upright capsule. The base chromosome number of the genus is seven, and, with the exception of Mitella nuda, a tetraploid with 2n = 28, all investigated species are diploids. Embryological studies have revealed variation in the mode of endosperm formation. In M. diphylla, endosperm formation cor- responds to the Helobial pattern of development, whereas in M. pentandra formation follows the ab initio Cellular sequence, and a nonaggresive haustorium comprised of four cells is formed in the chalazal chamber. Similar vegetative morphology, inflorescence structure, base chromosome numbers, and chemistry unite Mitella with Tiarella L. (x = 7), Heuchera L. (x = 7), Tolmiea Torrey & Gray (x = 7), Bensoniella Morton (Ben- sonia Abrams & Bacig.), Tellima R. Br. (x = 7), Conimitella Rydberg, and Elmera Rydberg in an apparently natural group within the Saxifrag- oideae characterized by two-carpellate, one-loculate gynoecia with parietal placentae. Only Mitella and Tiarella have Asian species, and with the exception of these two genera and Heuchera, the remaining are monotypic and endemic to the Cordillera of western North America. Savile (1954, 1961) has suggested that the genera of this assemblage have become segregated from one another through adaptations to differing ecological conditions, “They are accordingly largely distinguished [from one an- other] by seed-dispersal mechanisms or pollination mechanisms, which may mask relationships or suggest them in convergent forms.” (Savile, 1961, pp. 171, 172.) Continued cytological studies and hybridization experiments, as well as the application of other techniques will undoubtedly be of value in further elucidating generic relationships and may result in redefined generic concepts. None of the species of Mitella has reported economic value, although the plants are powerful astringents, and the roots of M. pentandra have been reported as effective antidotes for diarrhea. REFERENCES: Under family references see BAILLON; BENTHAM & Hooker; DAHLGREN, 1930; Danpy; De Borssteu; ENGLER, 1891, 1928; GRAENICHER; HAMEL, 1953; Ikuse; Jay, 1971; Li; Lussocx; Martin; MEISNER; RICKETT, pls. 93, 94; ROSENDAHL, 1905; SAVILE, 1954, 1961; ScHOENNAGEL; SKOVSTED; SMALL & RYDBERG; THOUVENIN; TorREY & GRAY; and Woop. prostrata Michx., an apparent “lost” third species from New England. | Forerste, A. F. On the relations of certain fall to spring blossoming plants. Bot. Gaz. 17: 1-8. pls. 1, 2. 1892. [Reports preformation of flower buds in M. nuda] 430 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Gray, A. Botanical notes. Bull. Torrey Bot. Club 13: 100, 101. 1886. [Sug- gests possible hybridization between Tiarella cordifolia and M. diphylla; also published as insert, Note on Tiarella cordifolia, Ibid.: 85. 1886.] Ouwi, J. Mitella of Japan. Acta Phytotax. Geobot. 1: 61-64. 1932. [Key to nd descriptions of eight species; see also, Flora of Japan, F. G. MEYER & E. H. WALKER, eds. Smithsonian Inst. ix + 1067 pp. 1965. Piper, C. V. New and noteworthy northwestern plants. II. Mitella trifida Gra- ham and its allies. Erythea 7: 159-163. 1899. RosENDAHL, C. O. A revision of the genus Mitella with a discussion of geographi- cal distribution and relationships. Bot. Jahrb. Supp. 50: 375-397. pl. 8. 1914 SaviLe, D. B. O. Splash-cup dispersal mechanism in Chrysosplenium and Mitella. Science 117: 250, 251. 1953. SAxENA, N. P. Studies in the family Saxifragaceae. V. Structure and develop- ment of the gametophytes in Mitella diphylla L. Jour. Indian Bot. Soc. 49: 38-41. 1970. 4. Tiarella Linnaeus, Sp. Pl. 1: 405. 1753; Gen. Pl. ed. 5. 190. 1754. Low, often evergreen, pubescent perennial herbs from scaly, creeping rhizomes, some producing leafy stolons. Basal leaves alternate from the rhizome and long petiolate; leaves simple [or trifoliolate], the blade palmately veined, cordate or broadly ovate in outline and usually pal- mately 3—7-lobed [or cleft], the margins dentate or crenate with broad, mucronate lobes; petioles dilated at the stipulate base, the stipules adnate to the petioles below, membranaceous above, tapering to acute or acumi- nate apices. Scapes naked or with 1-3 alternate, short-petiolate cauline leaves, like the basal leaves. Flowers perfect, short-pedicellate, in simple racemes, occasionally with paniculate branches at the base [or the inflo- rescence a panicle]; each pedicel subtended by a delicate, filiformly dis- sected bract. Sepals 5, imbricate in bud and connate at the base, forming a shallow, regular [or slightly irregular] floral cup, the calyx lobes spread- ing at anthesis, becoming reflexed in fruit. Corolla white or purplish- tinged, regular [or somewhat irregular], of 5 clawed, early deciduous petals exceeding [or slightly shorter than] the calyx lobes, inserted on the rim of the floral cup, the blades elliptical to oblanceolate [linear or subulate], the apices attenuate or often 3-toothed. Stamens 10, equal [or stamens opposite sepals longer than stamens opposite petals]; filaments persistent, slender and often clavate; anthers oval [or oblong], yellow or orange, dehiscing longitudinally. Gynoecium 2-carpellate, the carpels united below, forming a 1-loculate, superior ovary with 2 parietal placen- tae; ovules several to numerous, crassinucellar, and bitegmic; the carpels free above, unequal, and incompletely sealed along their adaxial sutures above the placentae, with the shorter of the two sometimes appressed and partially inclosed within the longer, the short styles (or only one of the two) terminated by inconspicuous, capitate stigmas. Fruit a 2-winged, valvate capsule, dehiscing between the free upper portions of the carpels, the expanded, membranaceous lamina of the carpels forming two unequal 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 431 Ficure 3. Tiarella. a-f, 7. Wherryi: a, flowering and fruiting plant, X 1/4; a" ee b, detail of raceme, X 1/2; c, flower, showing clavate anthers and superior ovary, X 5; d, longitudinal section of gynoecium showing parietal placenta, par- tially open lower carpel, and absence of stigma on upper carpel, X 8; e, mature ruit showing unequal open carpels, X 3; f, seed, X 16. g, T. cordifolia: gynoe- cium, illustrating stigmas and incompletely sealed sutures on both carpels, X 5. (sometimes subequal) wings. Seeds several, ovoid with a truncate base, the surface puncticulate, black and shining; embryo straight. Base chromo- some number 7. (Including Petalosteira Raf. and Blondia Necker ex Raf.) Lectotype species: T. cordifolia L.; see P. A. Rydberg, N. Am. Fl, 22: 117. 1905. (Name a diminutive of Greek, tiara, a high headdress, in allu- sion to the appearance of the capsules.) — FALSE MITREWORT, FOAM- FLOWER. A small genus of six, seven, or perhaps as few as three species adapted to cool, moist forests of mountainous and coastal regions of North America (variously one to four in Pacific Northwestern America, one or two in eastern North America), and one in eastern Asia (Tiarella polyphylla). Distinguished from allied genera of the Saxifragoideae by its gynoecium of two unequal and incompletely sealed carpels, the genus in North America has been divided into two geographical subgenera. The eastern American species comprise subgenus TIARELLA (subgen. Eutiarella Torrey & Gray), distinguished by simple racemose inflorescences, regular flowers with clawed petals, equal stamens, and anthers that are wider than long after dehiscence. The western American species, all included in subgen. 432 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 ANTHONEMA Nuttall ex Torrey & Gray, are characterized by paniculate inflorescences of somewhat irregular flowers with subulate petals, sta- mens of two lengths, and anthers longer than wide after dehiscence. This division of the genus appears less tenable, however, if the single Asian species, T. polyphylla D. Don, 2n = 14, ranging from the eastern Hima- layas through western and central China to Taiwan and Japan, is grouped with American species. The flowers of T. polyphylla are regular, but the petals are subulate to linear (or rarely lacking); the stamens are of one length, but anther dimensions are variable after dehiscence. In addition, like some forms of eastern Be Tiarella, its racemes often have short paniculate branches at the bas In the southeastern United "‘Sotes, two species of Tiarella? have been recognized, although specific lines have aroused controversy. Tiarella Wherryi Lakela, based on plants from southwestern North Carolina, Tennessee, and Alabama, was segregated from the wide-ranging north- ern and stoloniferous T. cordifolia L., 2n = 14, because of floral differences and a nonstoloniferous habit, yet some workers have maintained that T. cordifolia consists of both stoloniferous and nonstoloniferous forms.* Wherry (1940, 1949), while upholding T. Wherryi on the strength of floral and foliar characters, as well as a difference in blooming period, recognized nonstoloniferous forms from the Blue Ridge and Piedmont of Virginia and North Carolina as T. cordifolia var. collina Wherry. Fernald (1943), however, included all nonstoloniferous southeastern tiarellas in T. Wherryi and correlated the lack of stolons with a modified and less variable group of floral and inflorescence characters. Radford, Ahles, & Bell (1968), on the other hand, include all nonstoloniferous forms in T. cordifolia var. collina Wherry. In addition to the lack of stolons, Tiarella Wherryi differs from T. cordi- folia in its relatively broad and short capsule valves or wings with rounded tips, its tapering staminal filaments after anther disarticulation, and its sometimes purplish-tinged flowers with the lower pedicels 6-10 mm. long. In T. cordifolia the capsule wings are elongate and relatively narrow, terminating in a subacuminate tip, the staminal filaments have a blunt apex, the flowers are usually white, and the lower pedicels are 7-13 mm. long. The shorter of the two carpels in flowers of T. Wherryi lacks the small, capitate stigma present on both carpels in flowers of T. cordifolia. Wides pread from Nova Scotia and New Brunswick to Ontario and southward through the Appalachian Mountain states, Tiarella cordifolia reaches the southern limits of its range in the mountains of North and South Carolina, Tennessee, and Georgia. Occupying moist, forested slopes and woodland margins, the species ranges westward through West Virginia * A third species, Tiarella macrophylla, was described by Small (Fl. Southeast. U.S. 501. 1903) from Tryon Mountain, North Carolina. The type specimen, however, con- sists of basal leaves of Heuchera villosa Michx. and racemes of T. cordifolia (cf. Lakela, 1936). * Cf. Wherry (1940) for information on the Linnaean description of T. sth and the rouble on which it was based. This problem needs further clarification 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 433 and Pennsylvania to Ohio and Michigan and is known from a disjunct population in Wisconsin. Tiarella Wherryi is primarily a species of the Piedmont where it replaces T. cordifolia, but it also occupies habitats similar to those of T. cordifolia in the mountains of the Southeastern States. It has been impossible to determine from herbarium specimen labels, however, whether the two species occur in mixed populations or if they are restricted to different altitudinal zones in the mountains. On the Pied- mont, 7. Wherryi is known from scattered localities in Mississippi, Ala- bama, and Georgia, but the species is most frequently encountered in the Carolinas and Virginia. Problems encountered in the taxonomic treatment of southeastern tiarellas are undoubtedly partially due to variation within Tiarella cordi- folia sensu stricto, but are also related to lack of care in including stolons, when present, with pressed herbarium specimens. Attempts to delimit subspecific taxa within T. cordifolia by Lakela, who recognized two varieties (vars. tyvpica and austrina) on the basis of relative leaf dimen- sions and marginal dentation, have been largely unavailing since variation in these characters is continuous (cf. Fernald, 1943). Variants in floral features include plants with extremely small flowers and short stamens (2-3 mm. long), as f. parviflora Fernald, plants with the capsule valves subequal, as f. subaequalis Lakela, and plants with the petals neat a at the apex, as f. tridentata Lakela, Further studies of variation pattern ecology, and genetic compatibility to assess the obviously close A ineeip between 7. cordifolia and T. Wherryi are needed to substantiate the con- tinued recognition of two species in the Southeast. The three or four species traditionally recognized from northwestern North America are also closely allied to one another, and Kern (1966) has treated these taxa as comprising two subspecies and two varieties of one polymorphic species, Tiarella trifoliata L. (incl. subsp. trifoliata var. trifoliata, T. laciniata Hook. as var. laciniata (Hook.) Wheelock, and T. unifoliata Hook. (incl. T. californica Kell.) as subsp. unifoliata (Hook.) Kern). Together, these taxa which have overlapping ranges are distributed from southern coastal Alaska southward along the Pacific Coast to Santa Cruz County, California, and eastward in Oregon, Washington, and British Columbia to southwestern Alberta, northern Idaho, and western Montana. The taxa differ from each other in leaf form and marginal dissection (the floral features are stable throughout) and previously were suspected of hybridizing when sympatric. Kern documented genetic compatibility be- tween them and found no discontinuity in leaf variation. Taylor (1971a) confirmed the biological unity of the taxa chemically, but pointed out that the taxa are differentiated along an altitudinal gradient and often remain distinct in sympatric populations. He suggests that “breeding behavior and/or ecotypic distinctions are instrumental in preventing com- plete convergence.” The three investigated species, Tiarella cordifolia, T. polyphylla, and T. trifoliata sensu lato, are diploids with chromosome numbers of 2” = 14. Matsuura & Sut6é (1935), however, reported a gametic number of 9 for 434 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 plants of T. polyphylla and noted a tendency towards secondary chromo- somal associations in pollen mother-cells. Development of the ovule and female gametophyte in Tiarella cordifolia (T. Wherryi?) has been studied by Herr (1954), who found that the megaspores exhibit the T-shaped arrangement and the chalazal spore functions to give rise to an eight-nucleate embryo sac from which the mature Polygonum type sac develops. As in most species of Saxifraga, the polar nuclei fuse immediately, but in addition, the three chalazal anti- podals disintegrate prior to fertilization. Pollination ecology has not been studied in Tiarella, although one re- port lists the genus as proterogynous. Blooming from April through June, depending upon locality, the flowers of T. cordifolia and T. Wherryi ap- pear to have no obvious adaptations for self-pollination, yet nectaries that might attract pollinating insects are not evident. Kern found the western taxa to be self-compatible but lacking mechanisms to insure self-polli- nation. In all species of Tiarella the carpels are puberulent at least at the base with glandular hairs, but it is not known whether insect-attracting sub- stances are contained in the hairs. The adaptive value of the two unequal wings (carpels) of capsules of Tiarella appears to be associated with seed dispersal. Savile (1961) has described the propulsion of seeds from capsules when drops of water land on the longer of the two wings, thereby depressing that carpel and opening the valvate capsule. If this mechanism is to be effective, it would appear that the seeds must be released before the capsules become dry and mem- branaceous and the valves become stiff. Morphologically, Tiarella is unlike most genera of the Saxifragaceae in its stipulate leaves. This character, however, as well as floral and other vegetative similarities and base chromosome numbers, aligns the genus with Mitella, Heuchera, and several monotypic genera with two-carpellate, one-loculate ovaries with parietal placentation. Intergeneric hybrids in- volving Tiarella cordifolia and species of Heuchera and Mitella tend to support the postulated relationships and are discussed under these two enera. Tiarella has been claimed to be both diuretic and useful in the suppres- sion of urine and in treating gravel in the urinary tract. REFERENCES: Under family references see DANDY; ENGLER, 1891, 1928; HAMEL, 1953; Ixuse; Jay, 1971; Knutu; Lr; Raprorp, ef al.; Rickert, pl. 94; ROSENDAHL, 1905; SavILe, 1954, 1961; SCHOENNAGEL; SKOVSTED; SMALL & RYDBERG; THOUVENIN; TORREY & Gray; and Woop. Danrortu, C. H. A dimorphism in Tiarella cordifolia. Rhodora 13: 192, 193. 1911. [Notes plants in a Norway, Maine, population with either yellow or orange anthers. } a M. L. Some anomalous plants of Tiarella and Mitella, Rhodora 8: 0-92. 1906. [Report of intergeneric hybri . Tiarella cordifolia L. forma parviflora, nf. Ibid. 19: 132. 1917. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 435 . Virginia botanizing under restrictions. Jbid. 45: 445-480. pls. 783-795, 1943. [Notes on southeastern Tiarella, including a redefinition of T. Wherryi to encompass all nonstoloniferous forms, 445-449. Gray, A. Note on Tiarella cordifolia. Bull. Torrey Bot. Club 13: 85, as insert. 1886. [Also published as Botanical Notes. /bid. 13: 100, 101. 1886. Sug- gests possible hybridization between T. cordifolia and Mitella diphylia.| Herr, J. M., Jr. The development of the ovule and female gametophyte in Tiarella cordifolia. Am. Jour. Bot. 41: 333-338. 1954. [It is not specified whether the embryological materials were collected from stoloniferous or nonstoloniferous plants. | Kern, P. The genus Tiarella in western North America. Madrofio 18: 152-160. 19 LakeELA, O. The status of Tiarella macrophylla. Small. Torreya 36: 65. 1936. onograph of the genus Tiarella L. in North America. Am. Jour. Bot. 24: 344-351. 1937. Maycock, P. F. Tiarella cordifolia in Wisconsin. Rhodora 58: 309, 310. 1956. [Report of a herbarium specimen reputedly collected in Florence County some 200 miles west of the westernmost populations in Michigan. | Matsuura, H., & T. Suté. Contributions to the idiogram study in phanerogam- ous plants I. Jour. Fac. Sci. Hokkaido Univ. Bot. 5: 33-75. pls. 5-21. 1935. [T. polyphylia, n = 9.] STEARN, W. T. The Heucherellas. Gard. Chron. III. 115: 216, 217. 1944, [Ne otes sa aries the bigeneric hybrids between Heuchera and Tiarella. Tay or, R. J. Biosystematics of the genus Tiarella in the Washington Cascades. Nordliwest Sci. 45: 27-37. 1971a. Intraindividual phenolic variation in the genus Tvarella (Saxifra- gaceae); its genetic regulation and eiilies od to systematics. Taxon 20 467-472. 1b. Wherry, E. ue “ novelty in the genus Tiarella (Saxifragaceae), Not. Nat. 42. 4 pp. sy ete aes es on eastern Tiarellas. Bartonia 25: 70. 1949. Tiarella in Hartford County, Maryland. Castanea 17: 137. 1952. [Occurrence of a population of T. cordifolia.) 5. Heuchera Linnaeus, Sp. Pl. 1: 226. 1753; Gen. Pl. ed. 5. 106. 1754. Pubescent, perennial herbs, usually from stout rhizomes or woody caudi- ces [sometimes reproducing vegetatively by stolons]. Leaves mostly basal (occasionally cauline), long petiolate and stipulate, the stipules adnate to the petioles at the base; blades palmately veined, roundish-cordate or orbic- ular in outline, the margins lobed, notched, or toothed, usually with mucro- nate-tipped divisions. Flowers perfect, irregular or regular, in lax [or compact] pedunculate panicles of cymules, each cymule subtended by a small, persistent bract. Sepals 5, equal (and the flowers regular) or un- equal (and the flowers obliquely irregular), connate below, forming a 5- lobed, often gibbous, turbinate, campanulate, urceolate [or sometimes flat to saucer-shaped] floral cup. Petals 5 [or rarely absent], inserted on the inner surface of the floral cup between the calyx lobes, greenish-white, pinkish [or purplish], linear, oblanceolate, or spatulate in outline, some- times with coarsely fimbriate apices. Androecium of 5 included or exserted 436 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 stamens inserted on the inner surface of the floral cup, the filaments long and slender or short and cylindrical, tapering to the orbicular anthers. Gynoecium 2-carpellate, the carpels connate below with 2 slender filamen- — tous [or short and cylindrical] styles free above and either exserted or included, terminated by capitate stigmas; ovary adnate to the floral cup (the free portion of the floral cup above the adnate region shorter than, equal to, or longer than the adnate region), partially to almost completely inferior, 1-loculate with numerous bitegmic, crassinucellar ovules on 2 parietal placentae. Fruit a 2-beaked capsule, usually exserted from the persistent floral cup, dehiscing between the indurated styles; seeds numer- ous, small, brownish black, ovoid-ellipsoidal with smooth, echinate, or verrucose surfaces; embryo small, straight, embedded in endosperm of the ab initio Cellular type. Base chromosome number 7. (Including Oreanthus Raf, ex Ser., Oreotrys Raf., and Yamala Raf.) Type species: H. ameri- cana L. (Name commemorating Johann Heinrich von Heucher, 1677- 1774, professor of medicine in the University of Wittenberg and author of Index Plantarum Horti Medici Academiae Wittembergensis.) — ALUM- ROOT, ROCK-GERANIUM. ‘A strictly North American genus of perhaps as many as 55 species in five sections; seven spécies in the southeastern United States, two or three elsewhere in eastern North America, Heuchera Richardsonii R. Br. extend- ing across the northern Great Plains, and about 35 in western North America from southern coastal Alaska to southern Mexico. Primarily distributed in unglaciated areas, Heuchera has its greatest diversity in the semiarid regions of western North America where three of the five sections (sects. HruCcHERELLA (Torrey & Gray) Rosend. et al., HoLocHLOA (Nutt.) Rosend. e¢ al., and RHODOHEUCHERA Rosend. et al.) are endemic. A second area of diversity occurs in the Southern Appalachians where species of sects. HEUCHERA (subgen. Euheuchera Torrey & Gray, sect. Euheuchera (Torrey & Gray) Rosend, e¢ al.) and Herucuea (Torrey & Gray) Rosend. e¢ al. are concentrated. Several east-west species pairs in sect. HEUCHERA, however, establish links between the two regions. In the Southeast, Rydberg (1905) recognized 15 species of Heuchera, while Rosendahl, Butters, & Lakela (1936) reduced the number to seven but treated eight infraspecific taxa under five of the recognized species. Five of the species with nearly regular flowers and with the free, upper portions of the floral cups conspicuously shorter than or about equal to the inferior portion of the ovary belong to sect. HEuCHERA. Heuchera americana L. (H. Cortusa Michx., H. glauca Raf., H. Curtisii Torrey & Gray ex Gray, H. lancipetala Rydb., H. calycosa Small, H. hispida var. hirsuticaulis Wheelock, H. hirsuticaulis Rydb.), 2n = 14, a late spring-blooming species that grows on clayey banks and rocky forested slopes, ranges from eastern Connecticut to southern Illinois, Missouri, and eastern Oklahoma, and southward into South Carolina, northern Alabama, Mississippi, and Arkansas. A highly polymorphic species, it was divided by Rosendahl et al. into six varieties, of which five have been documented within our area. Al- 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 437 though included in sect. HEUcHERA subsect. HEUCHERA (flowers regular with the floral cup conspicuously shorter to about equal to the inferior por- tion of the ovary), H. americana intergrades over a wide area in Michigan, Indiana, Illinois, Missouri, and Arkansas with H. Richardsonii (included in sect. HERUCHEA subsect. PUBESCENTES), and introgressant forms have been recognized as vars. interior Rosend. et al. and hirsuticaulis (Wheel.) Rosend. et al. The second species of subsect. Hrucuera, H. hispida pected to occur in the northern mountains in our region. Restricted to limestone and sandstone substrata, H. hispida is also a spring-blooming species, and it was long confused with H. Richardsonii (cf. Wherry, 1933, 1934). Two of the three species included in subsect. ViLLosAE Rydb. ex Engler (flowers with the upper free portion of the floral cup much shorter than the inferior portion of the ovary and villous with long, white hairs, the leaves with acute lobes and seeds with long, sharp spines) occur in our area. An- other highly polymorphic species with pinkish flowers and a turbinate- campanulate floral cup, Heuchera villosa Michx. (H. macrorhiza Small, H. crinita Rydb.) is comprised of three varieties that intergrade with one an- other both morphologically and geographically. Occurring in acidic soils, H. villosa is distributed in the mountains of Virginia, the Carolinas, Ten- nessee, Georgia and Alabama, westward through the Ohio River Valley to the Mississippi, and into eastern Missouri. Heuchera arkansana Rydb., a species of shaded seepage slopes (usually on sandstone substrata) differs from H. villosa in its distinctly campanulate floral cup, leaves with ap- pressed fine, white hairs, and its generally shorter flowering stems. Closely allied to H. villosa var. macrorhiza (Small) Rosend. et al., H. arkansana is known from a few scattered localities in Newton, Benton, Washington, and Franklin counties, Arkansas. Subsect. MicraNntHaE Rydb. ex Engler includes Heuchera puberula Mack. & Bush and H. parviflora Bartling in our area. Heuchera puberula is a fall-blooming species that occurs on moist, shaded limestone bluffs from northern Arkansas and Missouri eastward to western Kentucky. A glandular-puberulent plant with flowers about 2 mm. long and the free portion of the floral cup about 0.3 mm. long, H. puberula is very close to H. parviflora, a more-or-less villous plant with smaller flowers (1.5 mm. long) that have an obsolescent floral cup. The typical variety of H. parviflora has white-villous petioles and is restricted to the mountains of North Carolina, while var. Rugelii (Shuttlw.) Rosend. et al. ranges from West Virginia southward through the mountains of North Carolina, Ken- tucky, and Tennessee, and westward into Indiana and Illinois. It is also known from Cherokee County, Georgia, and Etowah County, Alabama. As a subsection, the MicrANTHAE is similar to the VILLosAE, but the lobes and teeth of the leaves are rounded, and the spines of the seeds are short or obsolete. Additional species included in this subsection are western and link the two areas. Section’ Herucuea (Torrey & Gray) Rosend. e¢ a/. (plants with the 438 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 flowers moderately to very irregular and the free, upper portion of the wide, open-campanulate floral cup equal to or longer than the inferior portion of the ovary) is represented by two species, both in subsect. PUBESCENTES Rosend, e¢ al., in our area. Heuchera pubescens Pursh, a species with the styles becoming long exserted in fruit from the moder- ately gibbous floral cup is recorded from one collection in the mountains of North Carolina, while H. longiflora Rydb., a taxon with included or barely exserted styles and a strongly gibbous floral cup, is essentially restricted to our region. Outside of our area, H. pubescens occurs in Maryland, West Virginia, and Kentucky, while H. longiflora ranges from West Virginia and Kentucky, southward through the mountains of North Carolina and Tennessee and into northern Alabama. Both are plants of rocky ledges in deeply wooded regions and apparently prefer limestone substrata. Although Heuchera is easily distinguished from other genera of the Saxifragoideae by its distinctive flowers arranged in paniculate or thyrsi- form, spikelike inflorescences, few genera offer greater problems of delimitation and classification below the generic level. The early sub- generic classification of Torrey & Gray (1840), who recognized four sub- genera, has been followed by most workers, but Rosendahl, Butters, & Lakela (1936) interpreted the categories as sections and added the fifth. Rydberg (1905) placed the species in ten groups of undesignated rank that have subsequently been treated as subsections. At the specific and infraspecific levels, taxonomic problems are multi- plied not only because of variable diagnostic characters (e.g., pubescence, leaf shape, and floral structures that are subject to marked changes during anthesis) but also because of the extreme ease of gene flow between species and the production of intermediate and intergrading, often fertile hybrids. The stability of taxa within a given region appears to be a function of the strength of seasonal and ecological isolation. In the cordillera of the Pacific Northwest, the reticulate relationships between species are compli- cated further because of the topography and glacial history of the region (Calder & Savile, 1959). With the exception of one report of 2” = 28 for a Canadian population of Heuchera Richardsonii (Love & Ritchie, 1966), all chromosome num- bers recorded for the genus have been diploid, 2n = 14. Satellites have been observed in mitotic divisions in some taxa, but a more frequent phenom- enon is the occurrence of from one to four chromosomal fragments (or B chromosomes) in meiotic divisions. Jensen (1939) described irregularities in meiosis in H. americana that result in the noninclusion of one chromo- some in one of the two daughter nuclei in the first division, yet pollen ap- peared to be 100 per cent normal. The period of anthesis varies between species of Heuchera, with some blooming during the late spring and early summer, while others bloom later “Rosendahl, e¢ al. (1936, p. 1 in footnote) credit Engler (1928) with the valid publication of Rydberg’s groups as subsections. Engler a arranged the groups under the classical subgenera (sections) and crea earn indicated their rank, but he did not specifically state what rank the gro 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 439 in the summer and into the fall months. Insect pollination appears to be the general rule, and Graenicher (1907) noted a close relationship between the flowers of H. Richardsonii (as H. hispida) and a particular bee (Cal- letes aestivalis Patton). Nectar is secreted on the surface of the ovary and on the walls of the floral cup, the stigmas are receptive before the pollen is shed, and it appears that most species are adapted for cross pollination, al- though self-pollination may be expected in flowers with included anthers and styles. Hydathodes have been reported in the leaves of some species of Heu- chera, and the leaf petioles are distinctive in their three separate collateral bundles. A broad, continuous ring of fibers has been observed in the axis of investigated species, and the pith is likely to be very lacunar. Heuchera is generally close to Saxifraga anatomically, but it shows chemical similarities to Mitella and Tiarella, genera with which it shares a two-carpellate, unilocular ovary with parietal placentae, stipulate leaves, and a base chromosome number of seven. The sterile intergeneric hybrid between Heuchera and Tiarella cordifolia, < Heucherella tiarelloides (Lemoine) Wehrhahn, also substantiates the alliance of Heuchera with Tiarella. Although species of Heuchera are reputed astringents and were used as styptics in the treatment of ulcers by the Indians, current use of the plants is as garden ornamentals. Numerous garden hybrids have been developed, particularly in Europe, and most of the hybrids have as one parent H. sanguinea Engelmann, the pink-flowered coral-bells of old-fash- ioned perennial borders. REFERENCES: Under family references see BAILLON; DAHLGREN, 1930; DANDy; ENGLER, 1891, 1928; GRAENICHER; HAMEL, 1953; Jay, 1971; KnuTH; LUBBOCK ; Martin; Morr; Morvittez; Rickert, pls. 94, 95; RoBERTSON; ROSENDAHL, 1905; Savite, 1954, 1961; ScHOENNAGEL; SKovSTED; SMALL & RyDBERG; THOUVENIN; Torrey & Gray; and Woop. ANDERSON, F. W. Astringent qualities of Heuchera and Mitella. Bot. Gaz. 12: 65, 66. 1887. Arwoop, H. Coral bells. Lasca Leaves 8: 17, 18. 1958. [H. sanguinea. | Barton, W. P. C. Vegetable materia medica of the United States; or medical botany. Vol. 2. xvi+ 243 pp. pls. 26-50. Philadelphia. 1818. [H. ameri- cana, 159-164, pl. 40.] Bastin, E. S. Structure of Heuchera americana. Am. Jour. Pharm. 66: 467-473. Catper, J. A., & D. B. O. Savite. Studies in Saxifragaceae. I. The Heuchera cylindrica complex in and adjacent to British Columbia. Brittonia 11: 49- 67. 1959. Dantcren, K. V. O. Notes on the ab initio Cellular endosperm. Bot. Not. 1923: 1-24. 1923. Duncan, W. H. Preliminary reports on the flora of Georgia — 4. Notes on the distribution of flowering plants including species new to the state. Castanea 15: 145-159. 1950. [Extension of range of H. parviflora var. Rugelii.| 440 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 GAumANN, E. Studien iiber die Entwicklungsgeschichte einiger Saxifragales. Rec. Trav. Bot. Néerl. 16: 285-322. 1918/19. Grttespir, J. P. The genus Heuchera in Tennessee. Jour. Tenn. Acad. Sci. 31: 232-241. 1956. [Includes distribution maps. ] Hom, T. Medicinal plants of North America 65. Heuchera americana. Merck’s Rep. 21: 267-269. 1912.* Hooker, J. D. Heuchera sanguinea. Bot. Mag. 113: tab. 6929+ 2 pp. text. 1887 Horsrorp, F. H. Notes on see slate plants. Garden Forest 3: 240. 1890. [Garden use of H. americana. | JENSEN, H. W. On plants hs almost regular meiosis. I. Cytologia 9: 412-418. 1939, [Experiments with H. americana. | ; Lepien, F. Heuchera sanguinea Engelm., eine harte Staude fiir Blumenschnitt und Treiberei. Gartenflora 41: 617-619. pl. 1384. 1892 NIELSEN, E. L., & O. R. Younce. Observations on the distribution of Heuchera arkansana Rydberg. Am. Midl. Nat. 19: 595-597. Pace, L. Parnassia and some allied genera. Bot. Gaz. 54: 306-329. pls. 14-17. 1912. [Includes embryo-sac development in Heuchera. | Peacock, J. C., & B. L. Peacock. Further study on the tannin of Heuchera ameri ona L. Jour. Am. Pharm. Assoc. 16: 729-737. 1927.* ROSENDAHL, C. O. A new Heuchera from Missouri together with some notes on the Heuchera parviflora group. Rhodora 53: 105-109. pl. 1166. 1951. | Description is fet missouriensis as new and closely allied to H. parviflora.] os ERS, os Pct Identity of Heuchera hispida Pursh. Rhodors a5: wie 118. ; , & —. ut re oa of the genus Heuchera, Minn. Stud. Pl. Sci. 2: 1-180. 1936. Suarp, A. J. Further observations on vascular epiphytes in the Smoky Moun- tains. (Abstr.) ASB Bull. 10(2): 38. 1963. [Records H. villosa as epiphy- tic on the trunk of a large specimen of Tilia heterophylla. | STEARN, W. T, The heucherellas. Gard. Chron. 115: 216, 217. 1944 Stewart, M. G. A revision of the Heuchera rubescens group (Saxifragaceae) for the United States. Bull. So. Calif. Acad. Sci. 33: 42-49. Wuerry, E. T. The gardener’s pocketbook. ne es ae natives. 5. Alumroot, Heuchera pubescens Pursh, Natl. Hort. Mag. 9: 930. ———. Heuchera hispida Pursh rediscovered. cic: 35: 118, 119. 1933. ——. Heuchera hispida Pursh. Claytonia 1: 25-28. 4. . The genus Heuchera in Virginia. [bid. 4: 3-6. 1937. 6. Boykinia Nuttall, Jour. Acad. Nat. Sci. Phila. 7: 113. 1834, nom. cons. Leafy-stemmed, glandular-pubescent perennial herbs up to 8 dm. hi from slender, horizontal rhizomes [or thick, scaly rootstocks]. Basal leaves orbicular or reniform in outline, long petiolate and deeply 3—5(6)-cleft, the margins incised with acute teeth, each tooth terminating in a hydathode- like structure [or lobed and the margins crenate with broad, mucronate lobes], the venation palmate; petioles dilated at base, [exstipulate or] with [foliaceous or] bristlelike stipules; cauline leaves alternate, like the basal leaves but becoming short-petiolate or sessile and reduced upwards. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 441 . and compacted, becoming expanded, secund and stiffened in fruit. Calyx of 5 sepals, densely covered with capitate glandular hairs, the sepals con- nate below, forming a campanulate or turbinate floral cup, the 5 lobes erect, their apices acute. Corolla regular, of 5 white [or greenish yellow] petals inserted on the rim of the floral cup, sessile or clawed, and early deciduous (in ours, one vascular trace entering the base of the petal) [or persistent]. Stamens 5, inserted opposite the calyx lobes on the rim of the floral cup and not exceeding them in length, the anthers 2-locular, white, with an apiculate appendage above, dehiscing longitudinally and long per- sistent, but usually deciduous before maturation of the fruit. Gynoecium of 2 follicle-like carpels, connate below and adnate to the lower 1/3—1/2 of the floral cup; styles free and divergent [or incompletely connate and erect]; stigmas inconspicuous, capitate; ovary 2-loculate with numerous bitegmic, crassinucellar, anatropous ovules on axile placentae, the placen- tae restricted to the connate region of the carpels. Fruit a 2-beaked, septi- cidal capsule within the persistent urceolate floral cup; each carpel de- hiscing in the stylar region, the beaks widely spreading [or + erect]. Seeds numerous, small, black, and shining, the surface finely granulate. Endosperm present, of the a6 initio Cellular type; embryo small, straight; embryogeny corresponding to the Solanad type. Embryo sac of the Poly- gonum type. Base chromosome numbers 6 and 7. (Not Boykinia Raf., nom. rejic.; incl. Therophon Raf.; excl. Telesonix Rydb. and Peltoboykinia Hara.) Type spectes: B. aconitifolia Nutt. (Name commemorating Dr. Samuel Boyken, 1786-1848, planter, physician, and naturalist of Mill- edgeville, Georgia.) — BROOK-SAXIFRAGE. Flowers perfect, in terminal and subterminal cymose panicles, initially lax Cal As treated by Engler (1928), a small genus of nine or ten species; one in our region, four or possibly five in the Rocky Mountains, the Pacific Northwest, and California; one in Asia, Alaska, and the Yukon; and three indigenous to Japan. As defined by Rosendahl (1906) and Engler (1928), Boykinia consists of three sections, but both sectional and generic bound- aries are problematic, undoubtedly because some species referred to Boy- kinia are probably best considered as species of related genera in the Saxi- fragoideae. Boykinia aconitifolia Nutt. (Therophon aconitifolium (Nutt.) Millsp.), 2n = 12, of our area has been placed with Pacific Northwestern and Cali- fornian species in sect. BoyKIN1A, characterized by plants with divided or crenate leaves, five stamens, and gynoecia with the carpels not connate in the stylar region. Ranging in the mountains from northern Alabama and Georgia to Virginia and West Virginia, B. aconitifolia is a plant of moist, rocky stream banks and seepage slopes in wooded areas. Anthesis occurs in June and July, but in the vegetative phase plants of B. aconitifolia might easily be mistaken for plants of Geranium maculatum L, (Geraniaceae), which can occupy similar habitats. Plants from northern Alabama with a distinctly turbinate floral cup and rounded leaf lobes have been described 442 JOURNAL OF THE ARNOLD ARBORETUM [vor. 53 as Therophon turbinatum Rydb. but are probably only local variants of B. aconitifolia, The affinities of Boykinia aconitifolia appear to be with B. elata (Nutt.) Greene, 27 = 14, an extremely variable species that inhabits moist forests from middle elevations in the mountains to sea level along the coast from British Columbia south to northern California. Other western species in sect. Boyxrnia include B. rotundifolia Parry, endemic to the San Gabriel and San Bernardino Mountains of southern California, and B. major Gray, a plant that ranges from western Washington to California and northeast into Montana. Boykinia lycoctonifolia Engler, the sole representative of sect. Boykinia in Japan, has alternatively been placed in a monotypic genus, Neoboykinia Hara, because of its persistent greenish petals and an ovary that is almost completely adnate to the floral cup. The two or possibly three species historically included in sect. RENI- FOLIUM Rosend. are characterized by erect, partially united styles and reniform leaves. A tertiary relict endemic to unglaciated regions of Alaska, the Yukon, and the adjacent Siberian mainland, Boykinia Richardsonii (Hooker) Gray, 2m = 36, 84, has five stamens and white petals and ap- pears to be closely related to B. rotundifolia, a reniform-leaved species included in sect. Boykin1a. Other species of sect. RENIFoLIUM are B. Jamesui (Torrey) Engler, 2m = 28, and B. heucheriforme (Rydb.) A. Nelson, 2n = 14, which are closely allied (and perhaps conspecific). Both indigenous to the Rocky Mountains from British Columbia and Alberta south to Colorado, these alpine species are characterized by purple petals and ten stamens. Treated as members of the segregate genus Telesonix by Rydberg, plants of these taxa are probably best considered species of either Saxifraga or Telesonix, a procedure followed in most recent manuals. Two additional species from Japan, Boykinia tellimoides (Maxim.) Engl., 27 = 22, and B. Watanabei Yatabe, were included in sect. PELTO- BOYKINIA by Engler. Hara, however, has considered these species with peltate leaves, endosperm formation of the Helobial type, and yellowish petals covered with glandular trichomes as comprising a separate genus, Peltoboykinia. Other than the reported chromosome numbers and some embryological data, biological information concerning the genus is lacking. The short anther filaments might suggest that unlike most other genera of the Saxi- frageae the stamens in Boykinia do not exhibit the movements associated with proterandry during anthesis. Seed dispersal mechanisms have not been studied, although a mechanism in which the seeds are bounced out of the capsules by wind or passing animals is probably present. Studies of pollination ecology, dispersal mechanisms, genetic compatibility, and breeding mechanisms, as well as the use of other experimental approaches, are needed to determine the limits and homogeneity of the genus. Species of Boykinia are reported to be good wildflower garden plants, but other uses of the plants have not been found in the literature. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 443 REFERENCES: Under family references see DAHLGREN, 1930; DANDY; ENGLER, 1891, 1928; HAMEL, 1953; IkusE; Jay, 1971; Li; Rickert, pl. 95; ROSENDAHL, 1905; SAVILE, 1961; SkovsTeD; SMALL & RypgBerc; and Woop. Hamet, J. L. Notes préliminaires 4 une étude caryologique des Saxifragacées. I. Les chromosomes de Peltiphyllum peltatum (Torr.) Engler et de Boykinia tellimoides (Maxim.) Engler. Bull. Mus. Hist. Nat. Paris II, 20: 198-200. 1948. ARA, H. Two new genera of Saxifragaceae in Japan. Bot. Mag. Tokyo 51: 250-253. 1937. [B. lycoctonifolia segregated as Neoboykinia, and sect. Peltoboykinia raised to generic standing; notes on generic relationships of Boykinia. | Hutrén, E. Flora of Alaska and neighboring territories: A manual of the vascular plants. xxii+ 1008 pp. pls. 1-8. Stanford. 19638. [B. Richard- sonii, photograph, pl. 6.] 7. Saxifraga Linnaeus, Sp. Pl. 1: 398. 1753; Gen. Pl. ed. 5. 189. 1754. Perennial [rarely annual or biennial] scapose [or leafy-stemmed], often evergreen herbs, commonly caespitose [or mat-forming], and generally glandular-pubescent and semisucculent [becoming woody or thickened at the base in evergreen species]. Roots fibrous, often from a rhizome | many species reproducing vegetatively by stolons and bulbils in leaf axils or inflorescences]. Leaves basal [or alternate, rarely opposite and imbricate], often red-tinged; blades with pinnate or palmate venation, ovate, spatu- late, oblanceolate to lanceolate [or scalelike or peltate] ; margins entire or variously dissected, usually serrate or dentate, crenate, lobed [pin- natifid, or deeply cleft]; blade tapering gradually into the petiole or con- stricted from an obtuse, truncate, or cordate base; foliar glands (if present) superficial or on hairs; hydathodes often along the blade margin or at the apex [in some species secreting a solution of calcium bicarbon- ate]. Flowers perfect, complete [or apetalous and then 4-merous], borne in racemose cymes, bracteolate panicles [or rarely solitary]; inflorescences commonly tightly compacted, at least initially, often becoming diffuse during anthesis. Sepals 5, often reflexed and united below to form a saucer-shaped [to campanulate] floral cup. Corolla white [pinkish, purple, or yellow], regular or irregular, comprised of 5 [infrequently emarginate] clawed or sessile, sometimes yellow- [or purple-] spotted petals. Androecium of 10 stamens, obdiplostemonous; anthers dehiscing longitudinally, usually deciduous; filaments slender or often clavate, per- sistent: pollen 3-colpate, prolate or subprolate to spheroidal. Gynoecium 2-, or 3(—6)-carpellate, the carpels free or mostly connate below the middle; styles free; stigmas usually capitate and enlarged, occasionally smaller in diameter than the style, and not infrequently decurrent on the adaxial surface of the styles; ovary superior to semi-inferior and adnate to a poorly to well-defined floral cup, 1-loculate below with parietal placentae or 2- 444 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 iy: flow glandular disc at sides of ovary, X 10; m, c glandular disc (not sectioned), < 12; n, mature, dehisced capsule, x 6. ross section of ovary just above loculate with axile placentae and marginal placentae above; glandular nectar-secreting tissue sometimes present on ovary at base of styles, other- wise between ovary and the insertion of petals; ovules usually numerous, bitegmic and crassinucellar. Fruit [varying from plainly capsular and dehiscent across the top by ventral sutures of the stylar beaks (placenta- tion axile) to} follicular and dehiscent all or part of its length; stylar beaks 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 445 erect or widely spreading at maturity. Seeds numerous, small, spinulose ribbed or glandular roughened, otherwise smooth [or the testa occasionally loose and inflated or winged]; endosperm of the Helobial type, usually copious. Base chromosome numbers 8, 9, 10, 11, 12, 13, 14. (Including Aulaxis Haw., Dermasea Haw., Micranthes Haw., Spatularia Haw., Robert- sonia Haw., Hydatica Necker ex S. F. Gray, Hexaphoma Raf.) Lrcto- TYPE SPECIES: S. granulata L.; see J. K. Small, N. Am. Fl. 22: 127. 1905. (Name from Latin, saxum, a rock or stone, and frangere, to break, either in reference to the reputed medicinal value for kidney stones or the ability of many species to grow in rock niches and crevices. ) — SAXIFRAGES. A large genus of about 300 species grouped in from nine to 16 sections and widely distributed throughout the Arctic and North Temperate Zones. Towards the equator, Saxifraga gemmipara Franchet occurs in north- western Thailand, while S. kederifolia Hochst. is indigenous to Ethiopia; four additional species range from Colombia to Tierra del Fuego in the South American Andes. Primarily a montane genus in the North Temper- ate Zone, the greatest diversity and the largest number of species occur in the mountains of Europe, where most sections are well represented. Numer- ous species are circumboreal, and western North America and eastern Asia are also rich in species, although only two or three sections have good representation in these regions. Seven to ten species have been attributed to the southeastern United States; upwards of 60 are known elsewhere in North America. Numerous species have xerophytic adaptations, and the genus exhibits considerable ecological amplitude in arctic-alpine environ- ments. Saxifraga oppositifolia L., a polymorphic circumpolar species that occurs in alpine regions in the Temperate Zone, is known as far north in the Arctic (Kap Morris Jesup, Peary Land at 83°39’ N. lat.) as any species of vascular plant. The eight species in our region belong to sect. MicranTHEs (Haw.) D. Don (sect. Borophila Engler), which is separated from other sections by characters including subterranean rhizomes, somewhat inflated capsules that dehisce above the midpoint, and a distinctive reticulate pollen sexine. Engler & Irmscher (1916) further grouped our species in three of eight greges below the sectional level, but Johnson (1923), utilizing other criteria, revised the North American species of this section and placed the southeastern species in five of nine new or reestablished sections. A species of rocky outcrops in the Appalachians from Maryland to northwestern Georgia, Saxifraga Michauxii Britton (S. leucanthemifolia Michx. non LePeyr, Hexaphoma petiolaris Raf., Hydatica petiolaris (Raf.) Small, Spatularia Michauxii (Britton) Small) is distinguishable from other species by its irregular corolla with three petals clawed and larger than the other two, and by its ribbed spinulose seeds, Placed in sect. ARABISA Tausch by Johnson, the plants have multiseriate trichomes, superior ovaries, and elliptic capsules. In addition, the placentation is unusual, the placentae being restricted to the upper, unfused portions of the carpels and appearing marginal (cf. Fic. 4,e). 446 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Saxifraga micranthidifolia (Haw.) Steud. (Robertsonia micranthidi- folia Haw., S. erosa Pursh, Aulaxis micranthidifolia (Haw.) Haw., Mi- cranthes micranthidifolia (Haw.) Small), 2n = 22, restricted to moist seep- age slopes and the banks of cold mountain streams from northwestern Georgia to northern New Jersey, resembles S. Michauxii in habit, and the two species were placed in grex Stellares by Engler. Saxifraga micranthidi- folia is usually a larger plant than S. Michauxii and also differs in its regular corolla, clavate anther filaments, prominent disc at the base of the ovary, and smooth seeds. On the basis of its uniseriate trichomes and superior ovary, Johnson placed S. micranthidifolia in the monotypic sect. AULAXIS (Haw.) Johnson. Other species in grex Stellares, allied to our south- eastern species by Engler & Irmscher, are wide ranging in Pacific North America, Asia, and Europe. The central species of the grex, S. stellaris L., is highly polymorphic and circumboreal. Known in our region only from two disjunct populations in Watauga County, North Carolina, our plants of Saxifraga pensylvanica L. (Mi- cranthes pensylvanica (L.) Haw.), 2n = 56, 84, 112, are referable to subsp. pensylvanica (subsp. eupensylvanica Burns), 2 = 56. The only represen- tative of grex Integrifoliae Engler & Irmscher in the Southeast, this poly- morphic species is wide-ranging from Maine and Ontario to Minnesota southward to Missouri and Virginia. A characteristic species of bogs, low meadows and seepage slopes, it was placed by Johnson in the re-established sect. MicranTHEs (Haw.) Tausch non D. Don, distinguished by ascend- ing sepals and distinct carpels that are enveloped by a glandular disc or nectary. Grex Integrifoliae, as defined by Engler & Irmscher, is wholly North American, and the several midwestern and cordilleran species allied to S. pensylvanica constitute a well-defined subgrouping. The majority of our species was included in grex Nivali-virginiensis by Engler & Irmscher, and, with one exception, Johnson placed the same taxa in sect. DermAsEA (Haw.) Johnson. These species, as a group, are char- acterized by uniseriate hairs, subinferior two-loculate ovaries, shallow floral cups, and poorly developed glandular discs. Saxifraga virginiensis Michx. (Dermasea pilosa Haw., Micranthes virginiensis (Michx.) Small), 2n = 28, a plant of dryish to moist woodlands, ranges from New Bruns- wick to Minnesota, south to Georgia, Alabama and Mississippi. Primarily a species of the Piedmont in our area, S. virginiensis flowers early in the spring and has an initially compacted inflorescence that usually becomes lax and diffuse during flowering, ovaries partially adnate to the floral cup, and sepals that are ascending in fruit, It appears to be closely allied to S. texana Buckl. (S. Reevesii Cory, Micranthes texana (Buckl.) Small), which differs in its rhomboidal inflorescence, three- (or four-) carpellate ovary, and basal leaves with the margins entire and undulating rather than crenate, lobed, or shallowly toothed as in S. virginiensis. Segregated from other species and placed in the monotypic sect. TrIcarPuM by John- son, S. texana grows in acidic soils in sandy bottomlands, rocky glades, and on sandstone outcroppings from southwestern Missouri and adjacent Kansas southward through Arkansas, Oklahoma, and northeastern Texas. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 447 In Missouri, it is sometimes associated with the rare Geocarpum minimum Mack. (Caryophyllaceae). Saxifraga Palmeri Bush (S. virginiensis var. subintegra Goodman) is also closely allied with S. virginiensis. Endemic to western Arkansas and eastern Oklahoma (outside the range of S. virginiensis), S. Palmeri inhabits dryish rocky woodlands that are usually underlain by standstone. It differs from S. virginiensis in its entire or subentire leaves and from both that species and S. texana, in its glabrous or glabrate pedicels and flower- ing scapes. Although the ranges of S. Palmeri and S. texana overlap and seasonal isolation between them is incomplete, both flowering in spring, hybrids have not been reported. Crossing experiments and cytological studies within this group might help to clarify relationships. Saxifraga caroliniana A. Gray (Micranthes caroliniana (A. Gray) Small, M. Grayana (Britt.) Small, Saxifraga Grayana Britt.) and S. Careyana A. Gray (Micranthes Careyana (A. Gray) Small) are also phenetically very similar, and it has been suggested that they may be conspecific. Unlike other southeastern species of this section in their clawed petals, they differ from one another in several details. Petals of S. caroliniana are yellow- spotted, the anther filaments are clavate, the sepals are strongly reflexed in fruit, and mature capsules are ca. 4-5 mm. long, while in S. Careyana the petals lack spots, the filaments are filiform, the sepals are more or less spreading, and the capsules are 2.5-3 mm. long. Plants from Tennessee with capsules up to 4.5 mm. long have been recognized as S. tennesseensis Small, but are considered by Lord as local variants of S. Careyana. Both S. Careyana and S. caroliniana are rare, infrequently collected plants that are restricted to moist seepage slopes and rocky stream banks in the mountains of southwestern Virginia, North Carolina, and eastern Tennes- see. Saxifraga caroliniana has also been reported from the mountains of Kentucky. Crossing experiments and cytological studies of plants of these taxa are needed. Our Southeastern species of grex Nivali-virginiensis appear to be allied with numerous species in the Cordillera of western North America where several species appear to be vicariads of our species. Additional members of this grex are circumboreal, Eurasian, or restricted to the Himalayan region. The initial broad generic concept in Saxifraga established by Linnaeus was delimited more narrowly by Haworth (1812, 1821), who allocated the species among 16 genera. More recently, Small (1905) re-established several of Haworth’s genera, among others, placing the North American species of Saxifraga in 13 segregate genera. The broad Linnaean inter- pretation of the genus advocated by Seringe (1830), Hooker (1865), Engler (1872, 1891), and Engler & Irmscher (1916, 1919) has been widely accepted and is strengthened by the relatively uniform floral morphology throughout the genus. Subgeneric categories of Saxifraga have been problematic. Huber has revised sectional nomenclature and realigned the species into 14 sections, one of which, DiscocyNE Sternb., was treated as a separate genus (Zahl- brucknera Reichenb.) by Engler. Studies of pollen fine-structure have 448 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 mainly reinforced Engler’s sectional alignment, even though “. . . many of his [15] sections . . . are almost impossible to define or diagnose honestly” (Ferguson & Webb, p. 309). Chromosome numbers reported for species of Saxifraga range from 2n = 16 to 2m = 120, and an aneuploid series from 7 = 8 to n = 14 constitutes base numbers for the genus. Several sections have several base numbers, including sect. MicRANTHES in which x = 8, 10, 11, and 14. Advanced sections (sects. AIZOONIA, TRACHYPHYLLUM, PoRPHYRION, and XAN- THIZOON) generally have higher base numbers, but not necessarily the highest diploid numbers, since polyploidy has been frequent in most groups. Reports of hybrids in the genus are frequent, particularly those involv- ing species of horticultural interest, and interspecific hybridization and the formation of fertile (or only partially sterile) progeny that may become involved in the synthesis of secondary and tertiary hybrids has been documented (Webb, 1950) in the caespitosa group of sect. SAXIFRAGA (sect. Dactyloides Tausch.). Hybrids between S. rosacea Moench and S. granulata L. have resulted in the allotetraploid S. potternensis through nonreductional divisions at meiosis in F, generation plants (Mars- den-Jones & Turrill, 1930; Whyte, 1930). Observations on insect visitors to flowers of Saxifraga indicate that the often yellow- or purple-spotted petals and the nectar secreting region at the base of the styles or between the ovary and petals attract a predominance of dipterous pollinators. Although some species are homogamous and others are proterogynous, most species are proterandrous. In proterandrous flowers, the stamens are initially horizontal, but first the antisepalous and then the antipetalous stamens bend upward over the styles, and the an- thers dehisce. Soon after the pollen is shed, the anthers are usually deciduous. Development of the styles and stigmas is usually completed after the pollen is shed, and, in the “mossy” saxifrages, the stigmas become receptive six or seven days after the onset of anthesis (Webb, 1950). Both self-compatibility and self-fertility, however, have been recorded in the genus, although compatibility systems have not been studied sufficiently. In arctic regions species of Saxi fraga initiate flower primordia during the growing season before that in which anthesis occurs (Sgrensen, 1941), and Levin (1960) has noted well-developed flower buds in mid-November in New Jersey populations of S. virginiensis. In advanced flower primordia pollen tetrads may be developed before the onset of dormancy, while in other less advanced buds pollen mother cells that over-winter in some arrested stage of meiosis may be present. Meiotic abnormalities and re- duced pollen fertility in several species suggest a history of hybridization, particularly in polyploid taxa, but may result from environmental factors that disrupt microsporogenesis. Seed dispersal from dehisced capsules in Saxifraga is effected by short- period vibrations, which cause the seeds to bounce out of the capsules and to be carried short distances by the wind (Savile, 1961). The small seeds of S. virginiensis germinate in the spring on litter-free substrata, and it 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 449 has been suggested by Levin that this is a species of early stages of succes- sion. Aspects of megasporogenesis, embryo sac development, and embryo- geny are relatively uniform in investigated species, although both uniteg- mic and bitegmic ovules and linear and T-shaped megaspore tetrads have been reported. Twin embryo sacs have also been recorded, and two radicles from a single seed have been noted in other species. Endosperm formation usually coincides with the Helobial pattern, although some variation is evident in the formation of so-called “basal” endosperm. At maturity, the small, ee pe embryo usually lies surrounded by from three to five cell layers of endospe Vegetative reproduction by oie vivipary, stolons, and rhizomes has been used to some extent as a criterion in delimiting sections. In arctic and boreal species vegetative reproduction sometimes completely replaces sexual reproduction in some populations, but agamospermous seed produc- tion has not been documented in the genus The petioles of most species of Saxifraga are anatomically distinctive. One or more (usually three) concentric or hemiconcentric vascular bundles, each with an endodermis, are present. Medullary bundles have been re- corded in stems of S. pensylvanica and species of sect. A1zoonta, As in species of Parnassia and Lepuropetalon, species of sect. CYMBALARIA develop tanniniferous cells in the epidermis, but the ontogeny of these deviates from the developmental pattern in the other genera. The stomata, generally of the ranunculaceous type, have been considered of diagnostic value, since species groups can be defined by stomatal distribution pat- terns. Trichome types that vary from uniseriate or biseriate to multiseriate are also of taxonomic significance. Hydathodes along leaf margins and at sepal and leaf apices are also relatively common in the genus, and three types have been described. Termed chalk glands when their secretions con- tain calcium bicarbonate, the hydathodes exude a concentration of calcium that is apparently determined by soil chemistry (Kurt, 1929). Schmidt (1930) has observed that in certain European species guttation from hyda- thodes occurs only before the onset of anthesis and that the amount of fluid secreted is determined by a combination of light, temperature, and humidity factors. Pollen structure in the genus is heterogeneous, but the recent survey of pollen fine-structure by Ferguson & Webb (1970) revealed four distinct types of sexine pattern that aid in the division of the genus into sections and that should prove helpful in solving taxonomic and paleoecological problems. Section MicrANTHES, to which our species belong, is distinguish- able from other sections by a reticulate pollen sexine, and pollen from Saxi- fraga micranthidifolia is the smallest (polar axis 15-16.5 »m) noted in the survey. Pollen size, however, does not appear to be correlated with chro- mosome number (Sokolovskaia, 1958). Determination of the generic relationships of Saxifraga is dependent, in part, on the definition of generic bounds. Aside from the several segregate genera that rarely remain distinctive when considered in the over-all view 450 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 of the genus, Saxifraga appears to be most closely allied to the Asian genus Bergenia Moench, to Saxifragella Engler (two species in Tierra del Fuego), to Leptarrhena R. Br. (monotypic, ranging from Alaska to Washington), and to Heuchera. Because of the great species diversity and sectional representation, the mountains of Europe have been considered a primary center of evolution for the genus. The least specialized species, however, are probably in Japan and neighboring regions, and, judging from the great development of subfam. Saxifragoideae in western North America, Savile (1961) has postulated that Sexifraga had its origin in a primitive East Asian immigrant group that through a series of Cordilleran radiations into North America also gave rise to the several endemic North American genera. Many species of Saxifraga are cultivated in rock and alpine gardens, and numerous hybrids have been produced both spontaneously and through breeding programs. The easily hybridized species of Bergenia are also oc- casionally cultivated in rock gardens and may be confused as species of Saxifraga. Saxifraga stolonifera Meerb. (S. sarmentosa Schreder in Ellis), a native of China and Japan, is the old-fashioned pot plant known as straw- berry begonia, strawberry geranium, or mother-of-thousands. REFERENCEs: Under family references see BAILLON; BENTHAM & Hooker; Core; DAHL- GREN, 1930; Danpy; De Botssiev; ENGLER, 1891, 1928; ERDTMAN, 1952; Geutus: GRAENICHER: HAMEL, 1953; Huser; Ikuse; Jay, 1971; JONSSON; Knut; KUGLER, 195 5. 191. ia: Live & Love . LUBBOCK: Martin; Morr; Morvittez: Mutter ; RICKETT, pls. 93, 94; ROSENDAHL, 1905; SAVILE, 1954, 1961; SCHOENNAGEL: SEIDEL; Skovstep; SMALL & RypBERG; STEINER & Lérrier; THOUVENIN; WARMING; and Woop ALEXANDER, E. J. Micranthes micranthidifolia: Mountain-lettuce. Addisonia 18:1, 2. ph §77.1933. BeamisH, K. I. Studies of meiosis in the genus Saxifraga of the Pacific North- west. Canad. Jour. Bot. 39: 567-580. 1961. ————. A Pacific coast saxifrage with 10 pairs of chromosomes: meiosis, de- velopment of the female gametophyte, and seed production. Jbid. 45: 1797- 1801. 1967. & S. C. Lin. Fertilization and seed woes in Saxifraga integrifolia Hook. Ibid. ye 861-865. pls. 1 & 2. 1965. BRITTEN, J. Notes on Saxifraga. Jour. ik London 37: 66-70. 1899. Burns, G. W. The taxonomy and cytology of Saxifraga pensylvanica L. and related forms. Am. Midl. Nat. 28: 127-160. 1942 . The type of Saxifraga pensylvanica L., ssp. interior’ ssp. nov. Ibid. 30: 524. 1943. Busu, B. F. The Missouri saxifrages. Missouri Bot. Gard. Rep. 20: 138-140. 1909, igeice notes on S. pensylvanica, S. virginiensis, and S. texana.] species of Saxifraga. Am. Midl. Nat. 11: 213-235. 1928. [A veeatel review of Johnson’s 1923 revision, copious notes on distribution, and description of S. Palmeri as new.] 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 451 Catper, J. A., & D. B. O. SAvite. Studies in Saxifragaceae II. guint sect. Trackypkoliem in North America. Brittonia 11: 228-249. 19 Studies in Saxifragaceae III. Saxifraga geda iae and van and North American subspecies of S. punctata. Canad. Jour. Bot. 38: 409-435. 1960. CasteLt1, M. R. Flowering behavior in a uniform garden of wide-ranging, spring-blooming, woodland herbs of the eastern United States. Castanea 35: 260-277. 1970. [The rust Puccinia pruni-spinosa Pers. noted on plants of S. virginiensis; plants of S. virginiensis from Vermont, Massachusetts, West Virginia, and ir showed evidence of distinct race formation. ] CHAPMAN, M. The ovule and oe sac of Saxifraga virginiensis. Am. Jour. Bot. 20: 151- 158, pl. 4. Don, D. A enh hag of ont setishts Saxifraga. Trans. Linn. Soc. London 13: 341-452. 182 DUNCAN ‘inal? Notes on the stamens of seca Jour. Linn. Soc. Bot. 11 . 1871. [Communicated “id J. E. BAKER; notes on the movement and sed nit dehiscence of anther. ENcteErR, A. Beitrage zur Naturgeschic des Genus Saxifraga L. Linnaea 35: 1-124. Chart+ maps 1 & 1867. Index criticus specierum ts synonymorum generis Saxifraga L. 44 pp. ae. 1869. ———. Monographie der Gattung Saxifraga L. mit besonderer Berucksichtigung der eg ac Verhiltnisse. iv-+ 292 pp. Map 1. Breslau. 1872. yt pa ecg Il. Pflanzenreich IV. 117(Heft 69) (Pars ientcaicy: 1-47. 19 Saxifraga — I. Dh Arten von Saxifraga L. sect. Hirculus (Haw.) Tausch. § Hirculoidea Engl. et Irmsch. Pflanzenareale 1: [3] + Maps 1-3. 26. . Irmscuer. Revision von Saxifraga sect. Hirculus und neue Arten anderer Rsarmenits Bot. Jahrb. 48: 565-610. 1912. Saxifragaceae — Saxifraga I. — IV. 117(Heft 67): 1-451. 1916; II. [bid. (Heft 69): 449-709. Everett, T. H. Saxifraga tennesseensis. Jour. N.Y. Bot. et 36: 122-124. 1935. Faecrt, K., & L. vAN DER Pryt. The principles of pollination ecology. ix + 248 pp. Oxford. 1966. [Pollination-ecology case history of S. aizoides L., 155,456. Fencvsow, I. K., & D. A. Wess. Pollen morphology in the genus Saxifraga and s taxonomic significance. Jour. Linn. Soc. Bot. 63: 295-311. pls. 1-7. on GALLgE, O. Saxifragaceae. 2. The biological <-sobeagied of the Arctic species of Saxifraga. Medd. Grgnl. 36: 237-294. Genty, P., & BoucHarp. Saxifrages nis nouveaux. Bull. Soc. Bot. France 94: 234-238. 1947. Goopman, G. J. A new variety in Saxifraga. Rhodora 52: 183. 1950. [De- scription of S. virginiensis var. subintegra as new Gray, A. Notes on some North American specie at Saxifraga. Proc. Am. Acad. ae Sci. 20: rae 1885. [Notes confusion, sees S. nivalis and S. virginiensis; S. texana considered a form of S. virginie GUNTHAaRT, ve ‘Beitriige zur Bliithenbiologie der Cruciferen, i ceestdnctes und 452 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 53 der Gattung Saxifraga. Bibliot. Bot. 11(Heft 58): ix+97 pp. pls. 1-11. 1902. [Includes observations on the floral ecology of 58 species of Saxi- raga. Harmsen, L. Studies on the embryology and cytology of Saxifraga. Medd. Grgnl. 125(4): 1-15. 1939. HawortH, A. H. Miscellanea naturalia, sive dissertationes variae ad historiam naturalem spectantes. xv +189 pp. London. 1803, [“Dissertatio quarta; containing a new arrangement of the genus Saxifraga . . .” 145-168. ] Synopsis plantarum succulentarum. viii+ 334 pp. London. 1812. [Catalog of succulent plants cultivated in and around London; establish- ment of the segregate genera Micranthes, Robertsonia, and Miscopetalum, 320-323; enumeration of the species of Saxifraga, 324-326. | ————.. Saxifragéarum enumeratio. xx + 207 pp. London. 1821. Horm, T. Contributions to the knowledge of the germination of some North American plants. Mem. Torrey Bot. Club 2: 57-108. pls. 5-19. 1891. [S. virginiensis, 70, 71.] Hornrprook, M. Hybrid saxifrages. Jour. Roy. Hort. Soc. 51: 49-58. pls. 22- 34 INGWERSEN, W. E. T. Saxifraga. In: F. J. CHITTENDEN, ed. Roy. Hort. Soc. Dict. Gard. ed. 2. 4: 1877-1894. 1956. [See also, 505, 506 in SyncE, P. M., ed. Supplement, ed. 2. 1969.] Irvinc, W. The newer saxifrage hybrids: a short descriptive list. New FI. Silva 1: 67-70. figs. xvi & xvii. 1928. Jounson, A. M. A revision of the North American species of the section Bora- phila Engler of the genus Saxifraga (Tourn.) L. Univ. Minn. Stud. Biol. Sci. 4: 1-109, 1923. [Published simultaneously as Minn. Stud. Pl. Sci. 1: 1-109. 1923.] ———. Cryptomorpha, a new section of Saxifraga. Am. Jour. Bot. 10: 294-296. 1923. [S. eriophora from California, considered to be closely related to S. virginiensis, removed to a monotypic section. | . Studies in Saxifraga I. Section Hydatica. Ibid. 14: 323-326. pls. 40-47. 1927; II. Teratological phenomena in certain North American species of Saxifraga, Ibid. 18: 797-802. pls. 53, 54. 1931; III. Saxifraga section Mi- cranthes. Ibid, 21: 109-112. 1934, Kenoyer, L. A. Notes on variation in Micranthes texana. Proc. Iowa Acad. Sci. 21: 123, 124. pl. 17. 1914. [Variation in carpel number in flowers from a Kansas population; up to six carpels note ad Ktoprer, K. Beitrage zur floralen Morphogenese und Histogenese der Saxi- fragaceae 4. Die Bliitenentwicklung einiger Saxifraga-Arten. Flora B. 159: 347-365. 1970. KNABEN, G. Further studies of the life cycle of Norwegian species of the genus Saxifraga. (In Norwegian; English summary.) Blyttia 19: 148-157. 1961. Kurt, J. Uber die Hydathoden der Saxifrageae. Beih. Bot. Centralbl. 46: 203- 246. pls. 1, 2. 1929. Levin, M. H. Studies on the ecological life history of Saxifraga virginiensis. Bull. Torrey Bot. Club 87: 348-354, 1960. Lorp, L. P. The genus Saxifraga L. in the Southern Appalachians. Diss. Abstr. 21(7): 1726. 1961. Marsben-Jones, E. M., & W. B. Turrity. The history of a tetraploid saxifrage. Jour. Genet. 23: 83-92. pls. 3-6. 1930. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 453 Pucstey, H. W. The eer abana saxifrages. Jour. Linn. Soc. Bot. 50: 267-289. pls. 6, 7. Ranpuawa, A. S., & K. I. Anco 4 Sexual reproduction in Saxifraga ferruginea Graham. Syesis 1: 147-156. 1968 [1969]. SAXENA, N. P. om embryo sacs in Saxifraga diversifolia. Sci. Cult. 29: 614, 615. ScHmopt, H. oe Funktion der Hydathoden von Saxifraga. Planta 10: 314-344. 1930 Scutruorr, P. N. Zur Zytologie von Saxifraga. Jahrb. Wiss. Bot. 64: 443-499. pl. 1, 1925. SERINGE, N. C, Saxifraga. DC. Prodr. 4: 17-47. 1830. SmitH, H. Saxifraga of the Himalaya I. Section Kabschia. Bull. Brit. Mus. Bot. 2: 83-129. 1958; II. Some new species. Jbid. 229-260. pls. 13-21. SokoLovsKatA, A, P. On the correlation between the number of chromosomes and the size of pollen grains in the Arctic species of come aaa and Saxifragaceae. (In Russian.) Bot. Zhur. 43: 1146-1155. 1958. S@RENSEN, T. Temperature relations and phenology of the northeast Greenland flowering plants. Medd. Grgnl. 125: 1-305. pls. 1-15. 1941. SovugEcEs, R. Embryogénie des Saxifragacées. Développement de l’embryon chez le Saxifraga granulata, Compt. Rend. Acad. Sci. Paris 202: 240-242. 1936. STEYERMARK, J. A. The taxonomic status of Saxifraga Palmeri. Brittonia 11: 71-77. Temesy, E. Der Formenkreis von Saxifraga stellaris Linné. Phyton Austria 7: 40-141. maps 1-3. 1957. Wess, D. A. Biological flora of the British Isles: Saxifraga L. (section Dacty- loides Tausch). Jour. Ecol. 38: 185-213. 1950. Hybridization and variation in the robertsonian saxifrages. Notes Bot. School ae Coll. 5(2): 9-21. 1951. [Reprinted from Proc. Irish Acad. B.i33:8 1950. Wauvyte, R. 0. ‘Suuiey and floral abnormality in the tetraploid Saxifraga potter- ensis, Jour. Genet. 23: 93-121. 1930. Wiccins, I. L. Development of the ovule and megagametophyte in Saxifraga hieracifolia. Am. Jour. Bot. 46: 692-697. 1959. 8. Chrysosplenium Linnaeus, Sp. Pl. 1: 398. 1753; Gen. Pl. ed. 5. 189. 1754. Small, fibrous-rooted, sometimes semi-aquatic perennial herbs, decum- bent and creeping [or caespitose]; stems often becoming stoloniferous [or rhizomatous, sometimes producing bulbils]. Leaves opposite [or alter- nate], subsessile to distinctly petiolate; basal and cauline leaves isophyl- lous [or heterophyllous], blades simple, the margins crenate [or toothed], rarely entire. Flowers perfect, regular, inconspicuous, greenish-yellow [white or purplish], [terminal and solitary or] arranged in loose [or com- pact], often leafy dichasial cymes. Calyx of 4 spreading [or ascending] sepals, connate below and forming a campanulate [or saucer-shaped | floral cup; one pair of sepals larger than the other pair and overlapping it in the bud. Petals absent. Stamens (7) 8 [4 or 10], corresponding to the num- ber of disc lobes (if the disc is developed) ; filaments inserted on the mar- 454 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 FIGuRE 5. ar ep Sat 6 a-k, C. americanum: a, habit, X 1/2; b, flower with subtending leaf, < 6; c, flower in early stage of anthesis showing large, lobed disc, X 12; d, de tail of dehisced anthers, X e, capsule seen from above before dehiscence — note shriveled ac. xX 6; f,.s pile herenien ac cross section of Fag a Bart Sasi placentae, X 6; g, si ide view of two-beaked fo age : midiagrammatic vertical section of undehisced capsule, X 6; e- hisced espla iatie ae “eae e from above, X 6; j, seed, ok snide: matic longitudinal section of seed with cee ued 3 in endosperm, Xx 25. gin of the disc [or floral cup]; anthers red [or yellow], 2-loculate, dehisc- ing laterally; pollen 3-colporoidate, subprolate. Gynoecium of 2 carpels; styles free; ovary [superior to] inferior, + adnate to the floral cup [or free], surrounded by and partially embedded in the disc, uniloculate, with several anatropous ovules on 2 parietal placentae. Disc nectar-secreting, well developed [or absent], often lobed and colored purplish [or greenish-yel- low], becoming shriveled and obscured in fruit by the enlarging ovary. Fruit a subinferior capsule with 2 short, divaricate beaks, dehiscing between the styles along lines above the 2 placentae, flaring, and appearing cup shaped. Seeds numerous and small, oblong [to globose] with a distinct keel on one side, reddish brown [or black]: ; surface [smooth, ridged, or] verrucose, usually with papillae; endosperm of the Helobial type, copious, surround- ing the small, straight embryo. Base chromosome number 6 or 7. LECTO- TYPE SPECIES: C. oppositifolium L.; see J. K. Small & P. A. Rydberg, N. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 455 Am. Fl. 22: 83. 1905. (Name from Greek, chrysos, gold, and splen, spleen, in reference to the plant’s alleged medicinal attributes.) — GoLtpEN SAXIFRAGE A genus of about 55 species widely distributed in the Arctic and Temper- ate zones of the Northern Hemisphere and two species disjunct in the Southern Hemisphere, occurring in the Andes of Chile and in Patagonia and Tierra del Fuego. The greatest number of species occurs in eastern Asia where about 27 are indigenous to the Sino-Himalayan region and 10 are native in the Japanese flora. Chrysosplenium americanum Schwein- itz ex Hooker, 2m = 24, is the sole representative of the genus in our region, but five additional species are known from North America. Re- ported from scattered localities in the mountains of Georgia, North Caro- lina, and adjacent eastern Tennessee, C. americanum is a plant of moist stream beds, stream banks, and wet forests, a habitat to which all species of the genus appear to be restricted. Sometimes semiaquatic, plants of C. americanum often form extensive populations in the shallow waters of brooks and streams and at first glance might be mistaken for watercress, Nasturtium officinale (Rorippa Nasturtium-aquaticum, Cruciferae). Out- side our area, C. americanum is known from stations northward along the Appalachians into New England, Nova Scotia, and the Gaspé Peninsula, westward through Quebec to Saskatchewan and southward around the Great Lakes in Minnesota, Wisconsin, Illinois, Indiana, Michigan, and Ohio. The affinities of Chrysosplenium americanum do not appear to be with the only other opposite-leaved species in North America, C. glechomi- folium Nutt. ex Torr. & Gray, indigenous to the Pacific Coast States and British Columbia, but instead with C. oppositifolium L., widespread in Europe, and with C. ramosum Maxim., known from Hokkaido and the mountains of Korea north to the mountains of the Amur region. According to Packer (1963), the alternate-leaved North American species include C. tetrandrum (Lund) T. Fries, a circumpolar plant also known from dis- junct populations in the Colorado Rockies; C. Rosendahlii Packer, endem- ic to the Canadian Arctic Archipelago and the adjacent mainland; C. iowense Rydb., originally thought to be restricted to the driftless area of northeastern Iowa but later shown to be more widely distributed in the plains of southwestern Manitoba, central Saskatchewan, and Alberta, north to the vicinity of Great Slave Lake; and C. Wrightii Franchet & Savatier, known from Alaska and Herald Island off the arctic coast of Siberia. The taxonomy of C. tetrandrum, C. Rosendahlii, and C. iowense has been complicated inasmuch as the species form a complex that has close affinities with the Old World C. alternifolium L. Packer (1963), who has clarified this situation, says that biosystematic investigations are still neces- sary within the broadly defined species, C. alternifolium. The subject of three monographs, Chrysosplenium has been divided into two subgenera and several series by Maximowicz (1881); two sec- tions by Franchet (1890, 1891), and seventeen series by Hara (1957). 456 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Hara concluded that the subgenera Chrysosplenium (subgen. Gamosplen- tum Maxim.) and Dialysplenium Maxim. and the sections CHRYSOSPLEN- tuM (sect. Oppositifolia Franch.) and ALTERNIFOLIA Franch. are un- natural and polyphyletic assemblages based on variable floral and foliar characters, respectively. Franchet’s division of the genus into alternate- leaved and opposite-leaved species, although not formally recognized by Hara, is nonetheless employed by him and seems to be of value when dealing with the species on a world-wide basis. Furthermore, leaf arrange- ment has been correlated with two types of stomatal development (Moreau, 1971). Collectors should make notes from unpressed material, since the flowers are small and delicate, and diagnostic floral features are often obscured in the preparation of herbarium specimens. Little biological information is recorded about the species of Chryso- splenium. Chromosome numbers are known for only eight species, yet cytologically known taxa include all the North American species except Chrysosplenium glechomifolium. A wide range of numbers is evident: C. americanum, C. Wrightii, C. tetrandrum, and C. nepalense D. Don have reported numbers of 2m = 24, while other counts include 2m = 22 for C. Grayanum Maxim., 2n = 42 for C. Oppositifolium, 2n = 48 for C. alterni- folium, 2n = 96 for C. Rosendahlii, and 2n = ca. 120 for C. iowense. Although Hamel (1953) notes that the chromosomes are extremely small and difficult to distinguish from one another, he suggests on the basis of chromosome morphology that C. alternifolium is an autotetraploid. The few recorded observations suggest that the reproductive biology of Chrysosplenium is varied and probably subject to environmental and seasonal modification. In Greenland, Sgrensen (1941) observed develop- ment of flower primordia and pollen formation in the fall in plants of C. tetrandrum. Anthesis occurs in early spring, and the flowers have been variously described as homogamous, proterogynous, or proterandrous. The The seed dispersal mechanism in Chrysosplenium is related to the parietal position of the placentae. In C. americanum the fully dehisced capsule has the shape of a deep cup with a flared rim. The capsule opens along lines between the styles down to a point above the two placentae, thereby exposing the seeds on the wall below, and forming a “splashcup” that resembles the gemma cups of Marchantia. In a laboratory experi- 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 457 ment, Savile (1953, 1961) found that a drop of water landing in the cup caused the ripened seeds to be thrown horizontally up to a distance of 16 inches. In nature the mature capsules are held upright, positioned, as it were, to catch falling drops of rain; and as Savile points out, “Chryso- splenium is well suited to this dispersal mechanism since it grows frequently in the splash from waterfalls, below moist cliffs, or in swampy woods where even a drizzling rain will fall from the trees in large drops.” The seeds, which offer diagnostic characters due to their sculptured surfaces, have been germinated on moist filter paper after dormancy has been broken by chilling. The affinities of C hrysosplenium are not clearly evident within the Saxi- sistently been placed in the Saxifrageae DC., subtribe Saxifraginae Engler. The presence of hydathodes in leaves of some species also tends to support this placement, but chemically Chrysosplenium is remote from other genera of the subtribe (Jay, 1970). No reports of economic use of Chrysosplenium have been located, but the plants have been used as salad greens and might constitute an emer- gency survival food. REFERENCES: Under family references see BAILLON; DAHLGREN, 1930; DANDy; ENGLER, 1891, 1928; ErprMAN, 1952; HAMEL, dibeet an. Jay, 1971; JONSSON; KnutTH; Kucter, 1955, 1970; Li; Martin; Moreau; Morvittez; MULLER; RICKETT, pl. 95; ROSENDAHL, 1905; SAVILE, 1961; SCHOENNAGEL; ‘Skovstep; SMALL & RYDBERG; STEINER & LOFFLER; THOUVENIN; WARMING; and Woo Boropine, M. Zur vergleichenden Anatomie der Chrysosplenium-Blatte. Trav. Soc. Nat. St. Pétersb. 14(1): 32-46. 1883.* [For summary of and ref- erence to this article, see Bot. Centralbl. 19: 291-293. 1884.] Cuten, C. P. Tres species novae generis Chrysosplenii sinenses. Acta Phytotax. Sinica 8: 337-340. 1963. Drupe, O. Uber die natiirliche highest von Adoxa und Chrysosplenium. Bot. Zeit. 37: 665-672. pl. 8 . Uber die ee co von Adoxa zu Chrysosplen- ium und Panax. Bot. Jahrb. 5: 441-447. 1884. [The tribe Chrysosplenieae, ‘i dading Chrysosplenium and Adoxa, established within the Saxifragaceae. ] Ercuincer, A. Vergleichende Entwicklungsgeschichte von Adoxa und Chrysos- rag Mitt. Bayer. Bot. Ges. 2(5): 65-74. pls. 1-3. 1907; 2(6): 81~93. 908. pede? M. A. Monographie du genre Chrysosplenium Tourn. Nouv. Arch. oes a Hist. Nat. Paris III. 2: 87-114. pls. 3-6. 1890; 3: 1-32. pls. 1-7. Binds ae A vo on Chrysosplenium iowense. Rhodora 51: 191, 192. 1949. ——.. Synopsis of the genus Chrysosplenium L. (Saxifragaceae). Jour. Fac. Sci. Univ. Tokyo Bot, 7: 1-90. 1957. yellow-flowered variety of Chrysosplenium album. Jap. Jour. Bot. 33: 69. 1958. 458 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 S. Kurosawa. Cytotaxonomical studies on Japano-Himalayan ele- ments (1). (In Japanese; English summary.) Jap. Jour. Bot. 38: 71-74. 1963. Hepsers, I., & O. Hepperc. Chromosome counts in British vascular plants. Bot. Not 14: 397-399. 1961. [C. oppositifolium, 2n = 42. Heinze, A. Om Chrysosplenium alternifolium L. v. tetrandrum Lund och des utbredning i inom Skandinavien. Bot. Not. 1907: 231-237. 1907. Lae pe W.L. The North American species of Sacre ot Erythea 5: 99. 897. [An announcement of Rose’s revisio és. A. Cytotaxonomic evaluation of lke ae taxa. Vegetatio 5-6: 212- 224. 1954. [Includes discussion of C. tetrandrum and C. americanum. ] Lostna-LastnsKaja, A. Chrysosplenium. Fl. URSS 9: 200-215. 1939. [19 species recognized. | Maximowicz, C. J. Diagnoses plantarum novarum asiaticarum. Scripsit. Bull. Acad. Sci. St. Pétersb, III. 23: 305-391. 1877. [Adumbratio specierum generis Chrysosplenii L. ee 340-350. Diagnoses plantarum novarum asiaticarum. IV. Scripsit. Ibid. 27: 425— 560. pl. 1. 1882. Addenda ad Chrysosplenia, 467—475.] NakaokI, T., & N. Mortra. A new glycoside of Chrysosplenium japonicum Mak. (IIT). toqr Pharm. Soc. Japan 76: 320-322. 1956.* Packer, J. G. Some cytotaxonomic ieuemerpiee on the genus Chrysosplenium. (Abstr.) Proc. IX Int. Bot. Congr. 2; 291. 9. The taxonomy of some North a a species of Oe L., section Alternifolia Franchet. Canad. Jour. Bot. 41: 85-103. 1963. Rose, J. N. Preliminary ae of the North American species of Chrysosplen- tum. Bot. Gaz. 23: 2 ROSENDAHL, C. O. Code a in Chop, bee special psy to the taxo- nomic status and distribution of C. iow Rhodora 49: 25-36. 1947 SAVILE, D. B. O. sega hd opie nischantea $ in ans Sif and Mitella. Science 17: 250, 251. : 1953. SCHLOEMER, A. Wachectutaediaes Milzkraut. Nat. Volk 85: 90, 91. 1955. [Includes color plate of Chrysosplenium alternifolium.| Warminc, E. Saxifragaceae. 1. Morphology and biology. Medd. Grgnl. 36: 169-236. 1909. Chiericneaaa. 226-230. | Subfam. PARNASSIOIDEAE Dandy 9. Lepuropetalon Elliott, Sketch Bot. S. Carolina Georgia 1: 370. 1817. Diminutive, leafy, acaulescent or branched winter annuals from slender primary roots and poorly developed basal rosettes, usually forming small hemispherical tufts, Leaves alternate or subopposite, spatulate, the blade constricted into a broad petiole; margins entire, apex acute or obtuse; plants glabrous but with conspicuous golden- brown tannin sacs in the epidermis. Flowers perfect, small and inconspicuous, either appearing sessile in the basal rosette or terminal at the ends of short, leafy, and angled stems. Calyx of 5 broad, persistent, often unequal sepals with acute or obtuse apices, the sepals connate below, forming a campanulate, 5-ribbed floral cup. Corolla sometimes absent or of (4) 5 linear, bractlike, sub- microscopic white petals alternating with the sepals, inserted on the rim of 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 459 15; h, cross section of ovary, X 10; i, dehisced capsule with persistent calyx, X 12; j, seed, X 50; k, cross section of seed, X 75; 1, embryo oriented as in seed, oe the floral cup. Androecium dimorphic, of 5 staminodia opposite and in- serted with the petals and 5 functional stamens opposite the sepals, the staminodia dilated distally, the filaments of the functional stamens short; anthers globose, yellow, 2-loculate at anthesis, dehiscing longitudinally ; pollen 3-colporate, subprolate, finely reticulate. Gynoecium 3-carpellate, the 3 commissural stigmas free, the styles + connate below; ovary inferior, adnate to the floral cup, 1-loculate with numerous anatropous ovules on 3 parietal placentae. Fruit a loculicidal capsule, the seeds numerous, minute, cylindrical (0.15-0.2 mm. long X 0.07—0.1 mm. broad) with a short beak 460 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 at one end, the surface reticulate, red-tinged when immature, blackish when ripe; embryo straight, endosperm scanty. Type species: L. spath- ulatum (Muhl.) Ell. (Pyxidanthera spathulata Muhl.). (Name from Greek, lepyron, a shell or husk, and petalon, a leaf or petal, apparently in reference to the inclusion of the small petals within the persistent sepals.) One species, Lepuropetalon spathulatum (Muhl.) Ell., occurring in three widely separated regions: in the southeastern United States and Mexico, in central Chile, and in Uruguay. A diminutive herb up to 2 cm. high, L. spathulatum grows in wet sandy soils in cemeteries, pastures, along power lines, and at the edges of earth-filled depressions on rock outcrops from New Hanover County, North Carolina, southward on the Coastal Plain (rarely on the Piedmont) through South Carolina, Georgia, Alabama, Louisiana, and eastern Texas and along the Gulf Coast in Mexico. Un- doubtedly overlooked by collectors due to its small size and early flower- ing period, Lepuropetalon ranks as one of the smallest terrestial flowering plants. It is probably more abundant and widespread in the Southeast than herbarium collections indicate (cf. Thomas, 1971), and it may be found associated with species such as Hedyotis (Oldenlandia) crassifolia, Krigia virginica, Trifolium carolinianum, Plantago heterophylla, Tillaea aquatica, and Anagallis (Centunculus) minima. Essentially unknown biologically, the fact that Lepuropetaion is a winter annual suggests that the plants are self-compatible and that the flowers are probably self-pollinated. Observations on a series of plants preserved in alcohol support this hypothesis and show that at the onset of anthesis (Fic. 6d) the anthers dehisce while positioned directly over the stigmas. At progressively later stages of anthesis (Fics. 6e & f) the ovary swells and the anthers are “pushed” to the sides of the ovary. The mature capsules dehisce in spring, but germination of the numerous seeds is delayed until mid-winter when dormancy is broken, probably in response to increasing day length. Engler (1891) originally placed Lepuropetalon in the Saxifraginae of subfam. Saxifragoideae but later (1928) isolated the genus in a separate subfamily and tribe, the Lepuropetaloideae Engl., tribe Lepuropetaleae Engl. Nakai (1943) went a step further and established the family Lepuro- petalaceae. On the basis of morphological similarities (cf. Murbeck, 1918), Dandy (1927) placed Lepuropetalon with Parnassia when he established the Parnassioideae. With Parnassia, Lepuropetalon shares its dimorphic androecium, the presence of the same type of tannin sacs in the epidermis, and its commissural stigmas, a character known elsewhere only in the Papaveraceae,. A chromosome count might further substantiate this place- ment, REFERENCES: Under family references see DANDY; ENGLER, et 1928; ErDTMAN, 1952; RICKETT; ROSENDAHL, 1905; and SMALL & RYDBER 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 461 McVaucH, R. The vegetation of the granitic flat-rocks. Ecol. Monogr. 13: 119- 166. 1943. [Lepuropetalon, 155.] Mourseck, S. Uber die Organisation und verwandtschaftlichen Beziehungen der Gattung Lepuropetalon, Ark. Bot. 15(10): 1-12. 1918. Nakal, T. Ordines familiae. v + 256 pp. Tokyo. 1943. [Lepuropetalaceae, 243.] Tuomas, R. D. Collecting vascular plants in the habitat near the ground — or, locating and collecting “Belly Plants.” Castanea 36: 148, 149. 1971. 10. Parnassia Linnaeus, Sp. Pl. 1: 273. 1753; Gen. Pl. ed. 5. 133. 1754. Fibrous-rooted herbaceous perennials, usually developing stout rhi- zomes or rootstocks; plants glabrous, but with tannin cells evident in the epidermis, these often imparting a rusty appearance to the foliage. Basal leaves petiolate, the blades palmately veined, ovate, reniform, or orbicular, with entire margins and cuneate, truncate, or cordate bases, Flowers per- fect, borne individually on scapiform stems with 1 cauline leaf [or many cauline leaves in sect. FIMBRIPETALUM] like the basal leaves but reduced and sessile, often with a clasping base. Calyx of 5 broad, persistent sepals, connate below, forming a poorly defined floral cup. Petals 5, showy and white [yellowish or pink], the laminar surface furrowed over the con- spicuous, greenish, parallel veins; margins entire [or fimbriate at the base], sessile or clawed. Androecium consisting of 5 staminodia opposite the petals (the column of each divided into 3 [or many] filiform divisions with glandular-appearing apices) and 5 functional stamens alternating with the staminodial whorl and the petals; anthers large, tetrasporangiate, ovate or oblong with a cordate base; pollen 2-celled when shed, 3-colpor- ate, sometimes syncolpate or 4-ruporate, oblate spheroidal or prolate. Gynoecium [3- or] 4-carpellate, the stylar region short or absent, the [3 or] 4 commissural stigmas expanding and becoming distinct during an- thesis; ovary superior or partially inferior and adnate to the floral cup, ovoid, 1-loculate, with numerous bitegmic, anatropous, tenuinucellar ovules on [3 or] 4 dilated parietal placentae, the placentae continuous with the exposed stigmatic surfaces above. Fruit a loculicidal capsule opening at the apex by four valves; seeds numerous, sawdust-like, with a loose, wing-like testa; embryo small, straight, with short cotyledons; endo- sperm of the ab initio Cellular type scanty or absent; embryo sac of the Polygonum type; embryogeny corresponding to the Asterad type; germi- nation epigeous. Base chromosome number 9. TYPE SPECIES: P. palustris L. (Named after Mount Parnassus, the slopes from which Dioscorides re- ported and described a plant, the Grass-of-Parnassus. Cf. R. T. Gunther, The Greek herbal of Dioscorides. frontisp. xii + 701 pp. Oxford, 1934.) — GrAssS-OF-PARNASSUS, BOG-STARS, WHITE LIVERWORT. A genus of about 50 species widely distributed in wet and boggy soils in the Arctic and Temperate zones of the Northern Hemisphere; three late- summer blooming species in our region and eight or nine elsewhere in the United States and Canada. All of the North American species belong to sect. Parnassia (sect. Nectarodroson Drude), while other species in this 462 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 section are trans-Eurasian at high latitudes. Four additional sections of the genus as delimited by Drude (1875) and Engler (1928) are confined to the Old World where the genus exhibits its greatest diversity in the Himalayan region. In the Southeast, the most easily distinguished species is Parnassia asarifolia Vent., a plant with orbicular or reniform leaves that resemble those of Asarum (Aristolochiaceae). The flowers of P. asarifolia have clawed petals, and the staminodia, which are shorter than the functional stamens, have the divisions terminated by glandular-appearing knobs. A relatively rare species, P. asarifolia grows in wet boggy soils and on seepage slopes at scattered localities in the mountains from West Virginia south- ward through Virginia, Kentucky, Tennessee, the Carolinas, and onto the Piedmont in Georgia, Alabama, and west into eastern Texas, where it grows in broad-leaved-evergreen bogs. The two additional species in our area resemble one another closely (both have ovate basal leaves with cuneate or truncate bases, staminodia longer than the stamens, and sessile petals) but are largely separated geographically and by differences in floral structure. Parnassia grandi- folia DC. grows in wet, circumneutral soils from northern Florida westward to Louisiana, Arkansas, eastern Texas, and Missouri, and northward in the mountains in the Carolinas, Tennessee, and Virginia. It has also been reported from West Virginia. Flowers of P. grandifolia have the stam- inodial divisions distinctly knobbed at the apex, the ovary is greenish, and the anthers are reddish, while in flowers of P. caroliniana Michaux (P. floridana Rydb.) the staminodial divisions are tapered to an only slightly enlarged, often involute apex, the ovary is whitish, and the anthers are yellow. A plant of acid soils, P. caroliniana occurs infrequently on the lower Piedmont and on the Coastal Plain in savannahs in the Carolinas, Florida, and Mississippi. A fourth species of the eastern United States, Parnassia glauca Raf. (P. americana Muhl.) is widespread and relatively common from New Jersey and Pennsylvania, westward into Iowa and northward into eastern Canada. Considerable confusion has resulted because of its similarity to P. caroliniana and P. grandifolia (cf. Alexander, 1934; Wherry, 1936), but P. glauca can be distinguished from our species by its staminodia, which are shorter than the stamens. Cytological and biosystematic studies to eluci- date the close relationships between these three species would be desirable. Although chromosome numbers have not been determined for any of our species or for Parnassia glauca, cytological studies of seven or eight species have established the base number as nine and have disclosed the presence of triploids (2m = 27), tetraploids (2n = 36), and hexaploids (2m: = 5A), as well as diploids (2n = 18) within the genus. Parnassia palustris L. sensu lato (incl. P. obtusiflora Rupr.) includes plants of all polyploid levels (cf. Banach-Pogan, 1956; Erlandsson, 1942; and Love, 1950), but the tetraploid (circumpolar) and diploid (trans-Eurasian and Alaskan) races are most widely distributed. The occasional triploids and hexaploids have apparently resulted from hybridization and chromosome doubling. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 463 The intricacy of the flowers of Parnassia undoubtedly accounts for the long and continued interest in their floral biology. Early observers con- sidered the flowers proterandrous since there appears to be a distinct stam- inate phase (during which an individual stamen curves upward and inward over the stigma, its anther dehisces while over the stigma, and the stamen moves back into rank, followed by the next stamen until all five have completed the cycle) followed by the visible outfolding and apparent maturation of the stigmas. According to the so-called “classical” theory based on these observations, the flowers are considered to be adapted for cross-pollination. Experimental studies with P. palustris, however, have shown that the dichogamy is imperfect and that the stigmas are receptive when the anthers dehisce; self-pollination and self compatibility have been demonstrated, but the plants are cross-compatible as well. The flowers are visited by numerous species of Hymenoptera, Lepidoptera, and especially Diptera, which are probably attracted by the staminodia in combination with the prominently veined petals that together act as nectar guides. Nec- tar is not secreted by the often glandular-appearing staminodial apices (and the flowers have therefore been termed deceptive) but is produced either at the base of the staminodial columns or in two basin-like folds of tissue on the adaxial surface of the columns. Anatomically, the staminodia represent a whorl of the androecium, and e stamens of Parnassia are unique in the centripetal development of xylem in the single vascular bundle of the filament. Although it is perhaps not directly correlated with the stamen movements during anthesis, Arber (1913) has suggested that the hollow cylinder, which results from exarch xylem development, gives the filaments greater elasticity. The scapiform flowering stems of Parnassia have a well developed endodermis, the nodes are polystelic, and stomata, of the ranunculaceous type, occur on both surfaces of the leaves. Although somewhat anomalous within the family, Parnassia was ac- commodated in the Saxifragaceae by Engler (1891, 1930), who first placed the genus in a tribe, the Parnassieae Engl., in the Saxifragoideae but later included it in its own subfamily, the Parnassioideae Dandy. On the basis of close generic similarities, Dandy (1927) had included both Parnassia and Lepuropetalon (q.v.) in the Parnassioideae. First included in the Saxifragaceae by Bentham & Hooker (1865), Parnassia had previously n most consistently allied with Drosera (Droseraceae), although S. F. Gray had established the monotypic Parnassiaceae as early as 1821. Later workers, each using different evidence, have either supported previous placements for the genus or suggested other affinities. Pace (1912), using embryological evidence, concluded the genus was closely allied with the Droseraceae. Arber (1913), who interpreted the androecium of Parnassia to be a greatly reduced phalange of stamens like those in Hypericum, pointed to affinities within the Hypericaceae. Lebégue (1953), on the other hand, concluded that the embryogeny of Parnassia (of the Asterad type) is unlike that of either the Saxifragaceae or Droseraceae but conforms most Closely with development in Claytonia (Portulacaceae) and the Poly- 464 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 gonales, Chenopodiales, and Caryophyllales. Chemical evidence sum- marized by Jay (1971) shows great similarities between Parnassia and Hypericum. Undoubtedly because of the conflicting evidence of relation- ship and its morphological distinctiveness, the tendency in most recent manuals and floras has been to recognize the Parnassiaceae, although Cronquist has retained Parnassia within the Saxifragaceae. Takhtajan (1959, 1969), who recognizes the Parnassiaceae, considers the affinities between Parnassia and Drosera as indicating that his Nepenthales are possibly derived from within the Saxifragales. Leaves of Parnassia contain a catechol-type tannin, a bitter principle, traces of a resin, and rutin. The plants have occasionlly been used medi- cinally as sedatives for heart palpitation and epilepsy and in cases of diar- rhea and menorrhagia. Other uses of the plants have not been reported, al- though they make attractive late-blooming additions to boggy wild-flower gardens. REFERENCES: Under family references see BENTHAM & HooKER; CRONQUIST; DANDY; ENGLER, 1891, 1928; ErpTMAN, 1952; Hamel, 1953; Jay, 1971; KNUTH; KucLer, 1955, 1970; Morr; MULLER; RICKETT, pl. 95; ScHOENNAGEL; STEINER & LOrrLeR; TAKHTAJAN, 1959, 1969; ‘THOUVENIN; and Woo ALEXANDER, E, J. Parnassia caroliniana. Parnassia grandifolia, Addisonia 18: 43-46. pls. 598, 599. 1934. [Both species correctly illustrated. ] Arser, A. On the structure of the androecium in Parnassia and its bearing on the affinities of the genus. Ann. Bot. 27: 491-510. pl. 36. 1913. BartLey, W. W. Parnassia. Am. Bot. 16: 69. 1910 BANACH-Pocan, E. Karyological studies in Parnassia palustris L. Acta Soc. Bot. Polon. 25: 675-679. 1956. [Essentially a review of Erlandsson’s and Léve’s work, with the addition of chromosome counts from Polish populations in support of Léve’s treatment. | BenneEtT, A. W. Note on the structure and affinities of Parnassia palustris L. Jour. Linn. Soc. Bot. 11: 24-31. 1871. BuCHENAU, F. Einige Reo hachtungen aus dem Gebiete der Pflanzen-Teratologie. Bot. Zeit. 20: 305-310. pl. 10. 1862. [P. palustris, 307-309 BurkKILL, I. H. Teratological observations on Parnassia palustris L. Jour. Bot. 34: 12-15. 1896. [Variation in structure studied in ca. 5,000 flowers showed the ovary most Macinsi with 3-6 carpels. | CUNNELL, G, J. The arrangement of sepals and petals in Parnassia palustris L. Ann. Bot. II. 73(91): 441-453. 1959. CHopat, A> Possibilité Bs heres de la double yore observée sur Parnassia palustris L. Bull. Herb. Boiss. II. 3: 363, 364. 96 r f ameter! de Parnassia palustris. Compt. jon gi Phys. Hist. Nat. Genéve 21: 69, 70. 1904. Daumann, E. Uber die “Scheinnektarien” von Parnassia palustris und anderer Bliitenarten. Ein Beitrag zur experimentellen Bliitendkologie. Jahrb. Wiss. Bot. 77: 104-149. 1932. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 465 Uber die Bastaubungsékologie der Parnassia-Blite. II. Ein weiterer Beitrag zur experimentellen Bliitenforschung. /bid. 81: 705-717. 1935. . Uber die Bestiubungsékologie der Parnassia-Bliite: Ein weiterer Beitrag zur experimentellen Bliitendkologie. Biol. Pl. 2: 113-125. 1960. Drupe, O. Uber die Bliithengestaltung und die Verwandtschaftsverhaltnisse des Genus Parnassia, nebst einer systematischen Revision seiner Arten. Linnaea 39: 239-324. 1875. Ercuincer, A. Beitrag zur Kenntnis und systematischen Stellung der Gattung Parnassia. Beih. Bot. Centralbl. 23: 298-317. ERLANDSSON, S. Cytological and phytogeographic investigations of Parnassia palustris L. in onan (In Swedish, English summary.) Acta Horti Berg. 13: 117-148. 1942. GRAENICHER, S. Some bes on the pollination of flowers. Bull. Wis. Nat. Hist. Soc. II. 4: 12-21. 1906. Gray, S. F. Nat. Arrang. Brit. Pl. 2: 670-672. 1821. [Establishment of the Parnassiaceae, “Parnassiae,” 670. | Gris, M. A. Sur le mouvement des étamines dans la Parnassie des marais. Compt. Rend. Acad. Sci. Paris 67: 913-916. 1868. Kozor-PottanskI, B. M. On the autecology of Parnassia palustris L. (In Russian.) Bot. as 32: 15-22. 1 pl. 1947. KuLLENBERG, B. Some observations on insects visiting and pollinating the flowers of pied palustris L. (In Swedish, English summary.) Sv. Bot. Tidskr. 47: 439-448. 1953. LEBEGUE, A. Embryogénie des Parnassiacées. Développement de l’embryon chez Parnassia palustris L. Compt. Rend. Acad. Sci. Paris 236: 1693-1695. 1953. Love, A. Some innovations and nomenclatural suggestions in the Icelandic flora. Bot. Not. 1950: 24-60. 1950. [The P. palustris complex, 4 0-44. Martens, P. Fécondation directe et fécondation croisée chez Parnassia palustris L. Compt. Rend. Acad. Sci. Paris 201: 789, 790. 1935. . Pollination et om florale chez Parnassia palustris. Bull. Soc. Bot. Belg. 68: 183-221. Moore, D. M. Additional new records for the Arkansas flora, II. Proc. Arkansas Acad. Sci. 5: 91-93. 1952.* [P. grandifolia reported from Arkansas. ] Nexrassova, V. Les Parnassia de la section Nectarotrilobus. Bull. Soc. Bot. France 74: 635-655. 1927. Pace, L. Parnassia and some allied genera. Bot. Gaz. 54: 306-329. pls. 14-17. 1912 Rav, M. “h Review of recent work on the embryogeny of some families and genera of disputed systematic position, Pp. 75-80 im Plant embryology: a symposium. vi-+ 273 pp. New Delhi. 2. Rypserc, P, A. Parnassiaceae. N. Am. Fl. 22: 77-80. 1905. SAXENA, N. P. Studies in the family Saxifragaceae. II. Development of ovule and riban ogee gi in Parnassia nubicola Wall. Proc. Indian Acad. Sci. B. 60: 196- Be eee e P. On the correlation between the number of chromosomes and the size ize of pollen grains in the Arctic species of Ranunculaceae and Saxi- in re (In Russian.) Bot. Zhur, 43: 1146-1155. 1958. [Parnassia & Saxifrag STEINMETZ, Ee F. Parnassia palustris herba. Quart. Jour. Crude Drug Res. 6: 801-806. 1966. [History and properties of P. palustris; text in English, 466 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 German, and French. The German text also published in Acta Phyto- therap. 13: 141, 142. 1966.] THompson, H. S. Radical leaves of Parnassia palustris and Valeriana dioica. New Phytol. 16: 151-153. 1917. TrecHEM, P. van. Sur les Parnassiacées. Jour. Bot. Morot 13: 326-332. 1899. Wuerry, E. T. Further notes on three plants mentioned in Claytonia vol. 1, no. 1. Claytonia 1(2): 7, 8. 1934. . Parnassia in Virginia and adjoining states— another nomenclatural mix-up. Jbid. 1(6): 51-53. 1935. . The ranges of our eastern Parnassias and Sedums. Bartonia 17: 17-20. 1936. Subfam. RIBESIOIDEAE Engler 11. Ribes Linnaeus, Sp. Pl. 1: 200. 1753; Gen. Pl. ed. 5. 94, 1754. Erect, arching, trailing, or prostrate shrubs [occasionally becoming arborescent], often with conspicuous 3-forked or simple nodal spines and smaller, internodal bristles. Stems decurrently ridged from nodes and epidermis usually shredding in long strips; lateral buds small, solitary; leaf scars sessile or slightly elevated, crescent or U-shaped, with 3 bundle scars; pith terete or stellate, commonly spongy; nodes trilacunar. Leaves decid- uous [rarely persistent] alternate (particularly on new growth) and/or in clusters on short, alternate lateral shoots; petioles exstipulate but often dilated at base; blades plicate or convolute in bud, palmately lobed [or rarely unlobed], with palmate venation and serrate or serrulate [rarely en- tire] margins; hydathodes associated with the apiculate marginal teeth. Flowers perfect [or imperfect by abortion and the shrubs dioecious ], occurring singly, in few- to many-flowered, often bracteate axillary racemes [or in umbellate infloresences]; each pedicel with a clasping, foliaceous bract at about its midpoint, and jointed or unjointed beneath the ovary. Calyx of [4 or] 5 greenish, yellow [white or red to pink or wine-colored] sepals, adnate to the ovary and connate above, forming a turbinate, cam- panulate [rotate], or tubular floral cup, the calyx lobes usually reflexed. Petals [4] 5, colored like calyx or of contrasting color, often changing with age, smaller [or larger] than calyx lobes and inserted on inner surface of floral cup [or occasionally lacking]. Androecium of [4] 5 included or exserted stamens, inserted on inner surface of floral cup and alternate with petals; filaments pubescent or glabrous, thin and tapering to the oblong or subglobose anthers that dehisce by longitudinal slits; pollen yellow [or white], spheroidal, 8- or 9-aperturate, 2-celled when shed. Gynoecium 2-carpellate, syncarpous; styles 2, included or exserted, often surrounded by an annular disc at the base of floral cup, [free or] connate for all or part of their length, terminating in a single, capitate or bifid stigma; ovary inferior, 1-loculate with few to many anatropous, bitegmic, crassinucellar ovules in 2 or more series on each of 2 parietal placentae. Fruit a spiny, bristly, glandular-pubescent or glabrous red [yellow], greenish, or pur- plish-black berry, terminated by the withered remains of the flower and 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 467 . Ribes. rita = shipir ose a, branch with immature fruits soon after oe oe 1/4; b, flower, X 5; ¢ ertical section of flower showing insertion bedded in endosperm (stippled), X 10; i, seed with aril removed, X 10. j, R. curvatum: flower, showing exserted anthers and style and sessile glands on the ovary, X 5. disarticulating when ripe if pedicel is jointed. Seeds few to many, irregu- larly elliptic-ovate, with a hard, brownish seed coat; each seed coated by a pulpy, gelatinous substance (aril) developed from the funiculus, the pulp filling the cavities of the fruit between individual seeds; embryo small, straight, embedded in copious endosperm of the ab initio Cellular or Helo- bial type. Base chromosome number 8. LecToTyPE sPECIES: R. rubrum L.; see F. V. Coville & N. L. Britton, N. Am. Fl. 22: 193. 1908. (Includ- ing Grossularia Miller, Rebis Spach, Coreosma Spach, Chrysobotrya Spach.) (Name of uncertain origin, but probably originally derived from Arabic, ribds, the colloquial name for Arabian rhubarb, Rheum Ribes L., 468 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 a plant from which a red syrup, similar to currant syrup, was prepared for medicinal use.) —-CURRANTS, GOOSEBERRIES. A genus of about 150 species, alternatively placed in two genera (Ribes sensu stricto and Grossularia P. Miller) and widely distributed in the North Temperate Zone of Eurasia, in the Mediterranean region of north- western Africa, in North America from Labrador to Alaska, southward through Mexico, and in the Andes from Venezuela and Colombia to Pata- gonia. Assignment of currants (flower pedicels jointed) to Ribes and goose- berries (pedicels unjointed) to Grossularia is largely artificial, since other characters are incompletely correlated with the type of pedicel, and the gooseberries are now generally included in Ribes as subg. GRosSULARIA (P. Miller) Pers.’ Five additional subgenera and numerous sections, some based on segregate genera proposed by Spach and sections established by Berlandier, were recognized by Janczewski (1907) in the only monograph to date that considers the genus on a world-wide basis. The more recent treatments by Coville & Britton (1908), Berger (1924), and Rehder (1940) have either treated the species in groups of undesignated rank or _ have been synthetic, inasmuch as the older infrageneric classification has been maintained, although species have been realigned and subgeneric taxa have been merged and redefined. As a result, subgeneric classification, nomenclature, and synonymy are confusing, and the genus is in need of a thorough taxonomic revision, including typification, with the application of current rules of nomenclature. Although Rides is ubiquitous in the North Temperate Zone, the greatest concentration of species occurs in North America, where over 70 species are indigenous. Seven species in two subgenera are native in the south- eastern United States, but the bulk of the North American species, repre- senting four subgenera, is centered in the western Cordillera, and over 40 species have been attributed to the Californian flora alone. Of the six recognized subgenera, two comprised of dioecious species do not occur in North America and may represent more recent diversifications in secondary evolutionary centers. The 30 species of subg. PARILLA Jancz. are restricted to the Andes of South America, whereas the 20 species of subg. BERISIA Spach occur throughout the mountainous regions of Eurasia. Subgenus Grossuaria (P. Miller) Pers. (Grossularia P. Miller, pro gen.; sect. Eugrossularia Engler) includes five species in our area with small, 1—4-flowered pendulous racemes, unjointed pedicels, turbinate to tubular floral cups, nodal spines, and scarious bud scales. Ribes cynosbati L., 2n = 16, an arching shrub with bristly ovaries, included stamens with glabrous filaments, and nodal spines sometimes absent, is a species of rich woods and heath balds from New Brunswick southward in the mountains to the Carolinas, Tennessee, Georgia, and Alabama. Of more restricted distribu- tion, R. curvatum Small (Grossularia curvata (Small) Cov. & Britt., G. combination is commonly attributed to A. Richard (Bot. Med. 2: 488. ® This 1823), but Persoon (Synop. Pl. 1: 252, 1805) initially treated Grossularia as a sub- genus of Ribes. 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 469 texensis Cov. & Berger), occurs on granitic soils from the Piedmont of Georgia to Louisiana and eastern Texas. It is closely allied to the restricted endemic R. echinellum (Cov.) Rehder (Grossularia echinellum Cov.), 2n = 16, which also has exserted stamens with villous filaments (cf. Fic. 7, j). The ovaries of Ribes curvatum, however, are covered with sessile glands, while in R. echinellum the ovaries and fruits are densely covered with gland-tipped spines, and the mature fruits resemble cockleburs (Xan- thium). First discovered in the region of Lake Miccosukee in Jefferson County, Florida, R. echinellum has also been found growing as an under- story shrub along Steven’s Creek in McCormick County, South Carolina, some 250 miles to the north-northeast. The two additional species of subg. GROSSULARIA sect. EUGROSSULARIA in the Southeast have glabrous ovaries and purplish-brown fruits. Indige- nous to the eastern United States from Massachusetts and New York southward through the mountains into North Carolina and Tennessee, Ribes rotundifolium Michx. occurs in rich woods and on grassy balds. It is closely related to the widespread Middle Western species, R. missouriense Nutt. ex Torrey & Gray (Grossularia missouriensis (Nutt.) Cov. & Britton, R. gracile Pursh), 2n = 16, that occurs in moist to dryish upland woods from South Dakota and Kansas eastward to Minnesota and western Tennessee. Shrubs of both species often lack nodal spines, and the flowers have exserted stamens and nearly cylindric floral cups, but in R. mis- souriense the peduncles are commonly longer than the pedicels, and the style is 10-12 mm. long, while in R. rotundifolium the peduncles are usually shorter than the pedicels, and the styles are 6-8 mm. long. Subgenus Corgosma (Spach) Jancz. (included in subg. Rises [subg. Ribesia Berl.| by Rehder) includes two spineless species of the South- eastern states with jointed pedicels and perfect flowers usually in several- flowered racemes. Section Heritrera Jancz. is represented by Ribes glandulosum Grauer (R. prostratum L’Hérit.), 2n = 16, a sprawling, prostrate shrub with erect or ascending racemes of eight to twelve flowers with hispid ovaries that are covered with crystalline glands. In fruit, the ripe ovaries become a bright red. Indigenous to boreal regions from Labrador and Néwfoundland to the District of Mackenzie and British Columbia, it ranges southward to Minnesota, Michigan, New York, the New England States, and in the mountains into North Carolina and Tennessee where it grows at high elevations in Picea-Abies forests and on grassy balds. In boreal regions, however, the shrubs are usually found in swamps, bogs, and cold, wet, coniferous forests. Ribes odoratum Wendl., 2n = 16, included in sect. SYMPHOCALYX Berl. (sect. Siphocalyx Endl., Chrysobotrya Spach, pro gen.) with other species with yellow calyces and leaves convolute in bud, is an occasional escape from cultivation in eastern North America but is indigenous from Minnesota to Arkansas and west and south to South Dakota, Colorado, and Texas, where it grows on cliffs and rocky hillsides. Closely allied to and sometimes considered and treated as conspecific with R. aureum Pursh, of the western cordillera, R. odoratum is characterized by its frag- 470 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 53 rant flowers with tubular floral cups 11-15 mm. long, and globose, black berries. The vegetative anatomy of Ribes has recently been reviewed and studied by Stearn et al. (1970), who concluded that the genus is remarkably homogeneous anatomically and that no consistent anatomical differences support the recognition of Ribes sensu stricto and Grossularia. As in several genera of the Saxifragoideae, leaves of Ribes are characterized by the presence of hydathodes along the blade margins. A trimerous vascular skeleton, originating from trilacunar nodes, and the presence of bulbous- based, unicellular hairs are also consistent features of the leaves. Stomata, of the ranunculaceous type, are restricted to the lower leaf epidermis in all but a few species, and cells adjacent to the guard cells cap their inner sur- faces. Wood of Rides is diffuse or ring-porous, and the small to extremely small vessels of all but two investigated species consistently possess scalari- form perforation plates; in both R. aureum and R. americanum completely simple and simple plates with vestigial bars are also present. Intervascular pitting is scalariform to opposite, rarely alternate, and the pits are circular to elongate in outline. Rays, both homocellular and heterocellular, are of two distinct widths, and axial parenchyma is not present. Flowers of Ribes vary considerably in the depth and shape of the floral cup extending beyond the inferior ovary, the length of the anther filaments, the relative length of the style, and the coloration of the calyx and corolla, apparently in adaptation to cross-pollination by insects, and in some species, possible self-pollination. Moreover, homogamy, proterogyny, and proterandry, in addition to dioecism, are biologic characters that have been observed or inferred and employed in the division of the genus into in- frageneric taxa. Initiation of flower and inflorescence primordia occurs under short-day conditions during the growing season preceding the year of anthesis, usually, after extension growth of the stem has ceased. Terminal buds remain vegetative, and infloresences are produced in the axillary buds of the current year’s shoot, or alternatively in “spur” shoots that are axillary branches of shoots two or more years old. Anthesis occurs in spring, often before the foliage is fully matured, and both self-compatibility and self-incompatibility have been demonstrated for numerous species and horticultural varieties. Nectar production at the base of the floral cup is sometimes localized in an obvious disc, and ethereal oils secreted by epidermal glands on the calyx often make the flowers fragrant. Fly and bee pollination seems to be the general rule, but Grant & Grant have described “chin” pollination by hummingbirds of the pendulous, red flowers of Ribes speciosum that have elongated tubular floral cups. Coville (1924) described pollination of flowers of R. echinel- lum by carpenter bees and noted that the long hairs on the staminal fila- ments of that species enable the insects to alight and hold on as they probe for nectar in the pendulous flowers. Although Hamel (1955) reported a chromosome number of 22 = 32 for garden plants of the South American Ribes Gayanum (Spach) Steud., and 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 471 Russian investigators have reported tetraploid and triploid progeny in the F, generation of a red currant hybrid, polyploidy has not been an impor- tant evolutionary mechanism in the genus as is evidenced by the stability of the diploid number, 2” = 16, reported for all other investigated species (96 to date). The chromosomes are small and do not differ significantly in morphology between species, and a pair of satellited chromosomes occurs consistently in the karyotypes of all species studied. Chromosomal frag- mentation, translocation, and similar structural changes also appear to have been of little consequence in the genus, at least within sections since pair- ing during meiosis in numerous spontaneous and artificial hybrids is regular, univalents are rare, pollen abortion percentages are low, and sis is disrupted by the presence of univalents, and allosyndetic pairing when it does occur is often hampered by differences in homologue length. Sterility and pairing difficulties are easily overcome, however, through treatment with colchicine, and numerous tetraploid hybrids, often of horticultural value, have been produced experimentally. Cytological studies of colchicine-induced tetraploids of R. nigrum, the black currant, have shown great variation in the number of chromosomes in somatic cells, and Vaarama (1948) has interpreted the presence of only four chromosomes in some somatic cells to indicate that extant diploids are derived tetraploids and that the original base number of the genus was our. An embryological feature characteristic of the genus is the presence of apical “caps” on the hooked synergids, structures that have been reported elsewhere only for Helianthemum vulgare (Cistaceae) and Sedum terna- tum (Crassulaceae). Fruits of Ribes are eaten by mice and other small rodents, but the seeds are destroyed, while passage of seeds through the viscera of birds appears to speed germination, and species dispersal is probably largely by birds. In experimental studies, seed dormancy has been broken by stratification. In our area, fruits of both Ribes cynosbati and R. rotundifolium were eaten by Indians, and gooseberry pie made from these fruits was famous in colonial times. Numerous cultivated varieties of R. nigrum, the black currants, R. rubrum sensu lato, the red currants, and R. Grossularia, the gooseberries, have largely replaced wild species as eatables, although they are grown commercially only on a small scale and are of minor economic importance. Several species, including R. aureum and R. sanguineum, are sometimes cultivated in gardens for their ornamental value, and shrubs of both ornamental and edible species often escape cultivation and persist in abandoned door yards. Often associated during stages of secondary succession with the white pines of Pinus subg. Strobus Lemm. sect. Quingquefoliis Duhamel,® species *See E. L. Little, ¢ r., & W. B. Critchfield, subdivisions of the genus Pinus (Pines). US. Dep. Agr. Forest Serv. Misc. Publ. 1144. 51 pp. 1969. 472 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 of Ribes are the primary hosts of Cronartium ribicola J. R. Fischer, a long-cycled heteroecious rust and causative organism of pine blister, a disease that also affects five-needled pines. Basidiospores produced by promycelia of C. ribicola parasitic on the lower surfaces of the leaves of Ribes are capable of dispersal over relatively short distances to neighbor- ing pines, but aeciospores produced in aecia on the pine host can be carried considerable distances to receptive Ribes populations. As a result, species of Ribes have been subject to programs of eradication throughout much of their range in North America and Europe as a disease control measure, particularly where the shrubs are coincident with commercially valuable pine forests. Repeatedly introduced into New York State and New England from Europe at the turn of the century on inexpensive white pine nursery stock for reforestation programs, the rust had become widely distributed by 1920 throughout the northeastern United States. Although some workers have claimed that Cronartium ribicola was introduced into Europe on stock of Ribes aureum, it seems likely that the fungus was originally endemic on Pinus cembra in western Siberia and perhaps in the Swiss Alps. Although all of the five-needled pines are equally susceptible to the fungus that eventually causes the death of infected trees, different species and varieties of Ribes have been shown through artifical innoculations to differ con- siderably in their susceptibility. Snell reports that the economically and horticulturally important currants and gooseberries are sufficiently immune to the rust to eliminate the necessity of their eradication as a disease con- trol measure. As an amply distinct genus characterized by its woody habit, alternate and/or fasciculate leaves with palmate venation, inferior ovaries that develop into juicy or pulpy berries, and its distinctive, multiporate pollen morphology (Agababian, 1963), the affinities of Ribes within the Saxi- fragaceae sensu Engler are obscure. Engler (1891) originally placed the genus in its own subfamily, the Ribesioideae (the placement restored in the 1964 edition of the Syllabus), but in 1928 included Ribes in the tribe Ribesieae A. Rich. in the otherwise predominantly herbaceous subfamily Saxifragoideae. Others, convinced of its isolated position within the Saxifragaceae, have recognized the segregate family Grossulariaceae (DC. in Lam. & DC. FI. Franc. ed. 3. 5: 405. 1805, nom. cons.), but the place- ment of that family and its composition have remained problematic. Cronquist placed the Grossulariaceae in his Rosales (in which he in- cludes all of Engler’s saxifragaceous taxa) but enlarged the family to in- clude the woody, alternate-leaved genera that were treated by Engler as constituting the subfamilies Brexioideae, Escallonioideae, Montinioideae, Phyllonomoideae, Pterostemonoideae, Tetracarpaeoideae, and Iteoideae. Hutchinson (1959, 1967, 1969), by contrast, retained the Grossulariaceae as monogeneric but enlarged the Escalloniaceae to include (with the ex- ception of the Pterostemonaceae) Engler’s other subfamilies listed above, and placed these families in his Cunoniales. Takhtajan, on the other hand, treated each of the subfamilies as a segregate family and initially (1959) in- 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 473 cluded them in the Cunoniales but later (1969) reassigned them to the Saxifragales. With the exception of Jtea and five of the six subgenera of Ribes, the taxa aligned by Cronquist, Hutchinson, and Takhtajan are distributed primarily in tropical and subtropical regions of the Southern Hemisphere. As a group, Engler’s subfamilies illustrate sometimes strong, sometimes weak reticulate relationships with one another, yet they are undoubtedly polyphyletic. New criteria are needed to resolve more clearly their relation- ships with one another and the Saxifragaceae sensu stricto. REFERENCES: Under family references see BEAN; CorE; CroNQUIST; Davis; DIPPEL; A SCHOENNAGEL; STEINER & LOFFLER; TAKHTAJAN, 1969; THOUVENIN; TrPPo: and Woop. AGABaBIAN, V. S. Palynomorphology of the genus Ribes L. (In Russian.) Izv. Akad. Nik Armian. SSR Biol. 16: 93-98. pls. 1-3. 1963. Bates, J. C. Comparative anatomical research within roe genus Ribes. Univ. Karisas Sci. Bull. 21: 369-398. 1933 Bauer, R., F. Gruser, E. Keep, R. L. KNIGHT, & F. Nitsson. Obst-Ziichtung Beerenobst, Ribes ii Subgenera Ribesia, Coreosma und Grossularia. Pp. 439-476 in H. Karpert & W. Ruporr, eds., Handbuch der kia. ed. 2. Vol. 6. xxxvii + 913 pp. Berlin. 1962. [Includes extensive bibliogra- phies and reviews of systematics, pollination ecology, and breeding studies. ] Bercer, A. A taxonomic review of currants and gooseberries. N.Y. State Agr. Paper. Sta. Tech. Bull. 109: 1-118. pls. 1-8. 1924 BERLANDIER, J. L. Mémoire sur la famille des orissilariées. Mém. Soc. Phys. Hist. Nat. Genéve 3: 43-60. pls. 1-3. [1826?] 1828. BoprysHeva, A. M., & E. Z. OKNINA. Fall-winter development of flowering buds of Ribes nigrum L. (In Russian.) Bot. Zhur. 46: 377-386. 1961. Bunyarp, E. A. The history and Sc pela of the red currant. Jour. Roy. Hort. Soc. 42: 260-270. 6 pls. 19 ———. A revision of the red ees’ Gard. Chron. III. 62: 205, 206, 217, 232, 237. 1917. [Includes 8 figures on various pages; considers varieties of R. vulgare, R. rubrum, and R. petraeum, with information concerning origin, culture, and varietal synonymy. Carp, F. W. Bush- fruits: A horticultural monograph of raspberries, black- berries. dewberries, currants, gooseberries, and other shrub-like fruits. xii + 537 pp. ed. 5. New York. 1911. [The first volume in the Rural Science Series, ed. by L. H. Bailey. Part III. The Groselles, pp. 377-484, includes sections on all aspects of gooseberry and currant horticulture, in addition to a key and descriptions of 49 species of Ribes.] Cuurcu, A. H. Types of floral mechanism. [viii] + 211 pp. 52 pls. Oxford. 1908. [Ribes, 151-167.] Cotsy, A. S. Notes on self-fertility of some gooseberry varieties. Proc. Am Soc. Hort. Sci. 23: 138-140. 1927. [Self-pollination and self-compatibility confirmed for 55 varieties of 5 horticulturally important species. | 474 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Cooper, W. S. The ecological life history of certain species of Ribes and its ee to the control of the white pine blister rust. Ecology 3: 7-16. eal F. V. Grossularia rane: f sera tees gooseberry from Florida. Jour. Agr. Res. 28: 71-74. pl. 1 [Described as new; includes notes on pollination ecology, habitat, and cae species. | ———.. Origin of the genus name Ribes. Science 82: 491, 492. 1935. & N. L. Brirron. Grossulariaceae. N. Am. Fl. 22: 193-225. 1908. DILLEMANN, G. Note sur Videntification biochimique de deux groseilliers orne- mentaux et de leur hybride. Bull. Soc. Bot. France 104: 153-155. 1957. Dartincton, C. D. A comparative study of the chromosome complement in Ribes. Genetica 11: 267-272. 1929. [Studies of R. aureum, R. sanguineum, and R. X Gordonianum.]| Daruincron, H. T., & L. B. Cutver. Keys to the species of Ribes occurring in the Great Lakes region. Mich. Agr. Exper. Sta. Circ. Bull. 170. 24 pp. 1939. _ [Includes winter and summer keys to 13 species, including R. missouriense, R. cynosbati, and R. glandulosum FERNALD, M. L. Some lithological variations of Ribes. Rhodora 7: 153-156. 1905. [R. cynosbati var. cynosbati with soft-pubescent leaves is widespread in eastern North America, while var. caren Fern., ches scattered hairs Fivaz, A. E. Longevity and germination of seeds of Ribe es, particularly R. rotundifolium, under laboratory and natural solani U. S. Dep. Agr Tech, Bull. 261. 40 pp. pls. 1-5. Washington. 193 Fracker, S. B., & H. A. BriscHLE. Measuring the at distribution of Ribes. Ecology 25: 283-303. 1944. Gotpscumint, E. Cytologische Untersuchungen an diploiden Arten und Arten- hybriden der Gattung Ribes L. Hereditas 51: 146-186. 4, . Cytologische 1 an tetraploiden Pflanzen der Gattung Ribes L. Ibid. 52: 139-150. 1964. Grant, K. A. & V. Grant. Hummingbirds and their flowers. vii -+ 115 pp. pls. 1-30. New York. 1968. Gray, A. Our wild gooseberries. Am. Nat. 10: 270-275. 1876. HAMEL, J. L. Matériaux pour |’étude caryo-taxinomique des Saxifragacées. II. Un Ribes naturellement polyploide. Bull. Mus. Hist. Nat. Paris 27: 226- 231. 5: Heprick, U. P., et al. The small fruits of New Vork. xi + 614 pp. Frontisp. 94 pls. 33rd Annual Rep, N.Y. State Agr. Exper. Sta. Albany. 1925. [Currants and gooseberries, 243-354. HEtter, A. A. The genus Ribes in California. Muhlenbergia 1: 63-104. 1904. [See also C. R. Quick, Madrofio 4: 286-290. pl. 38. 1 938.] JANCZEWSKI, E. Monographie des groseilliers, Ribes L. Mem. Soc. Phys. Hist. Nat. Genéve 35: 199-517. 1907. [Includes references to his earlier papers 1903-1906.] . Sur les anthéres stériles — Ppa Bull. Acad. Sci. Cracov. Sci. Math. Nat. 1908: 587-597. pl. 1908. ———.. Ancétres des groseilliers 4 pa Bull. Soc. Natl. Acclim. France 56: 313-320. 1909. . Suppléments 4 la monographie des groseilliers. I. Espéces et hybrides nouveaux. Bull. Acad. Sci. Cracov. Sci. Math. Nat. 1909: 60-75. 1909; 1972| SPONGBERG, GENERA OF SAXIFRAGACEAE 475 II. Espéces et variétés de la Chine, nouvelles ou peu connues. III. Le groseil- lier doré (Ribes aureum) et ses formes affines. Bull. Acad. Sci. Cracov. Sci. Math. Nat. B. 1910: 67-91. 1910; IV. Hybrides nouveaux. /bid. 1911: 612-619. 1911; V. Remarques et corrections des diagnoses. bid. 1913: 714-741. 1913. . Hybride du groseillier dioique et hermaphrodite. /bid. 1916: 98-104. 16. Keep, E. Amphiplasty in Ribes. Nature 188: 339. 1960. [Only one, instead of two, satellited chromosomes in a hybrid between R. nigrum and R. Grossu- | Satellite and nucleolar number in hybrids between Ribes nigrum and R. Grossularia and in their backcrosses. Canad. Jour. Genet. Cytol. 4: 206-218. 1962. Komar, G. A. The development of flower and inflorescence in certain representa- tives of the Grossulariaceae. (In Russian; English summary.) Bot. Zhur. 55: 954-971. 1970. MevurmaNn, O. Cytological studies in the genus Ribes L. Hereditas 11: 289-356. 1928. Nasr, T. A. A., & P. F. Warernc. Studies on flower initiation in black currant. I. Some internal factors affecting flowering. Jour. Hort. Sci. 36: 1-10. 1 1961; II. Photoperiodic induction of flowering. /bid. 36: 11-17. 2 pls. 1961. NILsson, F. Amphidiploid species in the genus Ribes. 14th Int. Hort. Congr. Rep. 1: 697-711. 1955. Species hybrids in the genus Ribes, subgenus Coreosma. (In Swedish; English summary.) Bot. Not. 109: 33-49. 1956. Polyploidy in the genus Ribes. Genet. Agrar. 11: 225-242. 1960. Orrorn, H. R., C. R. Quick, & V. D. Moss. Self-incompatibility in several species of Rabe: in the western states. Jour. Agr. Res. 68: 65-71. 1944. [Includes notes on seed dispersal and germination. } Paviova, N. M. A survey of the literature on the genus Ribes. (In Russian.) Bull. Appl. Bot. Genet. Pl. Breed. 17: 463-513. 1927. [Entries in the ex- tensive, categorized bibliography in language of publication; also includes distribution maps of subgenera. . Revision on cytological-genetical grounds of the genus Ribes L. (In Russian.) /bid. 25: 104-119. PIsant, P. L. Rassegna nent cain sul ribes nero. Rivista Ortoflorofrutticolt. Ital. 45: 701-708. Pont, F. Zur cents unserer Beerenfriichte. Beih. Bot. Centralbl. 39: 206- 221. 1922. [Study of the development of the “arillus Rake, B. A. The classification of the gooseberries: I. Useful characters for the classification and identification of nee Bristol Univ. Annual Rep. Agr. Hort. a. 1957: 60-65. pls. 2, 3. 1957. SMALL, J. K. Grossularia curvata, southern gooseberry. Addisonia 5: 47, 48. pl. 184, 1920. [Notes that the species is hardy in southern Canada. SNELL, W. H. Afoot and afloat for blister-rust. Jour. N. Y. Bot. Gard, 51: 111— 115. 1950. 3 Gaiden red currants and white pine blister-rust; spare cultivated red currants — they are sufficiently anges to this fungus. /bid. 187-190. Spacu, E. Revisio Grossulariearum. Ann. Sci. Nat. Bot. II. 4: 16-31. pl. 1. 1835. SPAULDING, P. The blister rust of hes e pine. U.S. Dep. Agr. Bur. Pl. Industry Bull. 206. 88 pp. 1 pl. Frontisp. Washington. 1911. 476 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 . Investigations of the white-pine blister rust. Jbid. 957. 100 pp. pls. 1-6. Washington. 1922. [Includes extensive bibliography concerning blister rust.] STERN, W. L., E. M. Sweitzer, & R. E. Pups. Comparative anatomy and systematics of woody Saxifragaceae. Ribes. Pp. 215-237. 3 pls. in N. K. B. Rosson, D. F. Cutter, & M. Grecory, eds., New research in plant anatomy. xii + 250 pp. 46 pls. (Suppl. 1, Bot. Jour. Linn. Soc. 63) 1970. TamAs, P. Einige physiologische und ziichterische Probleme der Befruchtung in der Gattung Ribes II. Eine Schnellmethode zur Ermittlung des Selbst- kompatibilitatsgrades bei Schwarzen Johannisbeeren. Ziichter 37: 342-344. 1967. [Berry size in self-incompatible varieties is variable, whereas size is uniform in self-compatible species. ] & A. Porpaczy, Jr. Einige physiologische und ziichterische Probleme der Befruchtung in der Gattung Ribes I. Variabilitat der Kompatibilitatsver- haltnisse bei Schwarzen Johannisbeeren, /bid, 232-238. THory, C. A. Monographie ou histoire naturelle du genre groseillier, contenant la description, histoire, la culture et usages de toutes les groseilles connues. Xvi + 152 pp. 24 pls. Frontisp. Paris. 1829 TiscHier, G. Uber die Entwicklung des Pollens und der Tapetenzellen bei Ribes-hybriden, Jahrb. Wiss. Bot. 42: 545-578. pl. 15. 1906 ———. Chromosomenstudien bei Ribes Gordonianum und seinen Eltern. Planta 4: 617-650. 1927. VaaraMA, A. A triple species hybrid in the genus Ribes, Hereditas 34: 369, 370. 1948, . Cryptic polyploidy and variation of chromosome number in Ribes nigrum. Nature 162: 782. 1948. Spindle abnormalities and variation in chromosome number in Ribes nigrum, Hereditas 35: 136-162. 1949. Wuitine, G. C. Occurrence of shikimic acid in gooseberry fruits. Nature 179: 991, 1957, Yen, T. K. Floral development and vascular anatomy of the fruit of Ribes aureum. Bot, Gaz. 98: 105-120. 1936. ZIELINSKI, Q. B. Chromosome numbers and meiotic studies in Ribes. Bot. Gaz. 114: 265-274. 1953. Subfam. ITEOIDEAE Engler 12. Itea Linnaeus, Sp. Pl. 1: 199. 1753; Gen. Pl. ed. 5. 92. 1754. Erect, deciduous [or evergreen] shrubs [or small trees| with slender, virgate branches; pith terete, white, chambered; axillary buds superposed, with 3 or 4 overlapping scales; leaf scars elliptical to reniform with 3 prominent bundle scars. Leaves alternate, short petiolate, with 2 minute, simple, free, caducous stipules at base of petiole [or leaves apparently ex- stipulate]; blades simple, elliptic to obovate with an acute or short-acum- inate apex and a cuneate base, glabrous above, finely pubescent below, the venation pinnate, the margins serrate, serrulate, [spinose, or dentate] ; foliage (in ours) turning brilliant red in the fall. Flowers perfect, in many- flowered, often elongate [axillary or] terminal racemes that (in ours) be- come secund in fruit, the pedicels short, each with a membranaceous, subu- late bract and 2 lateral stipules on abaxial surface near its articulation with 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 477 rachis, both bracts and stipules caducous. Calyx of 5 erect sepals united be- low into a shallowly obconic [or campanulate] floral cup; calyx lobes subulate [or ovate], persistent. Corolla white to pinkish, the petals 5, linear, with incurved apices, valvate in bud, inserted on inner surface of rim of floral cup, suberect, but spreading during anthesis. Stamens 5, in- serted opposite calyx lobes at outer edge of an annular to slightly lobed Ficure 8. Itea. a-e, J. virginica: a, habit of flowering branch, X 1/2; b, flower, showing petals with incurved apices and confluent styles, 5; c, cross section of two- loculate ovary with ovules biseriate on axile placentae, x 15; d, two-beaked, two-valved septicidal capsule, X 3; e, see 15. disc at base of floral cup; filaments tapering gradually to the oblong, 2- loculate, yellow anthers that dehisce introrsely by longitudinal slits; pollen subisopolar, bilateral, and 2-porate, the apertures + circular to lolongate. Gynoecium 2-carpellate, the carpels connate below, incompletely connate above; styles 2, confluent and seemingly 1, becoming bipartite in fruit; stigmas capitate; ovary superior [or partially adnate to the floral cup below], pubescent, 2-loculate, with the numerous anatropous, crassinucel- lar, bitegmic ovules 2-seriate on axile placentae. Fruit a septicidally 2- valved, 2-beaked capsule, dehiscing between the styles. Seeds numerous, oblong to ovoid and irregularly compressed [or elongate and narrowly fusiform], the surface black and shining; embryo straight, relatively large; endosperm scanty. Base chromosome number 11. (Including Diconangia Adanson, Kurrimia Wall. ex Meisner, Reinia Franchet.) Type species: J. virginica L. (Itea, the Greek name of the willow, Salix, in allusion to the catkin-like appearance of the racemes of white flowers, or perhaps to the willow-like foliage.) —- VIRGINIA WILLOW, SWEET-SPIRE, TASSEL-WHITE. A well-marked genus of perhaps ten or eleven species restricted to east- ern North America and eastern Asia, one in the eastern United States and about ten distributed from the northwestern Himalayas to Taiwan and 478 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Japan southward into tropical regions in eastern India, Java, and the Philippines. One of the woody genera that allies the floras of eastern Asia and eastern North America, [tea formerly had a wider distribution and probably ranks as a Tertiary relict. Fossil pollen of one or perhaps two species has been detected in early and middle Miocene deposits from the Pacific Northwest and Alaska (Wolfe, 1970) and Takhtajan (1969) cites investigations that have referred macrofossils from the Baltic ambers to the genus. Itea virginica L., 2n = 22, of our area and the closely allied J. japonica Oliver, endemic to middle Japan, which constitute section IEA (§ Deciduae homa, and Texas. Although known from a few scattered localities in the mountains of Tennessee and western North Carolina, I. virginica is pri- marily a species of the Coastal Plain, adjacent Piedmont, and the Missis- sippi Embayment in our area, and indirect evidence from Pleistocene sediments in Georgia (Watts, 1970) suggests that the present ‘Coastal Plain disjunct” distribution pattern has been attained during postglacial time. Usually occupying moist or wet sites, where the shrubs may become arborescent and up to three meters high, J. virginica is most common in swamp forests and low woodlands; when it occurs in upland areas it usually occurs in late spring and early summer. The inflorescences usually are 3- 17 cm. long, but Fernald ( 1947) described shrubs from southeastern Virginia with extremely short racemes (1-2.2 cm. long) as forma abbre- viata. Although the pollination ecology of Jtea has not been studied closely, it is evident from an examination of flowers of /. virginica that the anthers dehisce soon after the flowers have opened while the stamens are stiffly erect and the anthers are positioned directly over the stigmas. With age, the stamens and petals gradually move downward, and the capitate stig- mas are exposed. Chromosome studies of five species (1. virginica, I. parviflora Hemsley, J. chinensis Hooker & Arn., J. macrophylla Wall., and I. ilicifolia Oliver) suggest that the genus is probably homogeneous cyto- logically, since 2n = 22 has been reported for each of these species. Compatibility relationships and breeding mechanisms have not been in- vestigated. Generic relationships and the position of /tea have been problematic, since the genus is distinctive in several features yet appears to have affini- 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 479 ties with both the Saxifragoideae and the Escallonioideae. Engler (1891) originally included Jtea in the Escallonioideae but later (1928) removed it to its own subfamily, the Iteoideae, while Hutchinson (1959, 1967) re- tained Jtea in the woody escalloniaceous group but elevated that group to the rank of family. Britton (1905) and Takhtajan (1959) considered Itea distinctive enough to merit segregation at the familial level, and res- urrected the Iteaceae, nom. cons., originally established by J. G. Agardh (1858). Cronquist (1968), on the other hand, included /tea in his broad circumscription of the Grossulariaceae (cf. Ribes). The distinctive and exceedingly uncommon pollen morphology (bilater- al, 2-porate) found in /tea has been one of the primary arguments used to justify recognition of the Iteaceae (see Agababian, 1960, 1961, 1964; Takhtajan, 1959) and has also reaffirmed the relationship alluded to by Bentham & Hooker between /tea and Choristylis Harv., a genus of one or two species that also exhibits “iteoid” pollen.” Restricted to East Tropical and South Africa, Choristylis has been placed in the Escallonioideae near Forgesia Comm. ex Juss., from Réunion, but is now generally placed with Itea (see Erdtman, 1956, 1969). Chemical data (Jay, 1971) suggest that the affinities of Jtea may be with Brexia Noronha ex Thouars, another East African genus isolated within the se aa that has, at times, been. segregated into its own monogeneric fami Itea differs from genera of the Ee in its distinctive chambered pith, stipulate leaves, and confluent styles and stigmas. More- over, Mauritzon (1933, 1939) has stressed that the bitegmic, crassinucellar ovules and the details of development and the structure of the embryo sac of Itea coincide most closely with the conditions met with in the Saxi- fragoideae, and he urges assignment of /tea either within or near that sub- family. The corresponding embryological characters in investigated escal- loniaceous species indicate that typical ovules are unitegmic and tenuinu- cellate. Endosperm formation, nucellar in the Escallonioideae and cellular in the Saxifragoideae, has not been described for /tea. The wood of Itea is diffuse-porous, averaging about 20 vessel elements per square millimeter, and the vessel perforation plates are scalariform, each with 40-70 delicate bars. In escalloniaceous genera intervascular pits are round and usually opposite, and pits to the ray and wood parenchyma are small and round, while in /tea intervascular pitting is scalariform, and the pits to ray and wood parenchyma are large and oblong. Itea ilicifolia, of central and western China, and J. yunnanensis Fran- chet, of Yunnan, are occasionally cultivated as garden ornamentals. Both evergreen shrubs with holly-like leaves, these two species are probably less hardy in northern gardens than /. virginica, which is especially attractive in the fall when its foliage turns a brilliant red. 7 Pollen closely resembling “iteoid” pollen has been described and figured for two saprophytic herbaceous genera of the Gentianaceae, Leiphaimos and Voyria (S. Nilsson & J. J. Skvarla. Pollen morphology of saprophytic taxa in the Gentianaceae. Ann. Missouri Bot. Gard. 56: 420-438. 1969). 480 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 REFERENCES: Under family references see AGABABIAN, 1961, 1964; BEAN; BENTHAM & OOKER; CRONQUIST; DippEL; ERDTMAN, 1952, 1969; ENGLER, 1891, 1928; HaMEL, 1953; HuTcHINson, 1959, 1967; IxusE; Jay, 1971; KaNnentra; Lt; Martin; Mavritzon, 1933, 1939; Moore; Morr; Morvitiez; REHDER; Sax; SCHOENNAGEL; TAKHTAJAN, 1959, 1969; THoUVENIN; and Woop. AGABABIAN, V. Palynosystematics of the Iteaceae. (In Russian.) Izv. Akad. Nauk Armian. SSR Biol. 13: 99-102. 1960. ARNOLD, R. E. /tea virginica. Gard. Chron. IMI. 119: 63, 64. 1946. [Introduc- duction to England of J. virginica as a horticultural subject occurred before 44,] Bowven, W. M. Diploidy, polyploidy, and winter hardiness relations in the flowering plants. Am. Jour. Bot. 27: 357-371. 1940. [Records » = 11 for I. virginica and I. ilicifolia; the former is hardier than the latter. | Britton, N. L. Iteaceae. N. Am. Fl. 22: 181. 1905. CLutTeE, W. N. The Virginia willow. Am. Bot. 31: 93, 94 + frontisp. 1925. Eimer, A. D. E. The genus /tea. Leafl. Philip. Bot. 2: 527-529. 1908. [l. maesifolia and I. luzonensis described as new. Erptman, G. Pollen morphology and plant taxonomy in some African plants. Webbia 11: 405-412, 1956. [Pollen of Choristylis compared with that of Itea, 407; also published as “reprint” from Webbia in Grana Palynol. II. 1(1), no. 5. 1954, apparently prior to issuance in Webbia. | Everett, T. H. Plant portraits. Gard. Chron. Am. 51: 246, 247, 258. 1947.* [Plate of 1. virginica.] FERNALD, M. L. /tea virginica L., forma abbreviata. Rhodora 49: 2225. 2997, Fox, H. M. Three hardy native American shrubs. Gard. Chron. III. 122: 89. 1947. [Notes on horticultural value of /. virginica. | JaNssonius, H. H. Mikrographie des Holzes der auf Java vorkommenden Baumarten. Vol. 3. 764 pp. 1914. [Itea, 280-286. ] KANEHIRA, R. Anatomical characters and identification of Formosan woods with critical remarks from the climatic point of view. 317 pp. pls. 1-50. Taihoku. 1921. [Itea, 104, 105.] Watts, W. A. The full-glacial vegetation of northwestern Georgia. Ecology 51: 17-33. 1970. [Jtea pollen appears postglacially. } Wotre, J. A. Neogene floristic and vegetational history of the Pacific North- west. Madrofio 20: 83-110. 1970. [Presence of Iteoid pollen in early and middle Miocene strata in the Pacific Northwest and Alaska: see Table 1, p. 94.] Subfam. HYDRANGEOIDEAE A. Br. Tribe PHILADELPHEAE Reichenb. 13, Philadelphus Linnaeus, Sp. Pl. 1: 470. 1753 ; en. Pl ed: Sy 214, 1754. Much branched, erect, arching [or subscandent], deciduous shrubs up to 5 m. high, often planted for their showy, often fragrant flowers. Branches opposite, terete or finely ridged, the cortex and periderm tight and persistent or flaking or exfoliating in thin strips, the periderm on Ficure 9. Philadelphus. a-d, P. inodorus: a, flowering branch with dehisced capsules from previous seaso 1/2; b, flower ~~ bud with valvate calyx lobes, X 1; c, detail of style and four- lobed stigma, X 6; d, cross section of ovary, xX 5. ed, P. hirsutus: e, immature frat ent after ‘disarticulation of corolla and androecium —note undivided stigma and ovarian disc, x 3; vertical section through ciiadee fruit showing auriculate placentae and pendu- lous ovules, < 5; g, mature, dehisced capsule, x 2; h, seed, X 6; i, avin section of seed showing embryo, ~ It embryo sid as in seed, ae older shoots gray, grayish-brown, or reddish, the epidermis on new growth greenish or chestnut-brown; pith spongy and pale white, terete; lateral buds solitary with 2 valvate scales [or several imbricate scales] ; leaf scars elliptic or cordate, transversely connected, with 3 bundle scars. Leaves short-petiolate, exstipulate; blades dark green above, pale green below, glabrous or variously pubescent, lanceolate, elliptic, or ovate in outline, with acuminate apices and attenuate, obtuse, or truncate bases; margins entire or irregularly serrate or dentate (the apiculate teeth seemingly as- ated with hydathode-like vein endings); venation pinnate with 3 or 5 major nearly parallel veins. Flowers short-pedicellate, usually with 2 erect, linear bracts subtending the flower; flowers borne singly or in 2’s or 3’s (a simple dichasium) at the ends of lateral branches, sometimes with paired 482 JOURNAL OF THE ARNOLD ARBORETUM - [voL. 53 flowers at successively lower nodes, often in the axils of leaves. Sepals 4, rarely 5, connate below and adnate to the ovary, forming a turbinate to subcampanulate floral cup; calyx lobes valvate in bud, deltoid, with acute or acuminate apices, often becoming reflexed in fruit but deciduous before capsule dehiscence. Petals 4, rarely 5, imbricate in bud, inserted on rim of floral cup, white or rarely purplish, orbicular, ovate, oblong, or spatulate in outline, forming a disciform or cruciform corolla. Stamens numerous, usually erect and surrounding the style, inserted on rim of floral cup; fila- ments subulate, tapering gradually to the basifixed oblong or subglobose anthers, the anthers with obvious connectives; pollen (of 1 examined species) finely reticulate, 3-colpor(oid)ate (longicolpate), subprolate. Gynoecium 4(5)-carpellate, styles 4 (rarely 3 or 5), completely connate, partially free above [or in some cultivated forms completely free], ter- minated by free clavate, linear, or cristate stigmas, or stigmas connate and columnar or subcapitate; ovary semi-inferior or inferior, adnate to floral cup, 4(5)-loculate, with numerous tenuinucellar, anatropous, unitegmic ovules pendulous on auriculate axile placentae, the placentae projecting from the upper or middle portions of the central axis. Fruit a corticate, loculicidal capsule; seeds numerous, oblong-cylindric, with a membran- aceous brownish or blackish testa, usually prolonged into a tail; embryo small, straight, embedded in fleshy, oily endosperm of the ab initio Cellular ype. Base chromosome number 13. (Syringa Adanson non L.) LrEcrTo- TYPE SPECIES: P. coronarius L.; see P. A. Rydberg, N. Am. Fl. 22: 162. 1905. (Name commemorating Ptolemy Philadelphus, King of Egypt, 283- 247 B.C., sometimes translated as brotherly love.) — Mock-ORANGE, SYRINGA. A taxonomically complex genus comprised of perhaps as many as seventy species in four subgenera exhibiting a disrupted arcto-Tertiary distribution in the North Temperate Zone and extending southward into Central America. Hu (1954, 1955, 1956) has recognized seven species representing two subgenera in the southeastern United States, and 39 species ranging in the Cordillera of western North America and extending southward through Mexico and into Panama. In Asia, 22 species are recorded as native from Korea, northeastern China, and Japan southwestward through China and into the Himalayan region, while three species are indigenous to south- eastern Europe and Asia Minor. According to Hu’s treatment of the genus, subg. PHILADELPHUS (subg. Euphiladelphus Hu) (shrubs with enclosed lateral buds, deter- minate branching, and determinate racemose or few-flowered inflorescences comprised of flowers with 25-90 stamens, clavate or linear stigmas, and caudate seeds) includes five species in our area. Three, P. intectus Beadle, 2n = 26, P. pubescens Loisel., 2n = 26, and P. Gattingeri Hu, are stiff- persistent calyx lobes, and short-caudate seeds. Philadelphus intectus (incl. var. pubigens Hu), up to 5 meters high, is characterized by its non- 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 483 exfoliating bark and inflorescences of five, seven, or nine flowers with disci- form corollas and glabrous floral cups. Growing on steep, rocky bluffs and _ bluff tops, P. intectus is known from a few localities along the Cumberland River in north-central Tennessee and southwestern Kentucky and more commonly in north- and west-central Arkansas and eastern Oklahoma. Ac- cording to Hu, the glabrous leaves and floral cups of P. intectus suggest a close relationship with P. Lewisii Pursh, of the Pacific Northwest, while in habit and leaf shape, it resembles P. pubescens. Philadelphus pubescens (incl. var. verrucosus (Schrader) Hu; P. Jati- folius Schrader) is not only similar in habit to P. intectus, but it occurs in the same geographic areas, viz., the Nashville Basin and the Ozark Plateau and Ouachita Mountains in Arkansas and eastern Oklahoma. It also occurs disjunctly in northeastern Texas, in Pope County, Illinois, and in Mc- Donald County in southwestern Missouri. Apparently a calciphile of steep, wooded slopes, the plant is characterized by its pubescent foliage and seven- to nine-flowered inflorescences, the flowers with strigose-villose floral cups and cruciform corollas of glabrous, obovate-oblong petals. Philadel- phus Gattingeri, known only from the type locality on the bluffs along the Cumberland River near Nashville, Tennessee, is doubtfully distinct from P. pubescens, but Hu has separated it on the basis of the pubescent abaxial surface of the petals, ovarian discs, and styles. Both Philadelphus floridus Beadle, 2n = 26, and P. inodorus L. (includ- ing vars. grandiflorus (Willd.) Gray, laxus Hu, carolinus Hu, and strigosus Beadle; P. grandiflorus Willd., P. gloriosus Beadle), 2 = 26, are arching shrubs one to three meters high, with solitary flowers or true dichasia of flowers with long-acuminate calyx lobes, apically attached placentae, ellip- soid or subovoid capsules with median, persistent calyx lobes, and long- caudate seeds. Highly variable in leaf shape, corolla form (cruciform, disciform, and intermediate types are known), and pubescence, P. inodorus has glabrous pedicels and floral cups, while P. floridus has pubescent ones. Philadelphus floridus is known in nature only from the type locality on the rocky slopes of the Coosa River in Floyd County, Georgia. Philadelphus inodorus occurs on rocky slopes in rich, moist woodlands in the Piedmont and in the mountains from southwestern Virginia through the Carolinas and into Georgia and Alabama (where it extends southward onto the Gulf Coastal Plain) and westward into central Tennessee. Localities in northern Virginia, northern Kentucky, and Pennsylvania are probably based on plants that have escaped from cultivation. Subgenus DeurziorEs Hu (shrubs with exposed lateral and terminal buds, indeterminate branching, and solitary flowers or ternate inflores- cences composed of flowers with 13-35 deve columnar or subcapitate four-grooved stigmas, and tailless seeds) includes one or possibly two species in the Southeast. Philadelphus hirsutus Nuttall (incl. vars. intermedius Hu and nanus Hu), 2n = 26, a low, spreading shrub up to two meters high, is a plant of rocky ledges and bluffs in the mountains of south- western North Carolina, northern Georgia and Alabama, and Tennessee, and it occurs disjunctly on the Ozark Plateau in north-central Arkansas. 484 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 It is characterized by flowers with pubescent floral cups, disciform corollas of ovate or suborbicular petals, and pubescent foliage. Philadelphus Shar pianus Hu (including var. parviflorus Hu) differs only in its glabrous floral cups, cruciform corollas of oblong petals, and sometimes glabrous undersurface of the leaves. Philadelphus Sharpianus is known only from a very local area in the Great Smoky Mountains in Tennessee and, like P. hirsutus, from a disjunct station in Van Buren County, Arkansas. Although the genus is relatively well defined, Hu’s hierarchy of sub- generic categories (including subgenera, sections, and series), which is based on the older, partial classifications of Koehne (1893, 1896, 1904) and Rehder (1940, 1949), needs a thorough review both to determine its biological validity and to bring the nomenclature into accord with the current Code. The apparent confusion that exists in intrageneric classifica- tion is largely due to the species problem in Philadelphus. Most species ex- hibit considerable morphological variation and consist of polymorphic population systems; determinations consequently are difficult. A plethora of specific and varietal names attests to attempts to recognize this variation, and the problem has been compounded by the description of many taxa from cultivation. All of the investigated species of Philadelphus (35 to date) are diploids with 2m = 26. Interspecific hybridization apparently occurs easily in horti- cultural collections, suggesting the possibility of hybridization in nature as well. Several hybrids of horticultural origin have been studied cytologically by Bangham (1929) and Janaki Ammal (1951) with differing results. Al- though Bangham reported no obvious meiotic disruptions in microsporo- genesis in hybrids, Janaki Ammal noted chromosomal lagging and other abnormalities that often resulted in high percentages of abortive pollen. Bangham recorded 2m = 26 for all the hybrids he studied, but Janaki Ammal found two triploid (2n = 39) cultivars of the trihybrid P. x pur- pureo-maculatus Lemoine and an aneuploid cultivar, P. ‘Beauclerk’ with 2n = 28. Additional aneuploid counts from the progeny of triploid culti- vars have been reported by Sampson (1955). ‘Anthesis occurs during May and June, and the odor of flowers of numer- ous species is suggestive of the fragrance of orange blossoms. Pollination is by numerous long- and short-tongued bees, dipterans, and lepidopterans that probe between the stamens for the abundant nectar secreted by the ovarian disc. Philadelphus coronarius L. has been observed to be protero- gynous, but self-pollination is undoubtedly possible once the anthers (which mature centripetally) begin to dehisce. Compatibility relationships, however, have not been investigated, and the maturation sequences in flowers of other species have not been observed. The petals and stamens are deciduous, but the styles persist for a longer period on the immature fruits, which gradually develop into woody capsules late in the growing season. The capsules remain on the shrubs for a considerable period before dehisc- ing, and the seeds may not be liberated until the following growing season. Dispersal of the ripe seeds is undoubtedly by wind. Ovules of Philadelphus are unitegmic, anatropous, and tenuinucellar, and 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 485 the archesporal cell undergoes meiosis directly to form a linear tetrad of megaspores. The chalazal megaspore develops into a Polygonum-type embryo sac that, during development, becomes partially extra-ovular by passing through the micropyle. An endothelium forms around the lower half of the embryo sac, and a micropylar haustorium of two or four elon- gated cells develops. Wood of Philadelphus is ring- or semiring-porous, typically with solitary vessels and paratracheal scanty parenchyma. Vessel perforation plates are scalariform or occasionally simple, and intervascular pits are small, round, and opposite. Stomata, of the ranunculaceous type, occur in foliage epider- mis, and the leaf teeth are glandular in certain species with dentate or serrate leaf margins. Typically, a single cell layer comprises the palisade mesophyll, and in transverse sections the petioles disclose an open, crescent- shaped vascular bundle. Initially placed in the Myrtaceae by A. L. de Jussieu, Philadelphus has not been allied with that family since D. Don (1826) established the Phila- delphaceae. De Candolle, Endlicher, and Lindley recognized the Phila- delphaceae as distinct but enlarged it to include Decumaria and Deutzia Thunb., a genus of about 40 species indigenous to south-central and eastern Asia. The affinities of Philadelphus with Hydrangea were alluded to by De Candolle, but Reichenbach (1828) was the first to include the above genera, along with Hydrangea, with other genera customarily included in the Saxifragaceae sensu lato. Together, Philadelphus, Deutzia, and several western North American genera (including Neodeutzia Small, Fendlera Engelm. & Gray, Carpenteria Torrey, Jamesia Torrey & Gray, Fendlerella Heller, and Whipplea Torrey) form a seemingly natural group, the tribe Philadelpheae Reichenb. of subfamily Hydrangeoideae in Engler’s system. As delimited by Engler, the tribe is usually included in the Hydrangeaceae § when that family is segregated from the Saxifragaceae. The recent sys- tems of Cronquist and Takhtajan (1966, 1969) both include the Phila- delpheae in the Hydrangeaceae, but Hutchinson (1959, 1967) has distri- buted the genera of Engler’s Hydrangeoideae rather indiscriminately be- tween the Hydrangeaceae and Philadelphaceae. Within the Philadelpheae, Philadelphus is morphologically most similar to Deutzia (a cytologically variable genus with recorded coun = 26, 78, 104, 127, and 130), Fendlera (2n = 22), a small genus of four xerophytic species in the southwestern United States and Mexico, and Neodeutzia, a genus of two species of northwestern Mexico which are often included in Deutzia. Differentiation between these genera, all comprised of shrubs, is based primarily on constant stamen numbers (Deutzia, Fend- lera) or variable ones (Philadelphus, Neodeutzia), on modifications in sta- men morphology, and on the numbers of parts comprising the calyx and corolla. Evidence from flavonoid chemistry (Jay, 1971) suggests that *Hydrangeaceae Dumort. (Anal. Fam. 36, 38. 1829), nom. cons., and not Phila- delphaceae D. Don (Edinb. New Phil. Jour. 1: 132-135. 1826), is used when Phila- delphus is eee Beto: Hydrangea in a segregate family. See Appendix II, Int. Code Bot. Nomencl. 486 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Deutzia is relatively unadvanced, while Philadelphus has attained greater chemical complexity, similar to that evident in Hydrangea. Disease-free, easily grown, and readily propagated from cuttings, species of Philadelphus have been widely planted in the North Temperate Zone as late-spring to summer-blooming ornamentals. About fifty species and varieties and numerous interspecific hybrids of horticultural value, pro- duced mainly by the Lemoine & Son Nursery in Nancy, France, are in cultivation. Philadelphus coronarius L., native to southeastern Europe and the Caucasus and one of the most widely cultivated species, may persist in old gardens and door yards. This species has also been used as parental stock for many of the hybrids available today. Aside from their horti- cultural value, species of Philadelphus are of no other economic importance, although water scented with the flowers was used as a perfume by the ancient Parthians. REFERENCES: Under family references see BEAN; Cronguist; DippeL; ENGLER, 1891, 1928; ERDTMAN, 1952; GeLius; HAMEL, 1953; HUTCHINSON, 1927, 1967; Jay, 1971; KnutH; Li; Lussock; Martin; Mauritzon, 1939; Moore; Morr: Mor- VILLEZ; Mier; REHDER; ROBERTSON; Sax; SCHOENNAGEL: STEINER & LOFFLer: TAKHTAJAN, 1966, 1969; THOUVENIN; and Woop. sage are es oe oranges. Garden London 38: 413. 1890; Ibid. 40: 288, 289. pl. 8 1891. ee Ww. The chromosomes of some species of the genus Philadelphus. Jour. Arnold Arb. 10: 167-169. 1929. BEADLE, C. D. Studies in Philadelphus. Biltmore Bot. Stud. 1: 159-161. 1902. [Includes P. intectus, P. floridus, and P. gloriosus as new species and P. inodorus var. strigosus. ren pe T. D. A. Carpolithes macrophyllus a Philadelphus. Torreya 19: 244. 919. [Reassignment of fossil fruit materials to Philadelphus; see Torreya $4:°235.: 2911. Creecu, J. L. Philadelphus notes. The Lemoine hybrids. Natl. Hort. Mag. 27: 111-123. Frontisp. 1948. [Includes hybrids of horticultural interest. ] Don, D. Observations on Philadelpheae and Granateae, two new families of plants. Edinb. New Philos. Jour. 1: 132-135. 1826. [Establishment of the Philadelphaceae; Philadelphus previously included in the Myrtaceae. ] Gris, A. Note sur quelques cas de monstruosité observés sur le Philadelphus speciosus. Bull. Soc. Bot. France 5: 330-334. 1858. Hitcucocx, C. L. The xerophyllous species of Philadelphus in southwestern North America. Madrofio 7: 35-56. 1943. Hv, S. Y. A monograph of the genus Philadelphus. Jour. Arnold a 35: 275— 333. 1954; Ibid. 36: 52-109, 325-368. 1955; Ibid. 37: 15-90 JANAKI Amat, E. K. Chromosomes and the evolution of bacden Pahoa Jour. Roy. Hort. Soc. 76: 269-275. 1951. JentscH, R. Untersuchungen an der rigid ansren agit einiger Saxi- fragaceen. Flora 144: 251-289. 1956/57 KoEHNE, E. Deutsche Dendrologie. xvi + 601 pp. Stuttgart. 1893. [Philadel- phus, 179-185.] 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 487 —. Philadelphus. Gartenflora 1896: 450, 451. 1896, . Zur Kenntnis der Gattung Philadelphus. Mitt. Deutsch. Dendrol. Ges. 13: 76-86. 1904. . Philadelphus. In: C. S. SarceNT, Pl. Wilsonianae 1: 4-6. 1911; 2: 145. 1912, [Publ. Arnold Arb. No. 4.] Moore, A. H. Concerning Philadelphus platyphyllus Rydb. Rhodora 16: 77. 1914. [Also see Rhodora 17: 121-123. 1915.] Nakal, T. Philadelphus Japano-Coreanae. Bot. Mag. Tokyo 29: 63-67. 1915. REHpER, A. New species, varieties and combinations from the herbarium and the collections of the Arnold Arboretum. Jour. Arnold Arb. 1: 191-210. 1920. [Philadelphus, 195-203; Fendlera, 203-206; Deutzia, 206-210.] REICHENBACH, H. G. L. Conspectus Regni Vegetabilis. xiv + 294 pp. Leipzig. 1828. Ricketr, H. W. Inflorescence of Philadelphus. Am. Midl. Nat. 31: 759-762. 1944 Sampson, D. R. Studies on the progeny of triploid Philadelphus and Forsythia. Jour. Arnold Arb. 36: 369-384, 1955. SCHLECHTENDAL, D. F. L. von. Pflanzen-Missbildungen. Linnaea 16: 463, 464. 1842. [P. coronarius. | TiecHeM, P. van. Structure de quelques ovules et parti qu’on eu peut tirer pour améliorer la classification. Jour. Bot. Morot 12: 197-220. 1898. [Saxi- fragaceae, 205-207, 216.] WetrtsTEIN, R. von. Ueber das Androeceum von Philadelphus. Ber. Deutsch. Bot. Ges. 11: 480-484. pl. 24. 1893. Wyman, D. The mock-oranges. Arnoldia 25: 29-36. 1965. [Includes list of the better species and cultivars for use as garden ornamentals, as well as notes on the origin and parentage of numerous hybrids. YamapayasHt, N. Types of vessel perforations in Korean woods. Trop. Woods 46: 20-22. 1936. Tribe HypRANGEEAE DC. 14. Hydrangea Linnaeus, Sp. Pl. 1: 397. 1753; Gen. PI. ed. 5. 189. 1754. Erect or ascending deciduous [or evergreen] shrubs or [small trees or high-climbing vines clinging by adventitious aérial roots]. Stems with reddish-brown exfoliating epidermis on new growth and dark grayish-black or brownish, sometimes exfoliating bark on older parts of shoots; pith round, relatively large, spongy and unlignified [or lignified]; lateral buds opposite and solitary, sessile or very short-stalked, globose-conical to ob- long and spreading, with 4-6 exposed scales; leaf scars crescent-shaped or irregularly 3—5-angled with 3 bundle scars; nodes trilacunar. Leaves opposite [or rarely whorled with 3 at a node], petiolate; blades thin and membranaceous [or coriaceous in evergreen species |, variously pubescent [or glabrous], ovate, lanceolate, or elliptic in outline, rarely deeply 4—6- lobed, with [entire] denticulate, or serrate margins, acuminate, acute or attenuate apices, and [attenuate, obtuse, or] truncate bases; venation reticulate, usually with a prominent midvein. Inflorescence a terminal (occasionally axillary) compound cyme or pyramidal thyrse [often appear- ing umbellate in evergreen species]; pedicels with numerous small, lanceo- 488 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 late bracts [or lacking bracts within the inflorescence, but the entire in- florescence inclosed in bud by several large, ovate, caducous bracts that leave distinct scars]. Flowers either fertile or neuter and inflorescences comprised of one or both types; neuter flowers lacking or present an terminating the branches in mixed inflorescences, or comprising entire inflorescences (particularly in cultivated forms), consisting of 3-5 en- larged, petaloid, bie [through pink to blue] persistent calyx lobes, often with an obvious “eye” (comprised of abortive floral parts and sometimes normal-appearing pi wekg fertile flowers (rarely lacking) central to the neuter flowers or comprising entire inflorescences, perfect [or imperfect by abortion, and the plants functionally dioecious]. Calyx of fertile flowers of 4 or 5 sepals, connate and adnate to the ovary, forming a ribbed, campanulate, hemispherical, or oblong patelliform floral cup with 4 or 5 persistent, deltoid calyx lobes projecting above. Petals 4 or 5 white, ovate or spatulate, inserted on the rim of the floral cup, valvate in bud and de- ciduous during anthesis [or rarely falling together as a calyptra at the onset of anthesis]. Stamens 8-10 [rarely to 20], inserted on the rim of the floral cup; filaments filiform and often contorted, tapering to the subor- bicular to oblong, tetrasporangiate anthers. Gynoecium 2—4(5)-loculate, the ovary adnate to the floral cup, inferior [or half-superior] with numer- ous unitegmic, tenuinucellar, anatropous ovules on 2—4(5) axile placentae; styles 2-4(5), free (when the ovary is inferior) [or united at the base] and projecting from the center of a flat [or conical] ovarian disc; stigmas 2-lipped, crescent-shaped, at the apices of the styles. Fruit a brown, strongly ribbed, poricidal capsule, dehiscing between the styles; seeds numerous and small, linearly elliptic [or rarely circular and winged or sometimes caudate]; embryo small, linear, embedded in endosperm of the ab initio Cellular type. Base chromosome number 16. (Hortensia Com- merson ex Jussieu, Cornidia Ruiz & Pavoén, Sarcostyles Pres] ex Seringe.) Type species: H. arborescens L. (Name from Greek, Aydor, water, and angeion, the diminutive of angos, a vessel or container, in allusion to the shape of the mature, dehisced capsules.) — GRAYBEARD, SEVENBARK, HORTENSIA. Twenty-four or twenty-five species customarily placed in two or three sections distributed in temperate and subtropical regions of eastern Asia, eastern North America, Mexico, Central and South America. In the Northern Hemisphere, the distribution of Hydrangea is that of a Tertiary relict with species in eastern Asia and eastern North America. The genus is also known from several Tertiary floras in China and western North America (including sites in Colorado, Idaho, California, Oregon, Wash- ington, and Alaska), but fossil or extant species are sa known from south- eastern Europe and Asia Minor. Nine species of sect. HypRANGEA (sect. Euhydrangea Maxim., subgen. Hortensiella Kichhe. pera Euhydrangea (Maxim.) Schneider) are distributed from the eastern Himalayas to central and southern China, southward in the Malay Archipelago to Java, and in 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 489 the Philippines and Ryukyu Islands and Japan, while two species are native in eastern North America. Endemic within our region, Hydrangea quercifolia Bartram, 2n = 36 an arching or sprawling shrub up to 4 m. high characterized by pyramidal inflorescences, numerous large neuter flowers, and large, deeply four- or five-lobed leaves, occurs eastward from the Mississippi River in Louisiana, Mississippi, and southwestern Tennessee through south-central Tennessee and Alabama and into western Georgia and the panhandle of Florida. Occasionally forming dense thickets along stream banks, the shrubs occur in wet, sandy soils and on the slopes of deep ravines and canyons in decidu- ous forests. Hardy and cultivated northward, plants of H. quercifolia are infrequent escapes or persistent relicts in abandoned dooryards in the Mid- and North-Atlantic states Hydrangea arborescens L.,® 2n = 36, an ascending shrub up to three meters high, commonly with numerous branches from the base, ovate to elliptic leaves with acute or acuminate apices, serrate margins, and the in- florescences broadly convex compound cymes, is widely distributed in east- ern North America, ranging from southern New York to Georgia and Alabama and westward into Illinois, Missouri, and eastern Oklahoma. In contrast to H. quercifolia, a relatively invariable species, H. arborescens is polymorphic, especially in the presence or absence of neuter flowers and in leaf shape, size, and vesture. Up to eight additional taxa in this com- plex have been attributed to eastern North America, but McClintock (1957a) has treated these as comprising three subspecies of H. arborescens differentiated on the basis of leaf indumentum and trichome type. Hydrangea arborescens subsp. arborescens (H. vulgaris Michx., H. arborescens var. australis Harbison) occurs throughout the range of the species. In our area it is primarily a shrub of the Piedmont and mountains, although populations are known from a few scattered localities on the Coastal Plain in North Carolina, Georgia, Alabama, Mississippi, and Louisiana. Leaves of this subspecies are glabrous on the under surface or have scattered, smooth-walled hairs along the mid- and secondary veins. The lower leaf surfaces of subsp. radiata (Walter) McClintock (H. radiata Walter, H. nivea Michx., H. cinerea Small, in part), by contrast, are white or silvery with a dense, felt-like indumentum of smooth-walled, uniseriate trichomes; in addition, the neuter flowers of this subspecies are sometimes bluish-white. Endemic within our area, subsp. radiata is known from the mountains and adjacent Piedmont of southwestern North Caro- lina, adjacent South Carolina and Georgia, and southeastern Tennessee. Plants having lower leaf surfaces with a grayish pilose or tomentose ap- ’ Species of Hydrangea are sometimes confused with species of Viburnum (Capri- foliaceae) with rae: neces including radiate marginal flowers, and Mackenzie (Torreya 27: 81-83. 1927) asserts that V. alnifolium Marsh, and V. americanum Miller are taxonom ae synonyms od et arborescens. For a further discussion Situation, see footnote 3, , 43 in I. K. Fercuson, The genera of the a — in the neat i phe States. Jour. Arnold Arb. 47: 33-59. 490. . ~ JOURNAL OF THE ARNOLD ARBORETUM _ [vor. 53 pearance imparted by uniseriate trichomes with the walls roughened by tubercles are referable to subsp. discolor (Seringe) McClintock (H. arborescens var. Deamii St. John, H. Ashei Harbison, H. arboréscens var. _ discolor Seringe, H. cinerea Small, in part). Less restricted in range than subsp. radiata, subsp. discolor is known from scattered localities from the mountains of the Carolinas, Georgia, and northern Alabama northwest- _ward through Tennessee and Kentucky and into southern Indiana, Illinois, Missouri, Arkansas, and eastern Oklahoma. Like all species of Hydrangea outside of cultivation, H. arborescens is a woodland shrub, occupying mesic slopes and ledges in deciduous forests, but both it and H. quercifolia appear to be sun tolerant, and they often grow at forest margins and in open areas when under cultivation. Although the ranges of both H. arborescens subsp. discolor and subsp. radiata occur completely within the range of subsp. arborescens, shrubs with inter- mediate foliage vesture are infrequent and usually appear to be inter- mediates between subspp. discolor and radiata. At one locality in Transyl- vania County, North Carolina, McClintock (1957a) found all three subspecies growing within yards of one another along with several aberrant individuals. She concludes, however, that intermediates “have not caused the character of [subspecies] populations to change.” Hydrangea arborescens and H. quercifolia comprise subsect. AMERICANAE (Maxim.) Engler (shrubs with white flowers, inferior ovaries, truncate capsules, and the seeds not caudate), but overall morphological similarities between H. quercifolia and H. sikokiana Maxim., a shrub with deeply four- to six-lobed leaves endemic to Honshu and Kyushu, Japan, and between H. arborescens and H. aspera D. Don, a wide-ranging east Asian species poly- morphic with respect to foliage and pubescence, suggest that the closest allies of our species are these Asian species and not one another. McClin- tock (1957a) has included both H. aspera and H. sikokiana in subsect. ASPERAE Rehder (a subsection differing from the AMERICANAE in flower- color range and caudate seeds), but Engler (1928) recognized the similari- ties between H. quercifolia and H. sikokiana and included the latter species in subsect. AMERICANAE, -Philippine Islands and Taiwan, species of sect. CorNIDIA (Ruiz & Pavén) Engler (evergreen vines, rarely erect shrubs, with broadly ovate bracts at the base of the inflorescence, inclosing the inflorescence in bud) are re- stricted to the New World. Two Mexican species have been recognized, but one or two additional species may remain undescribed from the state of Chiapas (Nevling & Gémez-Pompa, 1968). The eleven additional species in sect. Cornip1a are distributed in mountainous regions of Guate- mala, central Costa Rica, Panama, and at moderate elevations in the Andes from Colombia to central Chile and adjacent Argentina. That some 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 491 recognized species in this section may, however, actually represent either staminate or carpellate plants of functionally dioecious species has been pointed out by Nevling & Gémez-Pompa (1968). Widespread in eastern Asia and sometimes cultivated in western gardens, Hydrangea anomala D. Don, a high-climbing woody vine with deciduous foliage and with the petals of the flowers united at their tips and falling together as a cap at anthesis, has sometimes been placed in its own mono- typic section, CALYPTRANTHE Maxim. Alternatively, the species has been placed in subsect. CALYPTRANTHE (Maxim.) McClintock within sect. HYDRANGEA, a section that otherwise includes deciduous shrubs or small trees with flowers with petals free at their apices. With the exception of H. involucrata Siebold, endemic to Japan and Taiwan, all species of sect. Hyprancea, including H. anomala, are characterized by bracteate in- florescences that lack at the base the large bracts which are caducous be- fore anthesis. Diploid, 2n = 36, and one tetraploid, 2” = 72, chromosome numbers have been recorded for Hydrangea, and three of the four taxa in our region are diploids; only H. arborescens subsp. discolor remains uncounted. Cytological investigations have been limited to species of sect. HYDRANGEA, although low percentages of viable pollen noted by Nevling & Gémez- Pompa in some species of sect. Cornip1A might indicate disrupted meiotic mechanisms in those species. It is probable, however, as these authors suggest, that those species with low pollen stainability are carpellate plants of functionally dioecious species. Only one spontaneous interspecific hybrid in nature and Hie hybrids, both apparently inviable, arising in horticultural collections have been reported. Since most species bloom at the same time and occupy similar habitats, it appears that interspecific isolation is largely genetic. An- thesis occurs from about the end of June through August and into Septem- ber, although the persistent, often colorful neuter flowers may give a false - impression of the actual duration of anthesis. Observations have not been recorded, but it appears that our species are proterogynous and that pol- lination by a wide variety of insects, including long- and short-tongued bees, numerous dipterans, moths and butterflies, and beetles, is the rule. Plateau (1898), in observations of insects visiting inflorescences of Hydrangea macrophylla (Thunberg) Seringe (as H. opuloides) noted that insects flew directly to and landed on the center fertile flowers, bypassing the showy peripheral flowers. Fruit-set in inflorescences of H. arborescens subsp. arborescens in the southern Appalachians is normally high despite the usual absence of neuter flowers. The poricidal capsules ripen during the fall, but dehiscence and dispersal of the numerous seeds probably does not occur until sometime during late winter or early spring. Dispersal is undoubtedly by wind, and old in- florescences with dehisced capsules often remain on the plants for some. time before the lignified tissues of the peduncles decay. Wood of Hydrangea is ring- or semiring-porous with small, typically 492 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 solitary vessels. Perforation plates and intervascular pits are scalariform, and both uniseriate and multiseriate rays are present; tracheids are often in definite radial rows. The epidermis, which commonly exfoliates, often contains tannin bodies, and raphide sacs containing mucilage are common in most plant tissues, particularly the leaves. Cork cambium is formed in the outermost part of the cortex. Foliage stomata in some investigated species are paracytic, but Reyenga & Karstens determined that stomata occurring in the lower epidermis of the calyx lobes of neuter flowers of H. macrophylla are actinocytic. These workers have also described the disintegration of the calyx stomata (a phenomenon apparently previously undescribed) during the growth of the calyx lobes that accompanies a “greening” phase at the termination of anthesis. During the “greening” period, the calyx lobes lose their characteristic coloration and become a dull, drab greenish-brown. Numerous species of Hydrangea, all in sect. HypraNcra and including both H. quercifolia and H. arborescens (particularly H. arborescens f. grandiflora Rehder, a form with all neuter flowers, originally found growing spontaneously in Ohio) are widely planted in temperate regions as orna- mentals. Most of the forms in cultivation were introduced to European and American gardens from Asia where the shrubs have been favorites in Chinese and Japanese gardens for centuries. Haworth-Booth (1950), McClintock (1956b, 1957a & b), and Nevling (1964) have reviewed the available information concerning the origins, culture, and history of the ornamentals. Hydrangea macrophylla is the much-planted garden shrub and the “Hortensia” of florists with large globose inflorescences of white to pink and red through blue and purple neuter flowers. This species is well known to gardeners as a plant that will change its flower color depending upon the pH of the soil in which it is grown. Chromatographic separations of sepal pigments have shown that in both red and blue forms of H. macro- phylla the pigments are identical, viz. a diglycoside of delphinidin contain- ing glucose and arabinose. Additional phenolic compounds identified from sepal extracts include neochlorogenic and chlorogenic acids, kaempferol-3- glucoside, kaempferol-3-galactoside, and a kaempferol triglycoside contain- ing galactose, glucose, and xylose (Asen et al., 1957). The color or shade of the stérile flowers is dependent upon the ability of the plant to absorb and accumulate aluminum, the concentration of anthocyanin present in the sepals, and the ratio of the anthocyanin to the kaempferol glycosides. Aluminum has a marked effect on delphinidin diglycoside and the kaemp- ferol pigments, causing marked shifts of absorption spectra toward longer wave lengths with a resulting shift from red to blue (Asen & Siegelman, 1957). Although the availability of aluminum is correlated with soil pH, the relationship between soil chemistry and sterile flower color in Hy- drangea must still be clarified. Included within the Saxifragaceae in the tribe Hydrangeae DC., subfam. Hydrangeoideae A. Br. by Engler, Hydrangea and its allied genera have more commonly been segregated from the Saxifragaceae and treated as < 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 493 comprising the Hydrangeaceae Dumort. The Hydrangeaceae itself is often further divided (essentially along Engler’s tribal lines) into the Hydran- geaceae sensu stricto and the Philadelphaceae (cf. under Philadelphus). Within the Hydrangeoideae, Hydrangea seems to be most closely related to Platycrater Sieb. & Zucc., a monotypic Japanese genus that differs only in its greater number of stamens, and Cardiandra Sieb. & Zucc., an East Asian genus of about five herbaceous species with alternate leaves. In inflorescence, flower, and fruit characters Cardiandra and Platycrater are essentially identical, and their relationships with Hydrangea should be investigated. The dried root of Hydrangea arborescens contains hydrangin, an alka- loid, and is used medicinally as a diaphoretic and diuretic. In Japan, the wood of Hydrangea has been used in carving various small objects, and a sweet beverage, amacha, is prepared from steamed and dried young leaves of H. Thunbergii. Several accounts of both human and livestock poison- ing from leaves have been reported in our area. REFERENCES: Under family references see BEAN; DippEL; ENGLER, 1891, te HAMEL, 1953; HutrcHinson, 1927, 1967; Ikuse; Jay, 1971; Li; Martin RE; Morr; meee LLEZ; REHDER; ROBERTSON ; SAX; SCHOENNAGEL; THOUVENIN: and Woo AGABABIAN, V. S. On the palynomorphology of the Hydrangeaceae. (In Rus- sian.) Izv. Akad. Nauk Armian. SSR Biol. 14(11): 17-26. pls. 1-7. 1961. ALLEN, R. C. Factors influencing the flower color of hydrangeas. Proc. Am. Soc. Hort. Sci. 28: 410-412. 1932. . Influence of aluminum on the flower color of Hydrangea macrophylla DC. Contr. Boyce Thompson Inst. 13: 221-242. 1943 AsEn, S., & H. W. SircetMANn. Effect of aluminum on absorption spectra of the anthocyanin and flavonols from sepals of Hydrangea macrophylla var. Merveille. Proc. Am. Soc. Hort. Sci. 70: 478-481. 1957. ———, & N. W. Sruart. Anthocyanin and other phenolic compounds in red and blue sepals of Hydrangea macrophylla var. Merveille. Proc. Am. Soc. Hort. Sci. 69: 561-569. 1957. BEAN, W. J. Hydrangeas. New FI. Silva 9: 109-119. pls. 31-34. 1937. Briguet, J. Le stigmate et la biologie florale des Hydrangea américains. Compt. Rend. Soc. Phys. Hist. Nat. Genéve 36: 38-43. 9. Burkett, G. W . Anatomical aes within the genus Hydrangea. Proc. Indiana Acad. Sci. 41: 83-95. 193 Cuenery, E. M. Are aka flowers unique? Nature 158: 240, 241. 1946. [Concerns color changes of sterile flowers. ] Coucn, J. F., & J. Nacusxt. Isolation of rutin from Hydrangea paniculata var. grandiflora Sieb, Jour. Am. Chem. Soc. 67: 1419. 5.* Duncan, W. H. Preliminary reports on the flora of Georgia — 4. Notes on the distribution of flowering plants including species new to the state. Castanea 15: 145-159. 1950. [Report of H. radiata and H. arborescens occurring together at several stations in Rabun County, 153.] 494 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Eaton, R. J. Hydrangea paniculata naturalized in Massachusetts. Rhodora 31: 18, 19. 1929. [Report of a thoroughly naturalized population near Lincoln, Massachusetts, in Middlesex County. | . Propagation by seed in Hydrangea. Ibid. 254, 255. 1929. [Seeds of H. paniculata collected from the Lincoln, Massachusetts, population were “normally viable.”’] HamBLIN, S. F. American hydrangeas. Nat. Mag. 6: 164-166. 1925. HARsIson, T. G. Notes on the genus Hydrangea. Am. Midl. . te 255-257. 1928. [Includes H. arborescens var. australis and H. Ashei a HawortH-Bootu, M. The hydrangeas. viii + 185 pp. pls. 1-21. bonds 1950. Hetice, H., & F. WEBER. Stomata-zahl vergriinter Hydrangea-Kelchblatter. Phyton: Austria 2: 299-301. 1950. Hotticx, A. A new fossil species of Hydrangea. Bull. Torrey Bot. Club 52: 21, 22. ph. 20.1075. Hom, T. Vegetative reproduction in Hydrangea. Rhodora 31: 120. 1929. [Holm suggests that the Lincoln, Massachusetts, population of H. paniculata could have developed through propagation by root-shoots. Cf. R. J. Eaton. ] IpraHim, R. K., & G. H. N. Towers. Studies of hydrangeol in Hydrangea macrophylla. Ser. II. Biosynthesis of hydrangeol from C14-labelled com- pounds. Canad. Jour. Biochem. Physiol. 40: 449-453. JeENTscH, R. Untersuchungen an den Sprossvegetationspunkten einiger Saxi- fragaceen. Flora 144: 251-289. 1956/57. Kato, N. On the variation of nodal types in the woody plants. 2. (A contribu- tion to the nodal anatomy. 2.) Jap. Jour. Bot. 42: 161-168. 1967. [Nodes of Hydrangea are generally trilacunar, except for nodes of H. involucrata, which are variously tri-, tetra-, and pentalacunar. ] Kincspury, J. M. Poisonous plants of the United States and Canada. xiii + 626 pp. cast Cliffs, New Jersey. 1964. [Hydrangea, 370, 371. Maxrmowicz, C. J. Revisio Hydrangeearum Asiae orientalis. Mem. Acad. Sci. St. airs VIT. 10(16): 1-48. bls. 1-3. 1867. McCurntocx, E. New combinations in Hydrangea. Jour. Arnold Arb. 37: 373, 374. 1956a. . The cultivated hydrangeas. Baileya 4: 165-175. 1956b. . A monograph of the genus Hydrangea. Proc. Calif. Acad. Sci. IV. 29: 147-256. 1957a. ———. Hydrangeas. Natl. Hort. Mag. 36: 270-279. 1957b. McFancanp, J. H. Three distinctive American shrubs. New Fl. Silva 4: 221- 224. 3 pls. 1932. [H. arborescens var. grandiflora, H. quercifolia, ae Gordonia (= Franklinia) Alatamaha.]| NEvLING, L. I., Jk. Climbing hydrangeas and their relatives. Arnoldia 24: 17-39. 1964 . GOmez-Pompa. A new Hydrangea from Mexico. Jour. Arnold Arb. 49: 225-232. 1968. PerriraZ, J. Biologie florales des Hortensias. Bull. Soc. Vaud. Sci. Nat. V. 47: 51-63. 1911. hes 4 F. Nouvelles recherches sur les rapports entre les insectes et les fleur tude sur le réle de quelques organes dits vexillaires. Mem. Soc Zool. ‘France 11: 339-375. 1898, [Observations sur l’Hydrangea opuloides Lam., 361-374. REHDER, A. Hydrangea. In: C. S. SarcENT, Pl. Wilsonianae. 1: 25-41. 1911. [Publ. Arnold Arb. 4.] 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 495 REYENGA, T. W., & W. K. H. Karstens. Changes in the epidermal pattern of the sepals af Hydrangea macrophylla (Thunb.) DC., “Otaksa” in their successive growth phases. Formation of new stomata, disintegration of old stomata. Acta Bot. Néerl. 13: 340-45 1964 St. Joun, H. A critical revision of Hydrangea arborescens. Rhodora 23: 203- 209. 1921 SIWECKI, R. The genus Hydrangea L. in Poland and in particular in the collec- tion of the Kérnik Arboretum. (In Polish; English and Russian summaries.) Arb. Kornickie 16: 35-61. 1971. SMALL, J. K. Hydrangea quercifolia: gray-beard. Addisonia 5: 29, 30. pl. 175. & P. A. Rypperc. Hydrangeaceae. N. Am. Fl. 22: 159-178. 1905. | Hydrangea, 159-161. | Tay Lor, L. A. Plants used as curative by certain southeastern tribes. Part I. Plants used medicinally. xi + 88 pp. Bot. Mus. Cambridge. 1940. TIEGHEM, P. vAN. Structure de quelques ovules et parti qu’on eu peut tirer pour améliorer la classification, Jour. Bot. Morot 12: 197-220. 1898. Witson, E: H. The hortensias: Hydrangea macrophylla DC. and Hydrangea serrata DC. Jour. Arnold Arb. 4: 233-246. ; 15. Decumaria Linnaeus, Sp. Pl. ed. 2. 2: 1663. 1763; Gen. Pl. ed. 6. 238. 1764 Climbing or trailing woody vines with adventitious aérial roots and deciduous [or occasionally persistent] foliage. Stems with grayish or reddish-brown bark, terete or slightly ridged from the enlarged nodes; epidermis and cortex of young-shoots exfoliating; pith round or polygonal in cross section, greenish white and spongy. Lateral buds sessile, small and solitary with densely hirsute, reddish-brown, indistinct scales; leaf scars opposite, transversely connected, V shaped to linearly elliptic, inclosing 3 conspicuous bundle scars. Leaves opposite (occasionally 3 leaves at a node on vegetative runners), petiolate, the petioles short with slightly winged, exstipulate bases; blades dark green and glossy, glabrous [or with scat- =e o o = jo) = 0 oak & o O. fore “= - oe a ° 4 a o oo o =_ ° Lo 1 io) io” E ro) o =) a © co a i — i=) jo) c = ee =] oO = oh = o a | =. Ln | & irregularly dentate, or irregularly lobed margins and acute [obtuse, or rounded] apices and cuneate, obtuse, or truncate bases; venation pinnate with a prominent midvein. Flowers perfect, faintly fragrant, arranged in compact [or loose and broad], convex, terminal, many-flowered corymbs; each flower of the corymb usually subtended by an inconspicuous, linear bract. Calyx of 7-10(—12) sepals, connate and adnate to the ovary, for- ming a strongly 7—10(—12)-ribbed floral cup (often with additional ribs observable) with 7—10(—12) deltoid calyx lobes projecting above and persistent in fruit. Corolla of 7-10(—12) creamy-white, sessile, lanceolate petals inserted on the rim of the floral cup, valvate in bud. Androecium of 20-30 stamens, inserted on the rim of the floral cup; filaments often con- torted, subulate, tapering only slightly to the basifixed, suborbicular to ovate anthers, which have obvious connectives. Gynoecium 7-10(-—12)- loculate, the ovary inferior, adnate to the floral cup, with numerous, tenu- 496 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 5 Ficure 10. Decumaria and Hydrangea. a-i, D. barbara: a, shoot with terminal corymb, X 1/4; b, detail of flowers — note radiating stigmatic lines, X 3; ¢, detail of anther showing obvious connective, X 12; d, vertical section of flower (petals and anthers removed) showing stylar column and axile placentation, X 6; e, cross section of ovary, X 12; f, immature capsule with persistent style and stigma, x 5; g, completely dehisced capsule, 5; h, seed 12; i, embryo ? ? by , ’ oriented as in seed, 25. j, H. arborescens subsp. arborescens: mature dehisced capsule, seen from above, < 10. inucellar, anatropous ovules pendulous on axile placentae: style 1, thick and elongated [or very short and the stigma essentially sessile], projecting rom the center of the ovarian disc and terminating in a knoblike, capitate stigma with 7-12 radiate stigmatic lines. Fruit a prominently ribbed capsule, terminated by the persistent stigma and dehiscing between the ribs (through the decay of the tissue) ; seeds small and numerous, the small cylindrical embryo surrounded by scanty endosperm of the a6 initio Cel- lular type and a loose, membranaceous, lustrous yellow, often caudate seed coat. Base chromosome number 14. Type sprctgs: D. barbara L. (Name 1972] SPONGBERG, GENERA OF SAXIFRAGACEAE 497 from Latin, decimarius, relating to tenths, in reference to the usually ten- merous flowers.) —- CLIMBING HYDRANGEA, WOOD-VAMP. A genus of two recognized and perhaps one undescribed species of east- ern Asia and eastern North America. Decumaria barbara,’® 2n = 28, a high-climbing or occasionally trailing vine with ovate to obovate leaves with acute apices, flowers in compact rounded terminal corymbs, and the gynoecia terminated by single, toplike styles with radiating stigmatic lines, occurs in rich deciduous forests and swamps from southeastern Virginia into Florida and westward to Louisiana and Tennessee. Essentially re- stricted to the Coastal Plain in Virginia and North Carolina, the species is more widely distributed in the southern part of its range, and it occurs at low elevations in mountain valleys in southwestern North Carolina, north- ern Georgia, and eastern Tennessee. Adequately distinct from the American species, Decumaria sinensis Oliver is a low-climbing, often branched vine with semipersistent or per- sistent oblong leaves with obtuse or rounded apices; flowers in loose, open corymbs; and gynoecia with very short styles or completely sessile stigmas. Restricted to central China, D. sinensis apparently grows over a relatively narrow altitudinal range (1000-1300 meters) in the mountains of western Hupeh and eastern Szechuan. Variation within D. sinensis occurs in eastern Szechuan, and collections of plants with obviously persistent foliage and greenish flowers comprising small, tight corymbs may represent a third species. A single chromosome count, 2” = 28, has been recorded for Decumaria barbara by Sax (1931). No information concerning floral biology has been recorded, although the fragrance of the flowers probably attracts flying insects as pollinators. Depending upon geographic locality, anthesis occurs from April through June, and examination of preserved flowers sug- gests that they are either proterandrous or homogamous. The light weight of the caudate seeds and the climbing habit of the plant strongly suggest that dispersal is by air currents. Release of ripe seed probably occurs over a prolonged period (beginning in the fall) since decay of the intercostal tissue is slow, and the mature capsules and inflorescences remain on the vines through the winter. Assignment of Decumaria to the tribe Hydrangeeae DC. in subfam. Hydrangeoideae is generally accepted, since the genera comprising the tribe appear to be closely allied. Decumaria shows strong affinities with Schizophragma Sieb. & Zucc. (two or three species of eastern Asia with reported chromosome numbers of 2” = 28 and 72) and Pileostegia Hooker f. & Thomson (two species of eastern Asia), and generic lines merit re-evaluation. As in Decumaria, styles in flowers of both Schizophragma and Pileostegia are united and terminate in large, capitate stigmas, and the dehiscence of the ribbed capsules is effected by the decay of intercostal * The specific epithet, barbara, is in reference to the Barbary Coast. The species was originally thought to be indigenous to Africa. 498 JOURNAL OF THE ARNOLD ARBORETUM [voL, 53 tissue. Members of all three genera are woody climbers with opposite leaves and paniculate or corymbose inflorescences, but in Pileostegia the foliage is consistently persistent, and the corolla consists of fused petals that form a caducous calyptra. In Schizophragma, unlike Decumaria, the androecium consists of only eight to ten stamens, and the corolla is composed of four or five petals. Although two species of Schizophragma have large, showy, neuter flowers with one greatly enlarged calyx lobe (cf. Hydrangea for discussion of neuter flowers), S. crassum Hand.-Mazz. lacks these and appears to link the genus closely with both Decumaria and Pileostegia. Although both species of Decumaria have occasionally been grown as ornamentals in botanic gardens and arboreta, their use outside of warm- temperate regions is limited because of winter-tenderness and high water requirements. No economic or other uses have been reported. REFERENCES: Under family references see BEAN; DrppEL; ENGLER, 1891, 1928; HAMEL, 1953; HUTCHINSON, 1927, 1967; Jay, 1971; Lr; REHDER; SAX; THOUVENIN; and Woon. Contr. U.S. Natl. Herb. 5: x + 321-585. pls. 65-77. map. 1901 [Decu- maria, 416, 449, 454.] Neviinc, L. I, Jr. Climbing hydrangeas and their relatives. Arnoldia 24: 17-39. 1964. ARNOLD ARBORETUM HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 02138 1972] ANDERSON, STUDIES ON BIGELOWIA, II 499 STUDIES ON BIGELOWIA (ASTERACEAE), II. XYLARY COMPARISONS, WOODINESS, AND PAEDOMORPHOSIS Loran C, ANDERSON * MAJOR TRENDS IN XYLEM evolution discovered through the efforts of I. W. Bailey and his contemporaries have become generally accepted (see Carlquist, 1961; Esau, 1965; Fahn, 1967, for review). The apparent uni- versality of the xylem trends has been of great comfort to systematists in their application of anatomic data to phylogenetic considerations. How- ever, the concepts were developed by comparative studies, particularly among the Dicotyledoneae, in which the secondary xylem of woody shrubs and trees was used almost exclusively. This basis may provide an in- complete, perhaps biased, view of xylem evolution, as suggested by Carl- quist (1962) in his study of juvenility in rosette trees and woody herbs. He concluded that woods from such growth forms may not be comparable to those from woody shrubs and trees. His further studies (1969a, 1969b, 1970a, 1970b) were directed toward understanding relationships between woody anatomy of herbaceous and woody species of angiosperms with special emphasis on insular forms. Cumbie (1963) reported in detail on cambial activity in Hibiscus lasiocarpus; nevertheless, little more has been done to compare continental herbaceous taxa with related woody ones. The genus Bigelowia of the Astereae provides an excellent subject for such a study. These herbaceous rosette plants are frequently monocarpic (particularly so in the greenhouse), and appreciable amounts of secondary xylem are formed only in the caudex. The plants perennate through rhizomaty (in B. nuttallii) or produce offsets from the small caudices. (The caudex usually represents one year’s growth; two or three growth rings develop only rarely.) Herbaceous forms in the Astereae are prob- ably derived. Carlquist (1960) concluded his survey of the wood anatomy of the tribe with: “the evidence in Astereae, far from countering the well- established woody-to-herbaceous trend in dicots, seems to support it.” Data on chromosome numbers reinforce that view (Raven, et al., 1960; Solbrig, et al., 1964, 1969). Near relatives of Bigelowia appear to be Euthamia and other woodier taxa in the Haplopappus alliance (Anderson, 1970); therefore, I used these plants to examine evolutionary trends in a phylad containing woody and herbaceous taxa. MATERIALS AND METHODS Living caudices and flowering stems of all taxa in the genus were collected in the field by me or supplied by R. K. Godfrey, S. B. Jones, O. Lakela, or * Contribution no. 1143, Division of Biology, Kansas Agr. Expt. Station, Manhat- tan, Kansas 66506. This study was supported by National Science Foundation grant GB 8747 TABLE 1. Comparative xylary features from caudices of Bigelowia Vessel elements ; Fibers : g F m rt 2 g o i roe 3 os & a a re) pee tl a ee ee eS oO Fa ty a “= 2 ; i rv) ee %, 3) od ° 8 2 36 2 +o) & 5) Taxon and collection data 5 5 g z e e ef 5 g @ Hi g g E +. 8 Re 3 a . . oa Re Ge 5 ; ° 83 B. NUTTALLII Anderson 3173, 1 3.02 0.88 24.6 18.0 48.3 32 3-5 2.9 13.3 221.4 4.4 razos Co., Texas Anderson 3089, 1 Se 0.76 29.1 17.6 64.4 3. 4-5 5.2 14.7 199.1 4.5 Turner Co., Ga. Godfrey 63191, 1 3.78 0.88 29.1 20.5 76.1 2.9 4-6 4.3 13.5 180.0 4.5 Washington Co., Fla Thieret 25200 Z 5.80 2.02 44.8 26.1 84.2 3.5 4-5 3.8 Piz 192.6 S38 Vernon Par., La Thieret 25186, 2 6.04 2.14 42.6 26.9 124.3 Sel 3-5 4.1 12.6 172.7 3.9 Natchitoches Par., La. Anderson 3464, 2 8.00 215 67.2 31.4 141.3 4.2 4-5 4.1 17.8 362.1 5.1 Pinellas Co., Fla. 00s WOLAYOdUY GIONUV AHL JO IVNUNOL ¢§ “ToA] B. NUDATA ssp, NUDATA Jones 11071, Forrest Co., Miss. Anderson 3088, Anderson 3263, Onslow Co., N.C. B. NUDATA ssp. AUSTRALIS Anderson 3471, Pinellas Co., Fla. Lakela 30620 Collier Co., Fla. 1.26 23.4 26.8 23.5 27.6 25.4 28.9 118.7 4-7 4-7 274.4 238.9 193.6 286.7 221.4 271.4 II ‘VIMOTADIA NO SAIGNLS ‘NOSUAANV [ZL61 TOS 502 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 J. W. Thieret, whose help is gratefully acknowledged. The larger caudices from each population sample were selected to insure greater xylem de- velopment. Material also was taken from the base of the flowering stem and just below the inflorescence for qualitative and quantitative compari- sons with the caudex. Tissues were preserved in FPA, softened in HF, and processed further to prepare slides for examination as in earlier studies (Anderson, 1963; Anderson & Reveal, 1966). Tabular data on cell sizes were obtained by averaging 50 measurements; vessel pit diameters were based on fewer measurements, Each entry in the tables represents a single plant, although more than one was processed for certain collections. Voucher specimens are deposited in the Kansas State University Herbarium (Ksc). Addi- tional information on morphology for most collections used in this study is found in Anderson (1970). RESULTS Xylem production in Bigelowia is not extensive; radial development ap- parently does not exceed 3 mm. Description of the woods (TABLE 1, Ficures 1-12, and in the text) is based on study of the caudex. TABLE 2 compares tracheary elements from different regions of the plant body. Vessels. Average diameter and length of vessel elements in Bigelowia are relatively smaller than in most Astereae (Anderson, 1963; Anderson & Reveal, 1966; Carlquist, 1960). The presence of vascular tracheids (practically indistinguishable from narrow vessel elements in transection) can aggravate determination of the average diameter; fortunately they are not common in Bigelowia. Most vessels are round in transection; angular ones (fairly typical for Astereae) are infrequent, particularly in B. nudata. The relation of wall thickness to diameter in vessels ranges from 11.5 to 17.7% in B. nuttallii: 8.3 to 11.1% in B. nudata ssp. nudata; and 6.9 to 8.3% in B. nudata ssp. australis (cf. TABLE 1). The vessel elements all have simple perforation plates. Their end walls generally are transverse to somewhat oblique in orientation; a few of the natrower ones are caudate. Intervascular pits are alternate and have cir- cular to elliptic borders (FicurE 10); those of B. nudata ssp. mudata are particularly large for Astereae. Helical sculpture in vessel walls is absent or restricted to fine grooves in Bigelowia. Amount of vessel grouping is varied in Astereae. Bigelowia vessels occur Singly or in single radial chains (TaBLE 1; Figures 1—5 ) as in Euthamia leptocephala (Anderson, 1963, as Solidago leptocephala). However, in Chrysothamnus, a group once thought to be part of Bigelowia, vessels are in large, tangentially spread clusters (often more than a dozen per cluster). Libriform fibers. Dimensions of fibers in Bigelowia are most consistent with the average size for the tribe. The relation of wall thickness to fiber width is from 28.7 to 33.3% in B. nuttallii: 16.9 to 26.4% in B. nudata ssp. nudata; and 12.3 to 15.6% in B. nudata ssp. australis. In Bigelowia, fibers 1972] ANDERSON, STUDIES ON BIGELOWIA, II 503 are nonseptate; most retain active protoplasts, resulting in darker staining (Ficures 7-9). Carlquist (1969a) reported nucleated, septate fibers in the lobelioids. Presence of living fibers, considered an evolutionary advance, may be associated with a reduced amount of axial parenchyma in the xylem (Esau, 1965) or may be an expression of xeromorphy (Fahn, 1964). Fiber dimorphism with some fibers gelatinous occurs in some Bigelowia populations (more frequently in B. nudata). Gelatinous fibers can be determined from the amount of shrinkage in the secondary wall, differential staining, and lack of birefringence under polarized light. The gelatinous fibers are similar to the more typical wood fibers in all other ways (i.e., dimensions, original wall thickness, and living state). Axial parenchyma. Paratracheal parenchyma is scanty and abaxially vasicentric, which is typical for Astereae and the family (Carlquist, 1960, 1966). Apotracheal parenchyma is present (weakly) as initial bands only in the B. nuttallii samples with growth rings (FIcuRE 4). Growth rings. Few caudices live more than a year, those that do, possess growth rings which are not strongly defined. Three years’ growth in one caudex of my collection, Anderson 3464, is the greatest extent of wood found in the genus; other plants of that population had two years’ growth (cf. TaBLe 1). The wood is diffuse porous (FicurEs 3-5), and vessel diameter shows only slight fluctuation associated with annual ring formation. Fibers produced at the close of the first growing season, how- ever, are noticeably smaller than those produced at the beginning of the second season. Quantitative and qualitative changes in the rays also ac- centuate the demarcation line (FicurEs 4, 12). Wood rays. The very nature of the radial system seems to depend on a predisposition of the caudex to be annual or perennial. Differences be- tween rays of one-year-old and two-year-old samples are striking. Such dichotomy also is evident in the general xylary growth pattern in the caudex (see below). Rays of one-year-old caudices of all taxa always contain vascular traces to adventitious roots and basal leaves (FicurES 6-8, 11). These rays possibly are merely extensions of gaps in the primary vascular system. Parenchyma surrounding the traces in the ray is loosely oriented, similar to the ground tissue of the pith, and mostly cuboidal to vertically elon- gated when seen in radial view. Transfusion cells (FicuRE 11) similar to those in Brighamia (Carlquist, 1962) are present. The vascular traces and associated parenchyma hardly seem products of the cambium. In a sense then, the wood could be termed rayless. Wood in rhizomes of Bigelowia nuttallii (the only species having them), including the single two-year-old rhizome found in Anderson 3173 (F1GURE 5), is rayless. Rayless rhizomes have been reported for Euthamia leptocephala (Anderson, 1963). Woods from perennial caudices have relatively greater ray development. However, very few of the rays have included vascular traces; most have none. The rays are multiseriate, averaging four to six cells in width vee’ » re se » es @¢ ceet it ee e88 PRES wees " oe STS ES As ee eS Re gh Se GREENS ‘ae ae @* ge @ oh ys SU GH 2 ‘. ‘ ea ¥eee ng re" + eee: par, ean & * # te bee ty tie Be ae, Th v4 fie igi] St ow, $y, has irr : Wie » aioe? oO. ) ttr é a * i : P we , ge : ‘ =) @ P . ¥s * er rar™- es oF oe r oa ; in ard base of page except ige Transections of B Figures 1 =: ° —- = Pa 3 Q let ii, two-year caudex (Anderson 3464), s xylem; 5, B. nuttallii, ) izome (Anderson 3173 two-year rh > Xx 83 in first year g me ; australis (Lakela 30620); 8, same from rial of B. nuttallii (Godfrey 63191); - yf year-old sample of B. nuttallii ( per roerd 3464) with tall wood rays. All, 506 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 #,.® ‘ feseey yg: Oe gi TY Fe hee a ceellgs "4 FicurEs 10-12. 10, Bigelowia nudata ssp. australis, vessel elements with large pits (Lakela 30620), X 200; 11, radial section of one- year-old caudex of B. nuttalli: (Anderson 3089) with pith toward left side of page showing “ray’ consisting of vertically oriented parenchyma, transfusion ral (arrows), an ee See p B. nudata 0 B. nuttallii ——e 400- 300+ £ teen... oral : ns libriform fibers 200- > 100+ Bia Ee age © vessel elements Z ta See ee ee Ficure 13. Comparison of average tracheary element lengths from different zones af the same plant. Ordinate, cell length in microns; abscissa igen distance enter em or caudex in microns. Zone A, taxylem of upper stem (fibers absent); B, secondary xylem of upper stem; C, seco lower al stem; D, seconda f caudex. Plants studied, Anderson 3088 (Bigelowia nudata) and Anderson 3089 ee sak al were growing within a few meters of one another. nuttallii was on ry soil of sandstone outcrops; whereas B. mide was in a depression filled with peat moss at the margin of the outcrop. (Ficures 4, 9), and ranging from 0.6 to 1.5 mm. tall (the taller ones being in B. nuttallii). Ray parenchyma is predominantly procumbent (FicuRE 12) rather than being vertically oriented as are most ray cells of one-year- old caudices (FicurE 11). New rays, initiated at the beginning of each suc- cessive year of growth (FicurEs 2, 14), are wider in the spring wood than in summer wood. The propensity for raylessness in Bigelowia woods is also suggested by the increased production of rays after the initial year’s growth in the perennial caudices. The ontogenetic relationships between early rayless wood and later wood, in which rays had developed, was noted by Barghoorn (1941) GROWTH PATTERNS Study of growth forms should involve measurement of variability in tracheary elements among different organs of the plant and also determi- vascular trace, X 80; 12, radial section of three-year-old caudex of B. nuttallit (Anderson 3464) showing absence of wood rays in initial year’s xylem td and their increased abundance in each a sath ring, rays composed of predominantly procumbent cells, « 4 TABLE 2, Xylem data from different tissues in Bigelowia 2 Vessel elements Fibers =| x 3 ty g : we 2 ee we 3 Taxon = a § = a s : rl Z a ee ee = 2 v >) os eo v 2) 2 collection sample 2, I 3B a ee & ®¢ = ag < Ss Oo se FS (& #8 ch Sts 48 B. NUTTALLII Anderson 3173 high stem 1.38 0.20 30.2 22.1 2.4 185.3 3-4 2.3 1235 3/ 349.4 low stem 2.22 0.40 31.4 235 31 146.5 3-4 3.2 13.0 37 243.7 caudex 3.02 0.88 24.6 18.0 52 48.3 3—5 2.9 13.3 4.4 221.4 rhizome 1.93 0.52 38.1 Zit Bed! 94.8 3-4 2.0 11.4 3.4 347.2 Anderson 3089 high stem 1.50 0.14 40.3 30.6 3.4 246.4 45 ao 14.4 4.4 423.3 low stem 2.00 0.19 38.1 23.4 3.4 204.6 455 4.9 14.3 4.6 401.0 caudex 3.52 0.76 29.1 17.6 3.1 64.4 4-5 52 14.7 4.5 199.1 Godfrey 63191 high stem 1.80 0.32 44.8 30.2 3.0 177 4-5 3.4 12.5 4.6 342.1 low stem 1.92 0.38 29.1 213 2.9 131.8 4-5 53 135 4.4 291.2 caudex 3.78 0.88 29.1 20.5 2.9 76.1 4-6 4.3 13.5 45 180.0 80S WOLAYOTAV GIONUV AHL AO TYVNUNOL ¢¢ “I0A } B. NUDATA ssp. NUDATA Anderson 3088 high stem low stem caudex Anderson 3070 high stem ow stem caudex B. NUDATA ssp. AUSTRALIS Lakela 30620 caudex 267.7 198.0 #2.3 314.8 190.4 86.3 313.6 189.3 79.2 [ZL6T II ‘VIMOTADIA NO SAIGNLS ‘NOSUAANY 60S 510 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 nation of patterns of change in element length during growth in a given organ. Variability among different organs or plant regions is tabulated in TABLE 2. Growth patterns, as depicted by average tracheary element lengths in different plant zones, are shown for Bigelowia nuttallii and B. nudata in FicuRE 13; patterns of change during growth in element length in the caudex are shown in FicureE 14. Considerable dissimilarity might be expected in tracheary elements from different regions of the plant when their ontogeny and environment are examined. The aérial (flowering) stem grows rapidly in a single season and is in a more xeric environment than the caudex or rhizome. The high stem (cf. TABLE 2), just below the branches of the inflorescence, has limited cambial activity; the secondary xylem (zone B of Ficure 13), less abun- dant than the primary xylem, contains vessel elements and fibers, fibers being absent in the primary xylem (zone A, Ficure 13). The low stem zone, arbitrarily taken 2 cm. above the caudex has little secondary xylem {zone C, Ficure 13), but more than does the high stem. The caudex grows more slowly and accumulates the greatest amount of xylem in the plant (zone D, Ficure 13). The rhizome of Anderson 3173 (FIGURE 5) has wider vessels but narrower fibers than do other plant regions. In all taxa, vessel members from the aérial stems are generally wider and longer than those of the caudices, whereas fibers are narrower but longer. Vessel wall thickness is rather uniform throughout different zones of the same plant, although the vessels of the caudices, being narrower, have relatively thicker walls. The assumption that the caudex is more or less predestined to be either annual or perennial is reinforced in the growth patterns of the secondary xylem of both species (FicurE 14). The fusiform cambial initials of the annual samples and, consequently, the vessel elements they produce be- come progressively shorter during the growing season. Note that shorten- ing is not so pronounced in the caudices of Anderson 3088 and Anderson 3089 (Ficure 14) as in the different zones of the same plants (F1cuRE 13). Cambial initials and vessel elements of the two-year-old caudices in FIGURE 14, on the contrary, become successively longer during the first growing season, and the trend is continued in the second season. In all samples studied, the terminal growth of the two-year-old caudices contained longer vessel elements than did the one-year-old caudices (Tasie 1). Libriform fibers do not necessarily indicate the size of the cambial initials that pro- duce them because fibers exhibit varying amounts of intrusive growth during differentiation and maturation. DISCUSSION Trends in xylem evolution in Asteraceae seem rather closely controlled by environmental factors and paedomorphosis, i.e., shifts in wood anatomy related to change in lengths of cambial initials (Carlquist, 1966). Dis- cussion of these factors is an appropriate prelude to taxonomic consider- ations. B. nudata sss 0 q B. nuttallii ——* U T T T T th T T T T T 0.0 0.5 1.0 15 2.0 25 3.0 Ficure 14. Average vessel element length from gesbiiage positions within the caudex. ges inate, element in a raat abscissa, distance from caudex center in millimeters. Zon taxylem; B, first- tanned x ndary xylem C, last-formed ee lem of “ar st ee s growth (total growth for some) ; : D, last- e ond year’s growth. Not shift in growth pattern between annual a gee eg caudices during first year’s growth. Ecological anatomy. Fragmentary observations on the ecology of Bigelowia (Harper, 1911; Anderson, 1970) indicate that the general en- vironments of the taxa are similar. The range of B. nuttallii extends farther west (into Texas) than the range of B. nudata. (That western zone is some- what warmer and drier, as suggested from mean annual temperature and rainfall data.) The species are closely sympatric at one locality in Florida (Harper, 1911) and at scattered stations in southern Georgia, where they differ in habital preferences. The soils on rock outcrops where B. nuttallit grows are usually reddish-brown sandy loams with little organic matter (0.5-2.5%) and low cation exchange capacity. Soils where B. nudata grows are usually gray sandy oe with more organic matter (2.0-5.0% ) and greater cation exchange capaci The sunny, open sites where Pigelowla nuttallii grows on shallow, better- drained soils are definitely more xeric than those where B. nudata grows. Considerable xeromorphy is shown in root and leaf anatomy (Anderson, unpublished) of B. nuttallii and is reflected in these anatomical features of the wood: 1) narrow, frequently short vessel elements with thick walls and narrow pits; 2) a number of vessels per group; 3) a number of vascu- lar tracheids; and 4) narrow, short fibers with thick walls. The greater amount of living fibers might also reflect xeromorphism when compared to that of B. nudata, but the taller wood rays in the perennial caudices certainly do not. Helical sculpture on vessel walls tends to be more abun- dant or elaborate in xerophytic composites (Carlquist, 1966), but no taxa of Bigelowia have much helical sculpture. 512 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Differences in wood anatomy among different zones of an individual plant may relate to xeromorphy or to patterns of growth involving paedo- morphosis. For example, longer tracheary elements in the rhizomes than in the caudex of Bigelowia nuttallii (TABLE 2) probably reflect more mesic conditions in the underground as opposed to above-ground habitats. The production of longer vessel elements in higher zones of the aérial stem (FicureE 13) could relate to juvenilistic growth patterns (paedomorphosis) rather than to influences of the external environment. Carlquist (1970a) suggested this was the case in Echium pininana. Woodiness vs. paedomorphosis. Carlquist (1962), in promulgating his theory of paedomorphosis, pointed out that rosette trees and woody herbs commonly have different growth patterns than do woody plants. The shift in the growth pattern through loss of the marked reduction in cambial fusiform-initial length at the onset of secondary growth results in second- ary woods with longer vessel elements than found in related, truly woody plants (Ficure 15, adapted from Carlquist, 1966). types of = : . = paedomorphosis = a2 S a i normal (woody) curve rs “ “ a2 > aay secondary xylem Ficure 15. Patterns of change in average vessel element length during growth of plant (adapted from Carlquist, 1966). Such patterns are discomforting to phylogenists. Woody plants, being generally more primitive, generally have longer cambial initials and, there- by, longer vessel elements than do related, less woody plants. Yet through paedomorphosis, a derived herbaceous plant might have longer vessel elements than related more woody taxa. f Carlquist (1962) suggested that woods influenced by paedomorphosis cannot be compared with those from woody plants because “‘phylesis both toward and away from herbaceousness would involve juvenile expressions, so these could not be used as criteria of one ancestry or the other.” /f this proves universally true, then woods of woody and herbaceous mem- bers of a genus cannot be compared. However, these studies in Bigelowia 1972] ANDERSON, STUDIES ON BIGELOWIA, II $13 show that both juvenilistic and woody patterns are found in different plants of the same species (FicurRE 14). Annual caudices in Bigelowia show influence of paedomorphosis through their flattened growth curve (somewhat similar to pattern C of FIGURE 15); the resulting shorter vessel elements, compared with those of the perennials, represent a deviation from the patterns of paedomorphosis as Carlquist described them. These additional characters of the annual caudices also indicate paedomorphosis: pits with wide apertures on vessel elements, presence of transfusion cells, tendency toward raylessness with rays containing vascular traces, and erect ray cells. The normal (woody) pattern of change in vessel element length is found in the perennial caudices. In addition, woodiness as opposed to paedomor- phosis is indicated by longer vessel elements, absence of transfusion cells, and “typical” wood rays with predominantly procumbent ray cells. In the case of Bigelowia, a trend from woodiness to herbaceousness and paedomorphosis is apparent and is related to the shortening of the life of the caudex. Phylogenetic considerations. Bigelowia nuttallii and B. nudata are closely related; artificial, sterile hybrids between the species have been produced (Anderson, unpublished). If one species were to be considered more primitive, based on data from wood anatomy, it would be B. nuttallit which has generally longer vessel elements (even though probably shorten- ed through xeromorphism) ; also, a greater number of perennials have been found in that species. Woodiness may appropriately be designated a primitive feature for Bigelowia, as apparently it is for the tribe (Carlquist, 1960). Euthamia has been suggested as a close relative of Bigelowia (Anderson, 1970). Pat- terns of change in vessel element length during growth have been deter- mined for plants of E. graminifolia and E. leptocephala (Anderson, un- published); interestingly, both have a normal (woody) growth curve. More than a quarter of a century ago, Tippo (1946), in discussing the role of wood anatomy in phylogeny, stated: The next logical development which may be expected in this field of endeavor is a series of exploratory investiga- tions of the herbaceous Dicotyledons whose purpose will be not only to uncover new phylogenetic sequences, but to extend the lines of specialization already established into the relatively virgin territory. There is every reason to suppose that the trends already defined will be found to prevail in the secondary wood of the herbs, and there is also some basis for belief that similar sequences will be unearthed in the primary xylem. 514 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Knowledge of wood anatomy in herbaceous plants is still so limited that we do not know how universal paedomorphosis is or what other patterns may be found. Certainly they will not parallel the trends already established for woody dicots, as Tippo supposed they would. LITERATURE CITED AnvErSON, L. C. 1963. Studies on Petradoria (Compositae): anatomy, cytology, taxonomy. Trans. Kans. Acad. Sci. 66: 632-684. 1970. Studies on Bigelowia (Astereae, Compositae). I. Morphology and taxonomy. Sida 3: 451-465. . L. REVEAL. 1966. Chrysothamnus bolanderi, an intergeneric hydrid. Madrono 18: 225-233. : BarcHoorn, E. S. 1941. The ontogenetic development and phylogenetic special- ization of rays in the xylem of dicotyledons. III. The elimination of rays. Bull. Torrey Bot. Club 68: 317-325. CarLquist, S. 1960. Wood anatomy of Astereae (Compositae). Trop. Woods 3: 54-84. 1961. Comparative Plant Anatomy. Holt, Rinehart, & Winston, 1962. - theory of paedomorphosis in dicotyledonous woods. Phyto- elon Me 12: 30-45. . 1966. Wood anatomy of Compositae: A summary, with comments on factors controlling evolution. Aliso 6: 25-44. Wood anatomy of Lobelioideae (Campanulaceae). Biotropica 1: 47-72. . 1969b. Wood anatomy of Goodeniaceae a the problem of insular woodiness. Ann. Missouri Bot. Gard. 56: 3538— 970a. Wood anatomy of Echium (Giomisiatdees Aliso 7: 183-199. . 1970b. Wood anatomy of insular _— . ibaa and the problem of raylessness. Se Torrey Bot. Club 97: 3 Cumpig, B. G. 1963. The vascular cape = ir development in Hibis- cus ERE ‘Am. Jour. Bot. 50: 9 Sl. Esau, K. 1965. Plant Anatomy. ed. 2. ets Wiley & Sons, New York. Faun, A. 1967. Plant Anatomy. Pergamon Press, New York. Harper, R. M. 1911. Chondrophora [Bigelowia| virgata in west Florida. Torreya 11: 92-98. Raven, P. H., O. T. Sotpric, D. W. Kynos, & R. SNow. 1960. Chromosome numbers in Compositae. I. Astereae. Am. Jour. Bot. 47: 124-132. Sotsric, 0. T., L. C. ANpErson, D. W. Kynos, P. H. RAvEN, & L. RUDENBERG. 64. Chromosome numbers in Compositae. V. Astereae II. Am. Jour. Bot. 51: 513-519. Sotsric, O. T., L. C. ANDeRson, D. W. Kynos, & P. H. RAVEN. 1969. Chromo- some numbers in Compositae. VII. Astereae III. Am. Jour. Bot. 56: 348-353. Tippo, O. 1946. The role of wood anatomy in phylogeny. Am. Midl. Nat. 36: 3 12. Division or BioLocy Kansas STATE UNIVERSITY MANHATTAN, KANSAs 66506 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, III 515 NOTES ON WEST INDIAN ORCHIDS, III * Lestige A. GARAY AND HERMAN R. SWEET IN PREPARING THE generic descriptions of orchids for the projected Flora of the Lesser Antilles by Dr. Richard A. Howard of Harvard University, it became necessary to typify each generic name reported from that area in order to arrive at a correct circumscription. During our studies we have found that some of the early lectotypifications are incorrect due to random choices. Indirect lectotypifications through the removal of specific epithets by transfers have also been ignored. We have compared and have brought each known typification into line with the principles expressed in the “Guide for the determination of types,” published in the /nternational Code of Botanical Nomenclature, As a result of this procedure we are pre- senting with sundry comments a list of typifications of those generic names of orchids (whether currently accepted or synonymous) which are re- ported from the West Indies, the Greater Antilles and the Lesser Antilles, excluding Trinidad, Tobago, and the off-shore continental islands of northern South America. Achroanthes Raf. in Am. Monthly Mag. Crit. Rev. 4: 195. Jan. 1819. Type: Malaxis ophioglossoides Willd. OBSERVATION: This name is based on Malaxis subgenus Microstylis Nutt. he name was rejected in 1954 in favor of Microstylis (Nutt.) A. Eaton. It is a synonym of Malaxis Soland. ex Sw. Acraea Lindl. in Benth. PI. scales 155. Aug. 1845. Type: Acraea parvifolia Lindl OBsERVATION: It is a synonym ‘of Pterichis Lindl. Adnula Raf. FI. Tellur. 2: 87. Jan.—Mar. 1837. Type: Satyrium adnatum Sw OBSERVATION: A superfluous name based on the type of the genus Pelexia Poit. ex Lindl. It is a synonym of Spiranthes L. C. Ric anes Hoffmsgg. Verz. ees 20. 1842. PE: Ophrys Loeseli One. This is an n illegitimate name and a synonym of Liparis L. C. Rich Amalia Rchb. Herbarienb. Nomencl. 52. no. 1974. July 1841. LectotyPe: Bletia grandiflora La Llave & Lex. OssErRVATION: This is a new name for Laelia Lindl. 1831, not Adanson 1763. The name was rejected in 1935; it is a synonym of Laelia Lindl. * In this series, No. I was published in the oreags of the Arnold Arboretum 50; 462-468. 1969; UNO. Il, bid. 53: 390-398. 1972 516 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 fig ages et Lindl. ex Stein, oe apa 66. 1892. E: Epidendrum secundum pial cas It is a synonym of pace oe L. pemssiangere Hoffmsgg. Verz. tie 21. 1842. Epidendrum cochleatum pan ects It is'‘a synonym oe Epidendrum L. Arpophyllum La Llave & Lex. Nov. Veg. Descr. Orch. Opus 2: 19. 1825. Type: Arpophyllum spicatum La Llave & Lex. Auliza Salisb. ex Small, FI. eee 55. Apr. 26, 1913. Type: Epidendrum nocturnu OBSERVATION: It is a oles ei Epidendrum L. Aulizeum Lindl. ex Stein, Orchideenbuch 84. 1892. Type: Epidendrum ciliare L. OBSERVATION: It is a synonym of Epidendrum L. apo slate Schltr. in Fedde bne e Sp. Nov. 17: 77. Apr. 1, 1921. : Bletia sarcophylla Rchb Beadlea Small, Fl. SE. U.S. 319. July 1903. Type: Spiranthes Storeri Chapm. OBSERVATION: It is a synonym of Spiranthes L. C. Rich. Bletia Ruiz & Pav. Fl. Peru. Prodr. 119. ¢.26, Oct. 1794. Lectotype: Bletia catenulata Ruiz & Pav. OBSERVATION: Lectotype was chosen by Britton & Millspaugh, Bahama Flora 96. 1920. Brachionidium Lindl. Folia Orch. Brachionidium 1. Mar. 10, 1859. Lectotype: Restrepia parvifolia Lindl OBSERVATION: Lectotype was chosen by Garay in Canad. Jour. Bot. 34: 729 1956. Brassavola R. Br. in Ait. Hort. Kew. ed. 2. 5: 215. Dec. 1813. Type: Epidendrum cucullatum L. Brassia R. Br. in Ait. Hort. oo ed. 2,5: 215. Dec. 1813. Type: Brassia maculata R. Broughtonia R. Br. in Ait. Hort. Kew. ed. 2. 5: 217. Dec. 1813. Type: Epidendrum sanguineum Sw. Bolbophyllaria Rchb. f. in Bot. Zeit. 10: 934. Dec. 31, 1852. Type: Bulbophyllum bracteolatum Lindl. OBSERVATION: This is a synonym of Bulbophyllum Thou. Bulbophyllum Thou. Hist. Pl. Orch. hag tabl. esp. Mar. 11, 1822. LectotyPe: Bulbophyllum nutans Th OBSERVATION: The genus was nen in 1905. The lectotype was chosen by Green, Prop. Brit. Bot. 100. 1929. 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, III 517 Burlingtonia Lindl. in Bot. Reg. 23: #. 1927. Jan. 1, 1837. Type: Burlingtonia candida Lindl. OBSERVATION: This is a synonym of Rodriguezia Ruiz & Pav. Calanthe R. Br. in Bot. Reg. 7: sub ¢. 573. Oct. 1, 1821. Type: Orchis triplicata Willem. OBSERVATION: The genus was conserved in 1905. Calopogon R. Br. in Ait. Hort. Kew. ed. 2. 5: 204. Dec. 1813. Tyre: Limodorum tuberosum OBSERVATION: The genus was conserved i in 1905. Camaridium Lindl. in Bot. Reg. 10: t. 844. Nov. 1, 1824. Type: Camaridium ochroleucum Lindl. OBSERVATION: This is a synonym of Maxillaria Ruiz & Pav. Campylocentrum Benth. in Jour. Linn. Soc. Bot. 18: 337. Feb. 23, 1881. Tyre: Todaroa micrantha Rich. & Gal Cgenvation: This is a new name io Todaroa Rich, & Gal. 1845, not Parla- tore 1843 Carteria Small in Torreya 10: 187. Aug. 1910, not Diesing 1866. Type: Carteria corallicola Small. OBsERVATION: This is a synonym of Basiphyllaea Schltr. Cattleya Lindl. Coll. Bot. ¢. 33. 1821. Type: Cattleya labiata Lindl. Cattleyopsis Lem. Jard. Fleur. 4: Misc. p. 59, no. 165. Aug. 1, 1853. Type: Cattleyopsis delicatula Lem. Centrogenium Schltr. in Beih. Bot. Centralbl. 37(2) : 451. Mar. 31, 1920. Lectotyre: Neottia calcarata Sw. OBsERVATION: Lectotype was chosen by Correa in Darwiniana 11: 81. 1955. This is a synonym of Eltroplectris Ra Chloidia Lindl. Gen. and Sp. Orch. Pl. 484. Sept. 1840. Lectotype: Serapias polystachya Sw OBsERVATION: Lectotype was chosen indirectly by Cogniaux in Martius FI. Bras. 3(4): 276. 1895. This is a synonym of Tropidia Lindl. Chloraea Lindl. in Quart. Jour. Sci. 23: 47. Mar. 1827. Type: Cymbidium luteum Willd OBSERVATION: The genus Chioraea was first suggested by Lindley in his Collec- tanea Botanica Appendix No. 12. Jan. 1826. He clearly indicated that he was basing it on “Epipactis &c Gavilu, Feuillée.’” This reference is applicable only to two plants depicted by Feuillée in Jour. Observ. Bot. 2: 1714; the first on ¢t. 18 as Epipactis flore albo vulgo Gavilu, and the second on ¢. 20 as Epipactis amplo flore luteo, vulgo Gavilu. In describing the genus Chloraea, Lindley included among the nine species only one of Feuillée’s plants, namely the second polynomial mentioned on plate 20, calling it Chloraea Gavilu Lindl. Unfortunately this is an illegitimate name for Lindley sce eo eet luteum Willd., also based on the same polynomial, in synon 518 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The recent choice of Cymbidium virescens Willd. for a lectotype by Correa in Flora Patagonica 7(2): 200. 1970, is incorrect. — L. C. Rich. ex Lindl. Gen. and Sp. Orch. Pl. 409. Feb. 1840. : Limodorum grandiflorum Aubl. a... ei was originally suggested by L. C. Richard, Orch. Eur. Annot. 9. 1817 Cochleanthes Raf. FI. Tellur. - ‘@ (July) 1838. Type: Zygopetalum cochleare Lin Coelia Lindl. Gen. and Sp. Orch. Pl. 36. Apr. 1830. Type: Epidendrum tripterum Sm. Comparettia Poepp. & Endl. Nov. Gen. ac Sp. Pl. 1: 42. May 28, 1836. Lectotype: Comparettia falcata Poepp. & Endl. OBSERVATION: Lectotype was chosen indirectly by Reichenbach in Walp. Ann. Bot. Syst. 6: 689. 1863, when he transferred C. saccata Poepp. & Endl. to the genus Scelochilus. Therefore, the lectotypification by Britton & Wilson, Sci. Surv. Porto Rico 5(2): 211. 1924 is not tenable. Corallorrhiza Gagneb. in Acta Helv. 2: 61. Feb. 1755. Type: Ophrys Corallorhiza L. Corydandra Rchb. Herbarienb. Nomencl. 53. no. 2047. July 1841. Type: Galeandra Baueri Lindl ’ OBSERVATION: This is merely an illegitimate new name for Galeandra Lindl. ee apedaes Thou. in Nouv. Bull. Sci. Soc. Philom. Paris, ser. 2. 1: 318. Apr. 1809. Type: Corymborkis corymbosa Thou OBSERVATION: The type is provided in n Thou: Hist. Pl. Orch. ¢. 37. 1822. LEcToTYPE: Cranichis muscosa OBSERVATION: Lectotype was chosen by Acufia in Bol. Tec. Cuba 60: 48. 1939. Cranichis Sw. Nov. Gen. & Sp. Pl., Prodr. 8. July 1788. Sw Cryptarrhena R. Br. in Bot. pon 2:'F 5153. Nov: 1, 1816. Type: Cryptarrhena lunata R. B Cryptophoranthus Barb. Rodr. Gen. et Sp. Orch. Nov. 2: 79. 1882. Lectotype: Pleurothallis fenestrata Barb. Rodr. OBSERVATION: Lectotype was chosen by Acufia in Bol. Tec. Cuba 60: 115. 1939. Cybelion Spreng. Syst. Veg. 3: 679. Jan—June 1826. Type: Jonopsis pulchella H. B. K : OBSERVATION: This is an illegitimate new name for Ionopsis H. B. K. and is a synonym of that generic name. Cyclopogon Presl, Rel. Haenk. 1: 93. 1827. YPE: Cyclopogon ovalifolium Presl. OBSERVATION: This is a synonym of Spiranthes L. C. Rich. 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, III 519 Cycloptera Endl. ex Jackson in Ind. Kew. 1: 680. Sept. 6, 1893. PE: Cyclopogon ovalifolium Presl. OBSERVATION: This is a synonym of Spiranthes L. C. Rich. Cyrtopera Lindl. Gen. and Sp. Orch. Pl. i Jan. 1833. Lectotype: Cyrtopodium Woodfordii Sim OBSERVATION: Lectotype was chosen by i Richard ¢ in rhe rad Dict. Hist. Nat. 4: 561. 1844. This is a synonym of Eulophia R. B paclltelicn R. Br. in Ait. Hort. Kew. ed. 2. 5: 216. Dec. 1813. E: Cymbidium Andersonii Lamb. ex Andr. Dendrophylax Rchb. f. in Walp. Ann. Bot. Syst. 6: 903. Dec. 1864. Type: Dendrophylax hymenanthus Rchb. f. Dichaea Lindl. Gen. and Sp. Orch. Pl. 208. Apr. 1833. Type: Epidendrum echinocarpum Sw. OBsERVATION: The type was suggested by Lindley first in Bot. Reg. 18: sub t. 1530. Oct. 1, 1832. Lindley described Dichaea a second time in Hook. Jour. Bot. 1: 5. 1834 with D. panamensis as the type Dichaeopsis Pfitzer, Entwurf Nat. Anord. Orch. 93. 1887. LECTOTYPE: Epidendru um graminoides Sw. in hoc loco. OBSERVATION: This is a synonym of Dichaea Lindl. Dicrypta Lindl. Gen. and Sp. Orch. Pl. 44. May 1830. Type: Dicrypta Baueri Lindl. OBSERVATION: This is a synonym of Maxillaria Ruiz & Pav. Dienia Lindl. in Bot. Reg. 10: sub ¢. 825. Sept. 1, 1824. Type: Dienia congesta Lindl OBsERVATION: This is a synonym of Malaxis Soland. ex Sw. Dilomilis Raf. Fl. Tellur. 4: 43. tee) 1838. TYPE: Octomeria serratifolia Hoo Dinema Lindl. Gen. and Sp. Orch. Pl. 111. Aug. 1831. Type: Epidendrum polybulbon Sw. OBSERVATION: Type originally suggested by Lindley in his Collectanea Bo- tanica, Appendix No. 125. Jan. 1826. The generic name is a synonym of Epidendrum L. Domingoa Schltr. in Urban, Symb. Antill. 7: 496. Aug. 15, 1913. LectotyPe: Epidendrum kvmnenoiles Rchb. OBSERVATION: Lectotype was chosen by Acufia i in Bol. Tec. Cuba 60: 64. 1939. Dothilis Raf. Fl. Tellur. 2: 6 Jan.—Mar. 1837. Type: Neottia grandiflora H OBSERVATION: This is a een of Chloraea Lindl. Elleanthus Presl, Rel. Haenk. 1: 97. 1827. . LECTOTYPE: Ebleonthus lancifolius Presl. OBSERVATION: Lectotype was chosen by Britton & Wilson, Sci. Surv. Porto Rico 5(2): 203. 1924. 520 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Eltroplectris Raf. Fl. Tellur. 2: 51. Jan—Mar. 1837. Type: Neottia calcarata Sw. Encyclia Hook. in Bot. Mag. 55: t. 2831. June 1, 1828. Type: Encyclia viridiflora Hook. OBSERVATION: This is a synonym of Epidendrum L, Enothrea Raf. FI. Tellur. 4: 43. (July) 1838. Type: Epidendrum graminifolium L. OBSERVATION: This is a synonym of Octomeria R. Br. Epicladium Small, Fl. Miami 56. Apr. 26, 1913. Type: Epidendrum Boothianum Lindl. OBSERVATION: This is a synonym of Epidendrum L. Epidendrum L. Sp. Pl. ed. 1. 2: 952. May 1, 1753 (Aug. 16, 1753). Lectotype: Epidendrum nodosum L. OBsERVATION: Lectotype was chosen by Britton & Wilson, Sci. Surv. Porto Rico 5(2): 203. 1924. This generic name was rejected in 1930; it is a synonym of Brassavola R. Br. Epidendrum L. Sp. Pl. ed. 2. 2: 1347. July 1763. Lectotype: Epidendrum nocturnum Jacq. OBSERVATION: This generic name has been conserved with the above given type in 1930. Lectotype was chosen by Sprague in Prop. Brit. Bot. 53. 1929. Epistephium Kunth, Syn. Pl. Equinoct. 1: 340. Dec. 9, 1822. Type: Epistephium elatum Kunth. Epithecia Knowles & Westcott, Fl. Cab. 2: 167. Jan. 1839. Type: Prosthechea glauca Knowles & Westcott. OBSERVATION: The type of this genus has been confused with Epidendrum glaucum Sw. 1788, because of Lindley’s transfer of it to Epidendrum glau- cum (Knowles & Westcott) Lindl. in 1840. Hence this generic name was considered by Kuntze in 1903 and by Schlechter in 1914 to be a synonym of Dichaea Lindl. and Dichaeopsis Pfitz. respectively. It is, however, as Lindley’s transfer indicates, a synonym of Epidendrum L. Erythrodes Bl. Bijdr. 6: t. 1, f. 72. June 1825. Type: Erythrodes latifolia Bl. OBsERVATION: The type is given by Blume in Bijdr. 8: 411. July 1825. Eulophia R. Br. ex Lindl. in Bot. Reg. 8: ¢. 686. Feb. 1, 1823, not Agardh, 1822. Type: Eulophia guineensis Lindl. OBSERVATION: This generic name was conserved in 1905 and typified by Green in 1929 by Eulophia barbata Spreng. which was adopted by the 5th Inter- national Botanical Congress in 1930, However, Eulophia barbata is one 0 the main species upon which Pfitzer based his new genus Acrolophia in 1887. In segregating Acrolophia, Pfitzer explicitly typified Eulophia by E. guineensis Lindl. in his Entwurf einer natuerlichen Anordnung der Orchi- deen p. 59, 1887. Summerhayes and Hall in Taxon 11: 203. 1962 proposed changing the conserved type from E. barbata to E. guineensis. No action 1972 | GARAY & SWEET, WEST INDIAN ORCHIDS, II 521 has yet been taken by either the 10th or 11th International Botanical Congresses. Eurystyles Wawra in Oesterr. Bot. Zeitschr. 13: 223. July 1863. Tyre: Eurystyles cotyledon Wawra. Evelyna Poepp. & Endl. Nov. Gen. ac Sp. Pl. 1: 32. May 28, 1836. Lectotype: Evelyna capitata Poepp. & Endl. OBsERVATION: Lectotype was chosen by Pfeiffer, Nomencl. Bot. 1(2):. 1321. 1874. It is a synonym of Elleanthus Presl. Fractiunguis Schltr. in Anex. Mem. Inst. Butantan, Bot. 1(4): 55. Mar. Lectotype: Hexisea reflexa Rchb. f. OssERVATION: Lectotype was chosen by Acufia in Bol. Tec. Cuba 60: 89. 1939. This is a synonym of Reichenbachanthus Barb. Rodr. Fuertesiella Schltr. in Urban, Symb. Antill. 7: 492. Aug. 15, 1913. Tyre: Fuertesiella pterichoides Schltr. Galeandra Lindl. in Bauer & Lindley, Ill. Orch. Pl., Gen. ¢. 8. 1832. Type: Galeandra Baueri Lindl. Ghiesbreghtia Rich. & Gal. in Ann. Sci. Nat. ser. 3. 3: 28. Jan. 1845. Type: Ghiesbreghtia calanthoides Rich. & Gal. OssERVATION: This is a synonym of Calanthe R. Br. Goodyera R. Br. in Ait. Hort. Kew. ed. 2. 5: 197. Dec. 1813. LectTotyPE: Satyrium repens L. OBsERVATION: Lectotype was chosen by Sprague in Jour. Bot. 64: 113. 1926. This generic name was conserved with the above given lectotype in 1930. Govenia Lindl. in Lodd. Bot. Cab. 18: ¢. 1709. July 1831. Tyre: Maxillaria superba La Llave & Lex. Gyrostachys Pers. ex Bl. Coll. Orch. 127. 1859. LectoTyPe: Ophrys spiralis L. OssERvATION: Lectotype was chosen by O. Kuntze, Rev. Gen. Pi. 2: 663. 1891. This generic name was rejected in 1905; it is a synonym of Spiranthes L..C, Rich. Habenaria Willd. Sp. Pl. 4: 44. 1805. Lectotype: Orchis Habenaria L. OBSERVATION: The correct lectotype was chosen first by Kraenzlin in Engl., Bot. Jahrb. 16: 58. 1892. The choice of Orchis monorrhiza Sw. for the lectotype by Lindley in Bot. Reg. 18: sub ¢. 1499. 1832, does not carry the appropriate name. Hapalorchis Schitr. in Beih. Bot. Centralbl. 37(2): 362. Mar. 31, 1920. Lectotype: Sauroglossum candidum Kral. OBSERVATION: Lectot was chosen by Britton & Wilson, Sci. Surv. Porto Rico 5(2): 186. 1924. Schlechter at the time of publication recognized four species, but all four have been united under Spiranthes lineata Lindl. by Garay & Dunsterville, Venez. Orch. Ill. 4: 278. 1966. This is a synonym of Spiranthes L. C. Rich. JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 522 Harrisella Fawcett & Rendle in Jour. Bot. 47: 265. July 1909. Type: Aeranthus porrectus Rchb. f OxsERVATION: This is a synonym af Campylocentrum Benth. Heterotaxis Lindl. in Bot. Reg. 12: t. 1028. Dec. 1, 1826. Type: Heterotaxis crassifolia Lindl. OBsERVATION: This is a synonym of Maxillaria Ruiz & Pav. Hexadesmia Brongn. in Ann. Sci. Nat. ser. 2. 17: 44. Jan. 1842 Type: Hexadesmia fasciculata Brongn. nears Sens Rolfe in Hook. Icon. Pl. 25: t. 2461. May 1896. E: Homalopetalum jamaicense Rolfe Hormidium Lindl. ex Haynhold, Nom. Bot. 2: 880. 1840 Type: Epidendrum uniflorum Lindl. papers ce The type was chosen by Lindley in Bot. Reg. 25: Misc. p. 16. 1839. This is a synonym of Epidendrum L. Humboldtia Ruiz & Pav. Fl. Peru. Prodr. 121. Oct. 1794 (“Humboltia’), not Vahl 1794 LECTOTYPE: Humboldtia purpurea Ruiz & Pav., in hoc loco. OBSERVATION: The orthography has been comrectad by Ruiz & Pav. Syst. Veg. 233. Dec. 1798. For the explanation of lectotype see Stelis Sw. of which it is a synonym. Ibidium Salisb. ex Small, Fl. SE. U-S. ed. 2. 318. Apr. 23, 1913. Rich. LECTOTYPE: Ophrys spirals L. OBSERVATION : Lectotype was chosen by House in Bull. Torr. Bot. Club 32: 380. 1905. This is a synonym of Spiranthes L. C. Ionopsis H. B. K. Nov. Gen. et i> Pl. 1: 348. Aug. 1816. Type: lonopsis pulchella H. B. K Isochilus R. Br. in Ait. Hort. Kew. ed. 2. 5: 209, Dec. 1813. Lectotyre: Epidendrum lineare Jacq. OBSERVATION: Lectotype was chosen indirectly by Reichenbach in Bonplandia 422, 1854. easels Schltr. in Fedde Repert. Sp. Nov. Beih. 7: 123. Jan. 31, 920. by eciwe Epidendrum globosum Jacq. OBSERVATION: SN was chosen by Britton & Wilson, Sci. Surv. Porto Rico 5(2): 197. Koellensteinia Rchb. f. in ee rina 2: 17. Jan. 15, 1854. Type: Koellensteinia Kellneriana Rchb. f Laelia Lindl. Gen. and Sp. Orch. Pl. 115. Aug. 1831, not Adanson 1763. & Lex. Lectotype: Bletia grandiflora La Llave OBSERVATION: Lectotype was chosen by Dandy in Kew Bull. 86. 1935. This generic name was conserved in 1935. Laeliopsis Lindl. in Lindl. & Paxton, Fl, Gard. 3: 155. ¢. 105. Jan. 1853. YPE: Cattleya domingensis Lindl 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, III 523 Lankesterella Ames, Sched. Orch. 4: 3. May 4, 1923. Type: Lankesterella costaricensis Ames. Leochilus Knowles & Westcott, Fl. Cab. 2: 143. Nov. 1838. Type: Leochilus oncidioides Knowles & Westcott Lepanthes Sw. in Nov. Act. Soc. Sci. Upsala 6: 85. 1799. Lectotype: Epidendrum ovale Sw. (Lepanthes concinna Sw. nom. illeg.) OBSERVATION: Lectotype was chosen by Britton & Wilson, Sci. Surv. Porto Rico 5(2): 206. 1924. a i Ames in Bot. Mus. Leafl. Harv. Univ. 1(9): 3. Aug. 12, rane Pleurothallis floripecten Rchb. f. Leucohyle Klotzsch in Ind. Sem. Hort. Berol. App. 1. 1854. Type: Leucohyle Warszewiczii Kl. Limodorum L. Sp. Pl. ed. 1.2: 950. pda 1753 (Aug. 16, 1753). LectotyPe: Limodorum tuberosum OBSERVATION: Lectotype was en by Britton & Brown, Ill. Fl. No. U.S. and Canada, ed. 2. 1: 562. 1913. This generic name was rejected in 1935 and is a synonym of Calopogon R. Br. Liparis L. C. Rich. Orch. Eur. Annot. 21. Sept. 1817. Type: Ophrys Loeselii L. OBSERVATION: The genus was conserved in 1905, and the type was conserved Lycaste Lindl. in Bot. Reg. 28: Misc. p. 85, no. 96. Dec. 1842. Tyre: Lycaste plana Lindl. OBsERVATION: The description of Lycaste plana must be regarded as a com- bined “descriptio generico-specifica” according to Article 42 of the /nter- national Code of Botanical i ae ee Lindley’s newly published diagnosis in Bot. Reg. 29: Misc. p. 14. Feb. 1843. The choice of Lycaste macrophylla Lindl. by Acufia in Bol. oe Cuba 60: 165. 1939 is in- correct. Lysimnia Raf. Fl. Tellur. 4: 143. (July) 1838. Type: Brassavola cordata Lindl. OBSERVATION: This is a synonym of Brassavola R. Br. Macradenia R. Br. in Bot. Reg. 8: ¢. 612. Mar. 1, 1822. Type: Macradenia lutescens R. Br. Malaxis Soland. ex Sw. Nov. Gen. & Sp. Pl., Prodr. 8. July 1788. Lectotyre: Malaxis spicata Sw OBSERVATION: Lectotype was chosen by Britton & Brown, Ill. Fl. No. U.S. and Canada, ed. 2. 1: 570. 1913. The choice of Malaxis Rheedii Sw. by Ascher- son, Fl. Prov. Brandenb. 1: 699. Apr. 1864 is incorrect for it was not part of the original protologue. 524 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 gg eadge oe Ruiz & Pav. Fl. Peru. Prodr. 122. Oct. 1794. YPE: Masdevallia uniflora Ruiz & Pav. facsenen. The type was first suggested in Ruiz & Pavén, Syst. Veg. 238. 1798. Maxillaria Ruiz & Pav. Fl. Peru. Prodr. 116. Oct. 1794. LectotyPE: Mazillaria ramosa Ruiz & Pav., in hoc loco. OpsERVATION: A new lectotype is selected *hets, because the former typifica- tions are incorrect. The choice of Moazillaria longipetala Ruiz & Pav. by Acufia in Bol. Tec. Cuba 60: 171. 1939, is not tenable because of the transfer of this epithet to Lycaste longipetala (Ruiz & Pav.)Garay in Caldasia 8: 524. 1962. The choice of Mazxillaria platypetala Ruiz & Pav. by Brieger and Hunt in Taxon 18: 602. 1969 is equally untenable for the decision is based on a specimen not part of the original protologue. Brieger and Hunt state that “. . . it is known that Lindley received material from Ruiz and Pavén (Rchb. f. in Bonplandia 4: 210. 1856) and therefore this specimen could be considered to be at least an isotype of M. platypetala.” This argument would be justifiable were it not for Lindley who stated in Hooker, Jour. Bot. 1: 13. 1834, and in his Genera and Species of Orchida- ceous Plants 143. Dec. 1832, that his only specimen was collected by Colonel Hall (hab. s. sp. comm. cel. Hooker). On the other hand, of the 16 species originally assigned to Mazillaria, most of which are referable to other genera, there still exists an actual specimen of Mavxillaria ramosa in the Delessert Herbarium in Geneva, which is part of the original protologue; this was pointed out by Garay in Bot. Mus. Leafl. Harvard Univ. 21: 259. Apr. 28, 1967. Since this is the only Pavén material which corresponds to Mazillaria as it has been used in syste- matics and horticulture for the last 150 years, it must be chosen as the lectotype for the genus Mesadenella Pabst & Garay in Arquiv. Jard. Bot. Rio de Janeiro 12: 205. 1952. Lectotype: Spiranthes esmeraldae Linden & Rchb. f. OBsERVATION: Lectotype was chosen by Correa in Darwiniana 11: 68. 1955. This is a synonym of Spiranthes L. C. Rich. Mesadenus Schltr. in Beih. Bot. eee 37(2): 367. Mar. 31, 1920. LectotypPe: Spiranthes Galeottiana OBSERVATION: Lectotype was chosen be ae & Wilson, Sci. Surv. Porto Rico 5(2): 186. 1924. This is a synonym of Spiranthes L. C. Rich. ernie (Nutt.) Eaton, Man. Bot. ed. 3. 115, 353. Apr. 1822. : Malaxis opkioslossoides Willd Eset free This generic name was s conserved i in 1954. It is a synonym of alaxis Soland. ex Sw Myrmecophila Rolfe in Orch. Rev. 25: 50. Mar. 1917. Type: Schomburgkia tibicinis Batem. OBsERVATION: This is a synonym of Schomburgkia Lindl. Myrobroma Salisb. Parad. Londin. 2: t. 82. Sept. 1, 1807. Typ. E: Epidendrum rubrum Lamarck. OBSERVATION : This is a synonym of Vanilla Plum. ex Mill. 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, III 525 Narica Raf. FI. Tellur. 2: 87. ener 1837. Type: Neottia acaulis J. E. S OBSERVATION: This is a aa of Spiranthes L. C. Rich. sid Schltr. in Urban, Symb. Antill. 7: 495. Aug. 15, 1913. E: Trigonidium monophyllum Griseb. Neourbania Fawcett & Rendle in Jour. Bot. 47: 125. Apr. 1909. Type: Ponera adendrobium Rchb. f. N ee Raf. Fl. Tellur. 2: 89. Jan—Mar. 1837. YPE: Neottia glandulosa Sims. pdeshedon This is a superfluous name and a synonym of Ponthieva R. Br. Nidema Britton & Millspaugh, Bahama Flora 94. June 26, 1920. Type: Epidendrum Ottonis Rchb. f. OBSERVATION: This is a synonym of Epidendrum L. Ocampoa Rich. & Gal. in Ann. Sci. ins ser. 5. 32 31. Jan. 1845. Type: Ocampoa mexicana Rich. & G OBSERVATION: This is a synonym of eileen Sw. Octadesmia Benth. in Benth. & Hook. Gen. Pl. 3: 525. Apr. 1883. Type: Octomeria serratifolia Hook. OBSERVATION: This is a synonym of Dilomilis Raf. Octomeria R. Br. in Ait. Hort. Kew. ed. 2. 5: 211. Dec. 1813. Type: Epidendrum graminifolium L. Odontoglossum H. B. K. Nov. Gen. et Sp. Pl. 1: 350. Aug. 1816. Type: Odontoglossum epidendroides H. B. K. Olgasis Raf. Fl. Tellur. 2: 51. Jan—Mar. 1837. Type: Cymbidium triquetrum Sw OBSERVATION: This is a synonym of Oncidium Sw. Oncidium Sw. in Vet. Akad. Handl. Stockholm 21: 239. Sept. 1800. Lectotype: Oncidium variegatum Sw OBSERVATION: Lectotype was chosen by Garay in Taxon 19: 444. 1970. This choice corresponds to the guidelines provided in the International Code of Botanical Nomenclature for the selection of lectotypes, for this is the only species which Swartz illustrated in detail when he described the genus, and of which the holotype also exists in the British Museum (Nat. Hist e choice of Epidendrum altissimum Jacq. by Pfeiffer in his Nomencl. ‘Bot. 2: 497. 1874 is inappropriate for Oncidium (Epidendrum) altissimum (Jacq.) Sw. has been thoroughly misapplied up to the present time. (See a . Sweet in Jour. Arnold Arb. 53: 394, 395. 1972), The choice of Epidend carthagenense Jacq. by Britton & Wilson, Bahama Flora 97. 1920, is ened here because no material of the holotype exists. Opteron Raf. Fl. Tellur. 4: sh Bibi 1838. E: Maxillaria pallidiflora H nial RIBAS This is a le ie Xylobium Lindl. 526 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 wal hia Salisb. ex R. Br. in Ait. Hort. Kew. ed. 2. 5: 210. Dec. 1813. E: Epidendrum coccineum L. aa. The type was suggested by Salisbury in Trans. Hort. Soc. 1: 293. 1812. This is a synonym of Maxillaria Ruiz & Pav Ornithocephalus Hook. Exot. Fl. 2: ¢. 127. Oct. 1824. Type: Ornithocephalus gladiatus Hook. Pelexia Poit. ex Lindl. in Bot. Reg. 12: sub ¢. 985. June 1, 1826. Type: Satyrium adnatum Sw. OpsERVATION: This generic name was conserved in 1935, It is a synonym of Spiranthes L. C. Rich. Pentulops Raf. Fl. Tellur. 4: De (July) 1838. Tyre: Mazillaria decolor Lin OBSERVATION: This is a ae of Xylobium Lindl. Physurus L. C. Rich. ex Lindl. Gen. and Sp. Orch. Pl. 501. Sept. 1840. Lectotype: Orchis plantaginea Willd. OxpsERVATION: Lectotype was chosen by Britton & Millspaugh, Bahama Flora 87. 1920. It is a synonym of Erythrodes Bl. Pinelia Lindl. Folia Orch. Pinelia 1. Apr. 30. 1853. Type: Pinelia hypolepta Lindl. Platypus Small & Nash in Small, Fl. SE. U.S. 329. July 1903. Type: Platypus papilliferus Small & Nash. OBSERVATION: This is a synonym of Eulophia R. Br. Platystele Schltr. in Fedde Repert. Sp. Nov. 8: 565. Nov. 15, 1910. Type: Platystele bulbinella Schitr. Pleuranthium (Rchb. f.) Benth. in Jour. Linn. Soc. Bot. 18: 312. Feb. 21, 1881. Type: Epidendrum Dendrobii Rchb. f. OpsERVATION: This is a synonym of Epidendrum L. oe R. Br. in Ait. Hort. Kew. ed. 2. 5: 211. Dec. 1813. E: Epidendrum ruscifolium Jacq. Polyradicion Garay in Jour. Arnold Arb. 50: 466. July 15, 1969. Type: Angraecum Lindenii Lindl. Polyrrhiza Pfitz. in Nat. Pflanzenfam, 2(6): 215. Mar. 1889. Type: Limodorum funale Sw OBSERVATION: This is a synonym of Dendrophylax Rchb. f. Polystachya Hook. Exot. Fl. 2: ¢. 103. May 1824. Type: Epidendrum minutum Aubl. OgsERVATION: This generic name was conserved in 1905. The correct name of the type species in the genus Polystachya is Polystachya flavescens (B1.) J. J. Sm. (See Dunsterville & Garay, Venez. Orch. Ill. 3: 266. 1965). Ponera Lindl. Gen. and Sp. Orch, Pl. 113. Aug. 1831. Tyee: Ponera juncifolia Lindl. 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, III 527 Ponthieva R. Br. in Ait. Hort. Kew. ed. 2. 5: 199, Dec. 1813. Type: Neottia glandulosa Sims. Prescottia Lindl. in Hook. Exot. Bot. 2: ¢. 115. Aug. 1824. Type: Prescottia plantaginea Lindl. Prosthechea Knowles & Westcott, Fl. Cab. 2: 111. Sept. 1838. Type: Prosthechea glauca Knowles & Westcott. OBSERVATION: This is a synonym of Epidendrum L. er Rane. Lindl. in Jour. Linn. si 3: 64. Aug. 20, 1858. YPE: Pseudocentrum macrostachyum Lind ler eee Schltr. in Beih. is Centralbl. 37(2) : 369. Mar. 31, 1920. E: Goodyera Wrightii Rchb poeleireatiae This is a synonym i Spiranthes L. C. Rich. Psilochilus Barb. Rodr. Gen. et Sp. Orch. Nov. 2: 272. 1882. Type: Psilochilus modestus Barb. Rodr. Pterichis Lindl. Gen. and Sp. Orch. Pl. 444. Sept. 1840. Type: Pterichis galeata Lindl. Pteroglossaspis Rchb. f. Otia Bot. Hamb. pt. 1: 67. Apr. 1878. Type: Pteroglossaspis eustachya Rchb. f. Reichenbachanthus Barb. Rodr. Gen. et Sp. Orch. Nov. 2: 164. 1882. Type: Reichenbachanthus modestus Barb. Rodr. Rhynchadenia A. Rich. in Sagra, Hist. Isla Cuba 11: 248. 1841. Type: Rhynchadenia cubensis A. Ric OBSERVATION : This is a synonym of Macradenia R. Br. Rodriguezia Ruiz & Pav. Fl. Peru. Prodr. 115. Oct. thes Lectotype: Rodriguezia lanceolata Ruiz & Pav., in hoc loc OBSERVATION: Lectotype is chosen here. The holotype of R. lanceolata is in the Reichenbach Herbarium in Vienna, and it closely matches the illus- trations provided by Ruiz and Pavéon, part of the original ae Reich- enbach already noted this similarity in Bot. Zeit. 10: 772. oe Schltr. in Fedde Repert. Sp. Nov. 16: 427. July 31, ue Rodriguezia eleutherosepala Barb. Rodr., in hoc loco OBSERVATION: Lectotype is chosen here. Of the two equally characteristic species referred to this genus by Schlechter, R. eleutherosepala is the more common one and specimens referable to it are widely distributed in her- baria. The holotype is Barbosa Rodrigues’ original drawing. Sacoila Raf. Fl. Tellur. 2: 86. Jan—Mar. 1837. Type: Neottia aphylla Hook. OBSERVATION: This is a synonym of Spiranthes L. C. Rich. Sarcoglottis Presl, Rel. Haenk. 1: 95. 1827. YPE: Sarcoglottis speciosa Presl. OBSERVATION: This is a synonym of Spiranthes L. C. Rich. 528 JOURNAL OF THE ARNOLD ARBORETUM [voL, 53 . Type: Sauroglossum elatum Lin OBSERVATION: This is a synonym of Spiranthes L. C. Rich. Scaphyglottis Poepp. & Endl. Nov. Gen. ac Sp. Pl. 1: 58. May 28, 1836. Lectotype: Fernandezia graminifolia Ruiz & Pav., in hoc loco. OBSERVATION: Lectotype is chosen here. The holotype is in the Herbarium of Jardin Botanico, Madrid. The choice of Scaphyglottis parviflora Poepp. & Endl. by Pfeiffer, Nomencl. Bot. 2: 1068. 1874 is untenable because this epithet was transferred to Mazxillaria by Garay in Bot. Mus. Leafl. Harvard Univ. 21: 258. 1967. Schomburgkia Lindl. Sert. Orch. ¢. 10. Apr. 1, 1838. Type: Schomburgkia crispa Lindl. Seraphyta Fischer & Meyer in Bull. Sci. Acad. St. Pétersb. 7: 24. Mar. 10, 1840 Sauroglossum Lindl. in Bot. Reg. 19: ¢. 1618. Oct. 1, 1833. Type: Epidendrum diffusum Sw. OBSERVATION: This is a synonym of Epidendrum L. Spathiger Small, Fl. Miami 55. Apr. 26, 1913. Type: Epidendrum rigidum Jacq. OBSERVATION: This is a synonym of Epidendrum L. Specklinia Lindl. Gen. and Sp. Orch. Pl. 8. Apr. 1830. Lectotype: Epidendrum lanceola Sw., in hoc loco. OBsERVATION: Lectotype is chosen here because this species is the one which most approximates the generic characters given by Lindley: ‘“labellum petalis conforme,” “sepalis liberis, basi saccatis.” The holotype is in the British Museum (Nat. Hist.). This is a synonym of Pleurothallis R. Br. Spiranthes L. C. Rich. Orch. Eur. Annot. 20. Sept. 1817. Lectotype: Ophrys spiralis L. OBsERVATION: Lectotype was chosen by Pfeiffer, Nomencl. Bot. 2(2): 1238. 1874, and was conserved in 1930. The choice of Spiranthes aestivalis L. C. Rich. by Correll in Fl. Texas 3(3): 169. 1944, is inappropriate. This generic name was conserved in 1905. 1972] GARAY & SWEET, WEST INDIAN ORCHIDS, III 529 shown in Taxon 8: 258. 1959; the conservation was superfluous at that time. In 1929, Epidendrum ophioglossoides Jacq. was selected as the lectotype by Green in Prop. Brit. Bot. p. 100, and this typification was conserved in 1930. Unfortunately, Epidendrum ophioglossoides is based on Plumier’s polynomial of 1703 as well as on Plumier’s drawing of it published subse- quently by Burmann in 1759. A closer examination of Plumier’s drawing reveals at once that it is referable to the genus Pleurothallis R. Br. as we understand it, and is conspecific with Pleurothallis floribunda (Lindl.) Lindl. The above proposed retypification of Stelis is necessary to prevent a mass transfer of some 1,000 specific epithets from Pleurothallis which is inevit- able if the old typification is maintained Stellilabium Schltr. Die Orchideen 530. Nov. 28, 1914. Type: Telipogon astroglossus Rchb. f Stenoptera Presl, Rel. Haenk. 1: 95. 1827. Type: Stenoptera peruviana Presl. Stenorrhynchos L. C. Rich. ex aig Syst. Veg. 3: 677. Jan.—June 1826. Lectotype: Neottia speciosa Jac OBSERVATION: Lectotype was ent by Britton & Millspaugh, Bahama Flora 86. 1920. The choice of Satyrium orchioides Sw. as lectotype by Correa in Darwiniana 11: 70. 1955 is incorrect. Regarding orthography see Sprague in Kew Bull. 243. 1929. This is a synonym of Spiranthes L. C. Rich. Sturmia Rchb. Icon. Bot. Pl. Critic. 4: 39. 1826. Type: Ophrys Loeselii L. OBSERVATION: This is a synonym of Liparis L. C. Rich. pie hana! Rchb. f. in Bonplandia 2: 21. Jan. 15, 1854. CTOTYPE: Tetragamestus aureus Rchb. f. jabs Lectotype was chosen indirectly by Reichenbach in Linnaea 41: 85. 1876, when he transferred 7. modestus to Ponera. The choice of Scaphyglottis arundinacea H. Petrop. by Pfeiffer, Nomencl. Bot. 2(2): 1373. 1874 is not correct, for this latter species is not part of the original protologue. This is a synonym of Scaphyglottis Poepp. & Endl. Tetramicra Lindl. Gen. and Sp. Orch. Pl. 119. Aug. 1831. Type: Cymbidium rigidum Willd Thiebautia Colla in Mem. Soc. Linn. Paris 3: 161. 1824. Type: Limodorum purpureum Lamarck. OBSERVATION: This is a synonym of Bletia Ruiz & Pav. Todaroa Rich. & Gal. in Ann. Sci. Nat. ser. 3. 3: 28. Jan. 1845, not Parla- tore 1843 Type: Todaroa micrantha Rich. & Gal. OBSERVATION: This is a synonym of Campylocentrum Benth. i sesiobte a Schltr. in Beih, Bot. Centralbl. “ @ 2): 423. Mar. 31, 1920. OTYPE: Stenorrhynchus actinosophila Barb. R 7B Nis Lectotype was chosen by Acufia in tok Tec. Cuba 60: 43. 1939. This is a synonym of Eurystyles Wawra. 530 JOURNAL OF THE ARNOLD ARBORETUM [voL, 53 Trichopilia Lindl. Nat. Syst. Bot. ed 2. 446. May 1836. Type: Trichopilia tortilis Lindl. Trigonidium Lindl. in Bot. ate 23: #. 1923. Jan. 1, 1837. Type: Trigonidium obtusum Lind Triorchos Small & Nash in Small, Fl. SE. U.S. 329. July 1903. Type: Cyrtopodium ecristatum Fern. OBSERVATION: This is a synonym of Pteroglossaspis Rchb. f. Triphora Nutt. Gen. N. Am. Pl. 2: 192. June 1818. Type: Arethusa trianthophora Sw. Tritelandra Raf. Fl. Tellur. 2: 85. an —Mar. 1837. Type: Epidendrum secundum Jac OxsERVATION: This is a synonym a Epidendrum L. Tropidia Lindl. in Bot. Reg. 19: sub ¢. 1618. Oct. 1, 1833. Type: Tropidia curculigoides Lin OBSERVATION: The type was suggested by Lindley in Wallich’s Catalogue No. 7386. 1832 Tylochilus Nees in Verh. Gartenb. Gesellsch. Berlin 8: 194. #. 3. 1832. Tyre: Tylochilus flavus Nees. OBSERVATION: This is a synonym of Cyrtopodium R. Br. Ulantha Hook. in Bot. Mag. 57: sub t. 2990. June 1, 1830. Type: Neottia grandifiora Hook OBSERVATION: This is a synonym of Chloraea Lindl. Vanilla Plum. ex Mill. Gard. Dict. abridged ed. 4. 3: without page num- ber. Jan. 28, 1754. LECTOTYPE: ibalendram Vanilla L. OBSERVATION: Lectotype was chosen by Britton & Millspaugh, Bahama Flora 83. 1920. The correct name of the type is Vanilla mexicana Mill. (See Mansfeld in Kulturpfl. Beih. 2: 587. 1959.) : yee cteonant Rchb. f. in Bot. Zeit. 10: 635. Sept. 10, 1852. Lectotype: Warrea discolor Lindl. ree hlcasi ag Lectotype was chosen by Britton & Wilson, Sci. Surv. Porto Rico 5(2): 214. 1924. This is a synonym of Cochleanthes Raf. Wullschlaegelia Rchb. f. in Bot. Zeit. 21: 131. Apr. 17, 1863. YPE: Cranichis aphylla Sw. Xeilyathum Raf. FI). Tellur. 2: va Jan.—Mar. 1837. Type: Epidendrum altissimum Jac OBSERVATION: This is a synonym oF Oncidium Sw. Xylobium Lindl. in Bot. Reg. 11: sub ¢. 897. July 1, 1825. Type: Dendrobium squalens Lindl. OrcHID HERBARIUM OF OAKES AMES BoTaANICAL MusEUM Harvarp Un Y CAMBRIDGE, MASSACHUSETTS 02138 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 531 THE GENERA OF ZYGOPHYLLACEAE IN THE SOUTHEASTERN UNITED STATES ! DunNcAN M. PoRTER ZYGOPHYLLACEAE R. Brown in Flinders, Voy. Terra Austral, 2: 545. 1814, ““Zygophylleae” (BEAN-CAPER FAMILY) Annual to perennial [often suffrutescent] herbs [shrubs] or occasionally trees; branches usually divaricate, with angled or swollen nodes, growth sympodial. Leaves opposite [occasionally alternate], even-pinnately compound, [sometimes odd pinnate, occasionally simple or 2-foliolate, rarely 3—7-foliolate], often fleshy to coriaceous, persistent, petiolate [to subsessile]; leaflets entire [sometimes lobed], inequilateral, petiolulate to subessile; stipules paired, free, foliaceous [sometimes fleshy or spinescent], persistent or rarely deciduous to caducous. Flowers [4]5[6]-merous, per- fect, hypogynous, regular or occasionally slightly irregular; peduncles terminal or pseudaxillary, 1-flowered, solitary or occasionally few to many. Sepals [4] 5 [6], free or rarely slightly connate basally, imbricate in bud, persistent or occasionally deciduous. Petals as many as sepals, free [rarely connate basally], often clawed, sometimes twisted, imbricate or inal glandular disc usually present and conspicuous. Stamens in [1] 2 [or 3] whorls of 5 each, outer[most] whorl usually opposite petals, often alternately unequal in length or sterile; filaments free, subulate to filiform [rarely winged], frequently glandular [or appendaged] basally, outer whorl occasionally adnate basally to petals, inserted on or below disc; an- thers 2-loculate, subbasifixed to versatile, introrse, longitudinally dehiscent. Gynoecium [2—]5-carpellate, syncarpous; style terminal, usually simple; 1 This treatment was prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard Uni- versity made possible through the support of the National Science Foundation (Grant GB-6459X, principal investigator, C. E. Wood, Jr.). The format is that established in the first paper in the series (Jour. Arnold Arb. 39: 296-346. 1958). The area covered includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississippi, Arkansas, and Louisiana. The descriptions are based primarily on the plants of this area, with additional material from extraterritorial taxa in brackets. References that I have not seen are marked by an asterisk. I am very grateful to Dr. Wood for his numerous suggestions and additions, to Mrs. Nancy Dunkly for the preparation of the typescript, and to both for their help in re- i i muc 532 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 stigma minutely and obscurely lobed to distinctly ridged; ovary superior, [2—]5—10{—12]-lobed and -loculate, sessile or rarely on a short gynophore; ovules 1 or 2 to many in each locule, pendulous [or ascending], anatropous, placentation axile [rarely basal]. Fruit a [2—]5-lobed septicidal [or loculi- cidal] capsule or a schizocarp splitting lengthwise into 5—10[-12] hard, tuberculate to spiny [or winged] mericarps [rarely a drupe or berry]. Seeds 1 to several [to many] in each locule, endosperm present or absent; embryo with flat cotyledons. TypPE Genus: Zygophyllum L. Twenty-six genera and about 250 species, widely distributed, but mainly in the warmer, drier regions of the world. Eight genera are represented in North America: Guaiacum L., with one native and several cultivated species in southern Florida; Kallstroemia Scop., with six native and one in- troduced species, mainly in the Southwest; Peganum L., with P. mexicanum Gray native to Texas and P. hkarmala L. introduced into the Southwest from the Old World deserts; Larrea Cav., with L. tridentata (DC.) Coville widely distributed in the deserts of the southwestern United States; Por- lieria Ruiz & Pavon, with P. angustifolia (Engelm.) Gray in Texas; Tribu- lus L., with two introduced species; and Zygophyllum L., with the intro- duced Z. Fabago.L. Species of Guaiacum, Kallstroemia, and Tribulus occur in our area The Zygophyllaceae are a somewhat isolated family with no close rela- tives, except perhaps for the Balanitaceae. The family has recently been placed in the Geraniales (Takhtajan, 1959, 1969; Scholz, 1964; Thorne, 1968), Malpighiales (Hutchinson, 1959, 1969, in his Lignosae), and Sapindales (Cronquist, 1968). Airy Shaw (1966) thinks that they are closely related to the Rutaceae; Hutchinson (1969) supposes that they are derived from the Linaceae; and many authors place them near the Erythroxylaceae, Geraniaceae, Oxalidaceae, and other Geraniales sensu Engler, but these taxa do not appear to be closely related. The best in- dication of the families to which they are allied is that of Cronquist (1968), who places them with his Sapindales-Geraniales-Linales-Polygalales com- lex. The latest subfamilial classification of the family (Scholz, 1964), follows that originally published by Engler (1890, 1896); It is still a far from satisfactory treatment. The Zygophyllaceae at present are di- vided into seven subfamilies: Augeoideae Engler (Augea Thunb.), Chi- tonioideae Engler (Morkillia Rose & Painter, Viscainoa Greene), Nitrari- oideae Engler (Nitraria L.), Peganoideae Engler (Malacocarpus Fischer & Meyer, Peganum), Tetradiclidoideae Engler (Tetradiclis Stev. ex Bieb.), Tribuloideae D. M. Porter (Kallstroemia, Kelleronia Schinz, Tribulopsis R. Br., Tribulus),? and Zygophylloideae (Bulnesia C. Gay, Fagonia L., Guaiacum, Larrea, Metharme Phil. ex Engler, Miltianthus Bunge, Neolue- * These four genera, which form a natural group within the peg have been dis- cussed in detail elsewhere (Porter, 1969a, pp. 42-46). They deserve recognition as sub- family Tribuloideae D. M. Porter, stat. nov. (Based on ren “Tribuleae H. G. L. Reichenbach, Conspectus Regni Veg. 199. 1828.) 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 533 deritzia Schinz, Pintoa C. Gay, Plectrocarpa Gill., Porlieria, Seetzenia R. Br., Sericodes Gray, Sisyndite E. Meyer ex Sonder in Harvey & Sonder, Tetraena Maxim., and Zygophyllum). All but the last appear to be natural taxa, and further study undoubtedly will result in a rearrangement of the Zyeophylloideae. hoe (1967) has provided an excellent key to the genera of the fam Palynological nai (Erdtman, 1952; Agababian, 1964; and Kaligis- Walalangi, 1969) indicate a heterogeneity of pollen types within the family, especially within the Zygophylloideae as recognized above. A combination of data from palynology, comparative anatomy and morphol- ogy, and cytology should provide a more rational treatment of this sub- family. Pollen morphology and wood anatomy support the removal of Balanites Del. and the recognition of the family Balanitaceae, intermediate between Zygophyllaceae and Simaroubaceae. Reported chromosome numbers are Bulnesia Retamo (Gill.) Griseb. ex Hooker, 2” = 26 (x = 13); Fagonia (6 spp.), 2m = 18, 20, 22, ca. 24 (x = 9); Guaiacum officinale L., 2n = ca. 26; Kallstroemia pubescens (G. Don) Dandy in Keay, 2n = 32; Larrea (5 spp.), 2m = 26, ca. 26, 52, ca. 52 (x = 13); Malacocarpus crithmifolius (Retz.) C. A. Meyer, 2n = 24; Nitraria (3 spp.), 2n = 18, 24, 48, ca. 48, 60 (x = 12); Peganum Har- in, 2n = 22, 24; Tribulus (2 aah, 2n = 12, 24, 36, 48 (x = 6, 12); Viscainoa geniculata var. pinnata I. M. Johnston, an = 26; Zygophyllum (18 spp.), 22 = 16, 18, 20, ca. 20, 22, 44 (x = 8). Many studies anve beet made of the wood anatomy of various taxa in the family. The wood is highly specialized, and the species studied form a distinctive, natural group. The specialized features (cf. Metcalfe & Chalk, 1951, p. 291) are “a vessel member length of 0.1-0.2 mm., storied struc- ture, homogeneous rays and fusiform parenchyma cells, with several other features, such as solitary vessels, diffuse parenchyma and fibres with bordered pits, which are usually associated with an unspecialized struc- ture. The Zygophyllaceae have many economic uses; Hutchinson (1967, pp. 612, 613) gives a good summary of these (see also Guaiacum below). Kollswviewie hirsutissima Vail in Small and Peganum Harmala are re- ported (cf. Kingsbury, 1964) to poison cattle, and Tribulus terrestris (see below) to poison sheep. Alkaloids have been reported to occur in many genera, and it is these that are suspected in stock poisoning by the species noted above and others in the Old World. REFERENCES: AGABABIAN, V. S. The morphological types of pollen and taxonomy of the family Zygophyllaceae. (In Russian; Armenian summary.) Izv. Akad. Nauk rmian. SSR Biol. 17: 39-45. 1964. [Six basic pollen types occur. ] Artin, T. E. H., & J. R. Cannon. Distribution of alkaloids in some West Aus- tralian plants. Econ. Bot. 25: 366-380. 1971. [Includes Nitraria, Tribulus, & Zygophyllum. | 534 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Asprey, G. F., & P. THortncTon. Medicinal plants of Jamaica. Part II. West Indian Med. Jour. 3: 17-41. 1954. [Guaiacum, Tribulus, 26, 27.] BARNER, J. Die Nutzhélzer der Welt. Vol. 2. vii + 780 pp. Neudamm. 1942. [Zygophyllaceae, 321-327.] BatLey, I. W., & W. W. Tupper. Size variation in tracheary cells. I. A com- parison between the secondary xylems of vascular cryptogams, gymno- sperms, and angiosperms. Proc. Am. Acad. Arts Sci. 54: 149-204. 1918. [Guaiacum, Larrea, 169.] BaILton, H. Observations sur les Rutacées. Adansonia 10: 299-333. 1872. [Zygophyllaceae, 299-301, 312-318. ] tacées. Hist. Pl. 4: 415-423, 502-510. 1873. [English transl. M. M. Hartoc, The natural history of plants. 4: 422-432, 499-451. London. 1875.] . Traité de botanique médicale phanérogamique. Vol. 2. iv+ pp. 465- 1499. Paris. 1884. [Zygophyllaceae, 846, 882-886. | Bate-Smiru, E. C. The phenolic constituents of plants and their taxonomic significance. Jour. Linn. Soc. Bot. 58: 95-173. 1962. [Larrea, 111; Zygo- phyllum, 134.] BENTHAM, G., & J. D. Hooker. Zygophylleae. Gen. Pl. 1: 262-269. 1862; 998. 1867. Bosrov, E. G. Zygophyllaceae. Fl. URSS 14: 146-198, 723-731. 1949. [Mal- acocarpus, Nitraria, Peganum, Tetradiclis, T: ribulus, Zygophyllum. | BOLKHOVSKIKH, Z., V. Grir, T. Martvejeva, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. Ed. by A. Feporov. (Russian and English prefaces.) 926 pp. Leningrad. 1969. [Zygophyllaceae, 727.] Bray, W. L. On the relation of the flora of the Lower Sonoran Zone in North America to the flora of the arid zone of Chili and Argentina. Bot. Gaz. 26: 121-147. 1898. [Many Zygophyllaceae in both areas. | Britton, N. L., & J. A. SHarer. N. Am. Trees. x + 894 pp. New York. 1908. [Guaiacum, Porlieria, 564-566. | Bursipce, N. T. Dictionary of Australian plant genera. xviii + 345 pp. Sydney. 1963. [Kallstroemia, Nitraria, Peganum, Tribulopis, Tribulopsis, Tribulus, Zygophyllum. | Burwace, H. M. Index of plants of Texas with reputed medicinal and poisonous properties. v-+ 272 pp. Austin. 1968. [Kallstroemia, Larrea, Peganum, Porleria, Tribulus. | CANDOLLE, A. P. DE. Zygophylleae. Prodr. 1: 703-708. 1824. Cuattaway, M. M. Crystals in woody tissues; part II. Trop. Woods 104: 100- 124. 1956. [Bulnesia, Guaiacum, Larrea, Nitraria, Porlieria.| Cxopra, R.N., et al. Chopra’s indigenous drugs of India. ed. 2. xxxii + 816 pp. Calcutta. 1958. [Fagonia, 507: Peganum, 368; Tribulus, 430.] Coox, O. F., & G. N. Cottrns. Economic plants of Porto Rico. Contr. U.S. Natl. Herb. 8: 57-269. 1903. [Includes Guaiacum, Kallstroemia, Tribulus.] Cronguist, A. The evolution and classification of flowering plants. x + 396 pp. Boston. 1968. [Zygophyllaceae, 262-269, 274; Balanites in Simaroubaceae. ] Cruse, R. R. Recent highlights in the chemurgy of xerophytic plants. Econ. Bot. 13: 243-260. 1959. [Includes Larrea, Tribulus. | Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. 1966. [Zygophyllaceae, 279.] Descote, H. R., C. A. O’DonELL, & A. Lourreic. Revisién de las Zigofilaceas 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 535 argentinas. Lilloa 5: 257-352. 1940. [Bulnesia, Kallstroemia, Larrea, Plectrocarpa, Porlieria, Tribulus, Zygophyllum. | ——. yeophyllaceae. Gen. Sp. Pl. Argentinarum 1: 1-47. 1943, [Colored plates, tides maps. Don, G. Zygophylleae. Gen. Syst. Gard. Bot. 1: 768-775. 1831. ECKERT, G. Sc eviathes aelnaiiniie und bliitenanatomische Untersuchungen zum Problem der Obdiplostemonie. Bot. Jahrb. 85: 523-604. 1966. [Zygo- phyllaceae, 547-550. ] Euretiicu, P. R., & P. H. Raven. Butterflies and plants: A study in coevolution. Evolution 18: 586-608. 1964. [Larvae of Kricogonia feed on Guaiacum, those of Hypermnestra on Zygophyllum. | Ex-Hapipy, M. N. The pollen morphology of some Egyptian xerophytes. Bull. Fac. Sci. Cairo 34: 67-71. ENDLICHER, S. Zygophylleae. Gen. Pl. 1161-1166. 1840. Encter, A. Zygophyllaceae. Nat. Pflanzenfam. III. 4: 74-93. 1890; 353-357. 1896; ed. 2. 19a: 144-184. : _ Uber die geographische Verbreitung der Zygophyllaceen im Verhiltniss zu bre systematischen Gliederung. Abh. Akad. Wiss. Berlin Phys. 2: 1-36. 1896. [Distribution maps. | ErptMan, G. Pollen morphology and plant taxonomy. Angiosperms. Corrected reprint and new addendum. xiv + 553 pp. New York. 1966. [Zygophyl- laceae, 456-458; references to earlier papers. | GOMEz DE LA Maza, M. Ensayo de farmacofitologia cubana. Havana. 1889. [Guaiacum, Kallstroemia, Tribulus, 34, 35.] Gray, A. Contributions to American botany. 1. Revision of some Saget genera and orders precursory to the Flora of North America. Proc Acad. Arts Sci. 22:270-306. 1887. [Guaiacum, Kallstroemia, 306. | Zygophyllaceae. Synop. Fl. N. Am. 1: 352-356. 1897. [Notes by B. L. Rosrnson. | Gray, C. A., & H. W. BENNETTS. The toxic plants of Western Australia. Perth. 1956. [Nitraria, Tribulus, Zygophyllum, 120-122. | Hermscu, C., Jr. Comparative anatomy of the secondary xylem of the “Gruin- ales” sind “Terebinthales” of Wettstein with reference to taxonomic group- ing. Lilloa 8: 83-198. 1942. [Zygophyllaceae, 104-106; relationships with Linaceae and other families. Hevty, R. H., P. J. MEHRINGER, Jr., & H. G. Yocum. Modern pollen rain in the Sonoran Desert. Jour. Arizona Acad. Sci. 3: 123-135. 1965. [Kallstroemia, Larrea. | Hurcuinson, J. The genera of flowering plants. Vol. 2. xi+ 659 pp. London. 1967. [Zygophyllaceae, 268-270; excludes Balanites. Evolution and phylogeny of flowering plants. Dicotyledons: facts and theory. xxvi-++ 717 pp. 2 charts. London & New York. 1969. [Zygo- phyllaceae, 272-275; Sihies Balanites. | Inampar, J. A. E idermal structure, stomatal ontogeny and relationship of some Zygophyllaceae and Simaroubaceae. Flora B. 158: 360-368. 1969, J Guaiacum, Tribulus Jounson, B. Pollenmorfologien a ern Zygophyllaceae. Redog. Larov. Bromma Appendix 1951-52: 4, 5. 1952 permet peerst ae poucicabal survey of the Zygophyllaceae. vi+ 77 MSS. thesis, Western Illinois University, Macomb. 1969. [Three basic alin types occur. | 536 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Kine, L. J. Weeds of the world, biology and control. xxxii+ 526 pp. New York. 1966. [Peganum, Tribulus, Zygophyllum Kincssury, J. M. Poisonous plants of the United States and Canada. xiii+ 626 pp. Englewood Cliffs, New Jersey. 1964. [Kallstroemia, Peganum, Tribulus, 201-205.] Kuntze, O. Zygophyllaceae. Rev. Gen. Pl. 3: 29, 30. 1893. Lasser, T. Las zygophyllaceas venezolanas. Bull. Jard. Bot. Bruxelles 27: 381- 390. 1957. [Bulnesia, Guaiacum, Kallstroemia, Tribulus Linney, J. Zygophylleae. Nat. Syst. Bot. Ed. 2. 133, 134. 1836. Lussock, J. A contribution to our knowledge of seedlings. Vol. 1. viii + 608 pp. London & New York. 1892. [Zygophylleae, 294. | Martin, P. S., & C. M. Drew. Additional scattering electron photomicrographs of Southwesterti pollen erains, Jour. Arizona Acad. Sci. 6: 140-161. 1970. [ Kallstroemia, Tribulus Martinez, M. Las plantas mis ltiles que existen en la Republica Méxicana. ed. 3. Mexico City. 1959a. [Guaiacum, sina Larrea, 268-270. | . Plantas medicinales de México. ed. 4. 657 pp. Mexico City. 1959b. [ Guaiacum, 156, 157, 561, 597; Larrea, es 144; Tribulus, 369. Mauvritzon, J. Zur Esnbryolesie einiger Gralnales: Sv. Bot. Tidskr. 28: 84-102. .| 1934a. [Tribulus, Zygophyllum, 98- Etwas iiber die Embryologie der Zygophyllaceen sowie einige Fragmente iiber die der Humiriaceen. Bot. Not. 1934: 409-422. 1934b. MeEtTcaLFE, C. R., & L. CHALK. Zygophyllaceae. Anat. Dicot. 1: 285-292. 1950. Nar, N. C., & K. S. NATHAWAT. Vascular anatomy of the flowers of some species of Zygophyllaceae. Jour. Indian Bot. Soc. 37: 172-180. 1958. [Fagonia, Nitraria, Peganum, Tribulus, Zygophyllum. | Neat, M. C. In gardens of Hawaii, rev. ed. xix+ 924 pp. Honolulu. 1965. [Guaiacum, Tribulus, 475-477.]| Necop1, G. Lineamenti sulla cariologia delle Rutaceae e delle Zygophyllaceae. Arch. Bot. Forli 13: 93-102. 1937. [Peganum, Tribulus.] ———. Cariologia della Rutaceae e delle Zygophyllaceae. Sci. Genet. 1: 168- 185. 1939. [Fagonia, Tribulus. | Ozenpa, P., & P. QuezeLt. Les Zygophyllacées de l’Afrique du Nord et du hara. Trav. Inst. Rech. Sahariennes Univ. Algiers 14: 23-83. 1956. PEREZ ARBELAEZ, E. Plantas utiles de Colombia. 831 pp. Bogota. 1956. [Zygo- phyllaceae, 751-753.] Prittier, H. Manual de las plantas usuales de Venezuela. xvi + 458 pp. Caracas. 1926. [Bulnesia, 395; Guaiacum, 251, 252; Kallstroemia, 405; Tribulus, 93.] Porter, D. M. The taxonomy and distribution of the Zygophyllaceae of Baja California, Mexico. Contr. Gray Herb. 192: 99-135. 1963. [Fagonia, Guaiacum, Kallstroemia, Larrea, Tribulus, Viscainoa.| : Zyeophyllaceae. Flora of Panama, part VI, family 88. Ann. Missouri Bot. Gard. 56: 1-7. 1969. [Guaiacum, Kallstroemia, Tribulus. | Zygophyllaceae. Pp. 901-906 in D. S. Corrett & M. C. JOHNSTON, Manual of the vascular plants of Texas. xv + 1881 pp. Frontisp. 1 map. enner, Texas. Zygophyllaceae. “Pp. 772-778 in I. L. Wicctns & D. M. PorTER, Flora of the Galapagos Islands. xx + 998 pp. Stanford, Calif. 1971. [Kallstroemia, Tribulus. | Rau, M. A. Review of recent work on the embryogeny of some families and 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 537 genera of disputed systematic position. Pp. 75-80 in Plant embryology, a symposium. New Delhi. 1962. [Peganum, Tribulus, Zygophyllum. | Recorp, S. J. Storied or tier-like structure of certain dicotyledonous woods. Bull. Torrey Bot. Club 46: 253-273. 1919. [Bulnesia, Guaiacum, Larrea, Por- lieria, 272. See also Trop. Woods 9: 13-18. 1927.] . Lignum-vitae: A study of the woods of the Zygophyllaceae with refer- ence to the true lignum-vitae of commerce. Yale School Forestry Bull. 6: 1-48. 1921. [Bulnesia, Guaiacum, Porlieria. | . Keys to American woods (continued). Jbid. 73: 23-42. 1943 [Zygo- phyllaceae, 24]; Ibid. 74: 17-43. 1943. [Guaiacum, Porlieria, 34]; Ibid. 76: 32-47. 1943. [Bulnesia, Guaiacum, Larrea, Porlieria]; Ibid. 77: 18-38. 1944 [Bulnesia, Larrea, Porlieria]. _W. Hess. Timbers of the New World. xv + 640 pp. pls. 1-58. New Haven, Conn. 1943. [Zygophyllaceae, 554-559. ] Riwtey, H. N. The dispersal of plants throughout the world. xx + 744 pp. 22 pls. Ashford, England. 1930. [Nitraria, 359; Tribulus, 584, 585.] Rossins, W. W., M. K. BeLtve, & W. S. BALL. Weeds of California, ed. 2. Sacramento. 1951. [Kallstroemia, Tribulus, Zygophyllum, 287-291.] Ror, J. Phytogeography of Central Asia. Bull. Fan Mem. Inst. Biol. Bot. 11: -35. 1941. [Peganum, Tribulus, Zygophyllum. | Ross, J. N. Notes on useful plants of Mexico. Contr. U.S. Natl. Herb. 5: 209- 259. 1899. [Guaiacum, 255; Larrea, 229.] Sapet, Y. S. Bacterial nodules in the Zygophyllaceae. Nature 157: 656, 657. 1946. [Fagonia, Tribulus, Zygophyllum. | SaBnis, T. S. The physiological anatomy of the plants of the Indian desert. Jour. Indian Bot. Soc. 1: 183-205. 1920. [Fagonia, Seetzenia, Tribulus, Zygophyllum, 183-187. ] Saunpers, E. R. On carpel polymorphism. VI. Ann. Bot. 48: 643-692. 1934. [Augea, Guaiacum, Peganum, Tribulus, Zygophyllum, 673-683.) . Floral morphology. Vol. 1. viii + 132 pp. Cambridge, England. 1937. [Guaiacum, Peganum, Tribulus, Zygophyllum, 86-88.] Scurerser, A, Zygophyllaceae. Prodr. Fi. Siidwestafr. 4(65): 1-19. 1966. Scuoiz, H. Zygophyllaceae. Jn: H. MELCHIOR, Engler’s Syllabus der Pflanzen- familien. ed. 12. 2: 251, 252. 1964. [Includes Balanites. | Suaw, F. H., et al. A phytochemical register of Australian plants. Vol. 1. 294 pp. Melbourne. 1959. [Tribulus, Zygophyllum, 189.] Stapkov, A. N. Description morphologique du pollen des Zygophyllaceae de Turkménie. (In Russian.) Trudy Inst. Geogr. Akad. Nauk SSSR 61: 157-167. 1954.* Smitey, N. Tropical planting and gardening for South Florida and the West Indies. 202 pp. Coral Gables. 1960. [Guaiacum, Tribulus. | Sperry, O. E. Poisonous range plants in Texas. IV. Poisonous plants of the cal- trop family. Sheep Goat Raisers’ Mag. 36(12): 18. 1956.* Steenis, C. G. G. J. VAN. Zygophyllaceae. Fl. Males. I. 4: 64. 1949. [Fagonia, Tribulus. | Taxutasan, A. Die Evolution der Angiospermen. viii + 344 pp. Jena. 1959. [Zygophyllaceae, 61.] phe Flowering plants, origin and dispersal. (Transl. C. JerrRey.) x +310 pp. Edinburgh. 1969. [Zygophyllaceae, 226; excludes Balanites, Nitraria, Peganum. | 538 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 THORNE, R. F. Synopsis of a putatively phylogenetic classification of the flowering plants. Aliso 6: 57-66. 1968. [Zygophyllaceae, 61, 62; excludes es. —S L. A. Maderas y bosques argentinas. xxvii-+910 pp. Buenos Aires. 1956. [Bulnesia, Larrea, Guaiacum, Porlieria. VarL, A. M. & P. A. RypBerc. Zygophyllaceae. N. Am. Fl. 25: 103-116. 1910. [Fagonia, Guaiacum, Kallstroemia, Larrea (Covillea), Morkillia, Peganum, Porlieria, Sericodes, T ribulus, Viscainoa. See also Bull. Torrey Bot. Club 22: 228-231. 1895.] VASILYENN, I. T. Weed seedlings. (In Russian.) Leningrad. 1965. [Peganum, Tribulus, Zygophyllum, 229-231.] Wareurs, E. F. Taxonomy and relationship in the Geraniales in the light of their cytology. New Phytol. 37: 130-159, 189-210. 1938. [Zygophyllaceae, 205-208. ] Wart, J. M., & M. G. Breyer-Branpwijk. The medicinal and poisonous plants of southern Africa. Edinburgh. 1932. [Tribulus, Zygophyllum, 85-88.] WEBERLING, F. Weitere Untersuchungen zur Morphologie des Unterblattes bei den Dikotylen. III. Convolvulaceae. V. Zygophyllaceae. Beitr. Biol. Pflanzen 33: 149-161. 1957. [Zygophyllaceae, 150-161.] WELKIE, G. W., & M. CaLpweELL. Leaf anatomy of species in some dicotyledon families as related to the C, and C, pathways of carbon fixation. Canad. Jour. Bot. 48: 2135-2146. 1970. [Kallstroemia & Tribulus have specialized leaf anatomy associated with C, pathway; Fagonia, Larrea, Peganum, Porlieria, & Zygophyllum with C, pathway. ] WittamaN, J. J., & H. L. Li. Alkaloid-bearing plants and their contained alka- loids, 1957-1968. Lloydia 33(3a). vii + 286 pp. 1970. [Zygophyllaceae, ——, & B. G. Scuusert. Alkaloid hunting, Econ. Bot. 9: 141-150. 1955. [Zygophyllaceae, 3 genera give positive reactions, 3 negative & . Alkaloid-bearing plants and their contained alkalol ds. US. i Agr. Agr. Res. Serv. Tech. Bull. 1234. 287 pp. 1961. [Zygophyllaceae, eee SS R. O., & R. O. WitttaMs, Jr. The useful and ornamental plants in Trinidad and Tobago. ed. 4. 1951. [Guaiacum, 131; Tribulus, 301, 302.] Wars, 1 Cc. dictionary of the flowering plants and ferns. ed. 7. (Revised by Y SHAW.) xii + 1214 pp. + liii pp. (Key to the families of flower- ing sents.) Cambridge, England. 1966. [Zygophyllaceae, 1213.] Wvyo.er, H. Ueber die symmetrische Verzweigungsweise dichotomer Inflorescen- zen. Flora 34: 353-365, 385-398, 401-412, 417-426, 433-448. pls. 7-9. 1851. [Peganum, Tribulus, 360, 361.] ZAITSCHEK, D. V., et al. The saponin content of Israeli Zygophyllaceae. Lloydia 34: 163, 164. 1971, [12 spp. in 6 genera; saponins in 4 spp. of Fagonia ZumsrucH, H. J. Ueber die Bedeutung des Saponins fiir die svsteniatische Gliederung der Zygophyllaceen. Ph. D. Diss., Berlin. 1931.* Key To THE GENERA OF ZYGOPHYLLACEAE IN THE SOUTHEASTERN UNITED STATES General -characters: Nodes angled or swollen; leaves opposite, even-pinnately compound, stipulate; flowers terminal or pseudazillary, 5-merous, hypogynous, perfect, with an extrastaminal and/or intrastaminal disc. 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 539 A. Plants herbaceous; flowers yellow; stamens unappendaged; fruit a schizo- Carp separating at maturity into 5 or i B. Fruit tuberculate, at maturity separating into 10 mericarps, beak per- sistent on receptacle; intrastaminal glands absent. .. 1. Kallstroemia. B. Fruit spiny, at maturity separating into 5 mericarps, beak falling with mericarps; intrastaminal glands present. ................ 2. Tribulus. Plants woody (trees or shrubs); flowers blue or purple; stamens append- aged; fruit a 5-lobed septicidal capsule. .................. . Guaiacum, pe Subfam. TRIBULOIDEAE Porter 1. Kallstroemia Scopoli, Introd. Hist. Nat. 212. 1777. Annual [occasionally perennial] herbs; stems herbaceous [to suffrutes- cent], diffusely branched, prostrate to decumbent or ascending, terete, somewhat succulent, becoming striate on drying, densely pubescent to glabrate, spreading radially from a central tap root to 1 m. or more long. Leaves opposite, elliptic to broadly obovate, abruptly even-pinnate, one of each pair alternatingly smaller or sometimes abortive; leaflets (2) 3-5 (6)[-10] pairs, opposite, entire, subsessile, elliptic to broadly oblong or obovate, somewhat unequal in size, those on one side of rachis slightly smaller, lowest pair markedly unequal, terminal pair directed forward and more falcate, pubescent to glabrate; stipules foliaceous. Flowers solitary, pseudaxillary, regular; peduncles emerging from axils of alternately smaller leaves. Sepals 5 [6], free, pubescent, persistent [rarely caducous]. Petals 5 [6], white to orange, the same basally or green to brighter than apex, free, spreading (the corolla cuplike), fugacious, [usually] mar- cescent, convolute. Disc fleshy, annular, obscurely 10[—12]-lobed. Sta- mens 10 [12], the 5 [6] exterior ones opposite petals, somewhat longer and adnate basally to petals, the 5 [6] opposite sepals subtended abaxi- ally by a small bilobed gland; filaments filiform or subulate [rarely basally winged], unappendaged, inserted on disc; anthers globose or ovoid [to linear-oblong] or rarely linear, those opposite sepals rarely aborting. Style simple, cylindrical to [broadly] conical, more or less 10[—12]-ridged, persisting to form a beak on fruit; stigma capitate or oblong [or clavate], 10[—12]-ridged or -lobed, papillose [rarely coarsely canescent], terminal [rarely extending down almost to base of style]; ovary sessile, 10[—12] -lobed and -loculate, ovoid or pyramidal [globose or occasionally conical], glabrous to pubescent; ovules 1 per locule, pendulous, placentation axile, sometimes 1 or more aborting. Fruit 10[—12]-lobed, ovoid [occasionally conical or rarely pyramidal], glabrous or pubescent, at maturity dividing septicidally and separating from a persistent styliferous axis into 10[-12] or occasionally fewer mericarps; mericarps 1-loculate, 1-seeded, obliquely triangular, wedge shaped, more or less tuberculate or rugose abaxially. Seeds oblong-ovoid, obliquely pendulous; seed coat membranaceous; embryo straight; endosperm absent; germination epigeal. TyPE SPECIES: Kallstroemia maxima (L.) Hooker & Arn, (Tribulus maximus L.). (Origin of name unknown, although it has been suggested that it is derived from that of Anders Kallstrém, a contemporary of Scopoli.) 540 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 The largest New World genus of the family, with 17 species (see Porter, 1969a), these ranging from Illinois to Argentina, in open, disturbed habi- tats. Even in the tropics, Kallstroemia species invariably will be found in the driest localities of a particular area. Nine species are basically of a southwestern United States-northern Mexican distribution, and the least specialized species of the genus, K. perennans Turner, occurs in the Big Bend region of Texas. Three species are basically Caribbean in distribu- tion, and five are South American. Three species occur in our area. The most common Kallstroemia in the Southeast is K. maxima. This primarily Caribbean species occurs from Colombia and Venezuela north to Sinaloa, Mexico, on the west, and to South Carolina, on the east, with a waif found as far north as Pennsylvania (Porter, 1970). In our area it has been collected on sandy, disturbed soils near the sea in Florida, Georgia, and South Carolina. It is easily distinguished by its glabrous fruits. Kallstroemia parviflora Norton has been collected once in Oktibbeha County, Mississippi, where it apparently did not form a persistent colony. It has also been found as a waif in Maryland (Porter, 1970). It is indige- nous to grassy, disturbed areas from Illinois, west to Arizona, and south to central Mexico, and it is an introduced weed in Peru. The strigose to glabrous beak on the fruit is two to three times as long as the ovoid, strigose body. Kallstroemia pubescens (G. Don) Dandy in Keay, 2” = 32, is known in the United States only from Apalachicola, Florida, where it may or may not have persisted. This basically Caribbean species, known from Peru northward to Sinaloa, Mexico, and the Antilles, was first described from what is now Ghana. It has also been collected in the Ivory Coast and in India (Bennett, 1965). In the Greater Antilles it is known only from Jamaica (Proctor, 1967) and Puerto Rico. The short-pilose to glabrous beak on the fruit is as long as the pyramidal, densely appressed short- pilose body Four species grown in the greenhouse have been found to be self-com- patible (Porter, 1969a), a decided advantage for weedy plants. They (and apparently all others, except Kallstroemia perennans) also can be self-pollinated by an upward and inward movement of the stamens and petals as the flower, which is open in the morning, closes in early after- noon. Little is known of natural relationships within the genus. The only chromosome number known is that of Kallstroemia pubescens (Datta, 1968). Hybridization between K. maxima and K. pubescens, in Colombia, and between K. maxima and K. Rosei Rydb. in Vail & Rydb., in Mexico, is suspected (Porter, 1969a). REFERENCES: Under family references see AGABABIAN; BRAY; BURBIDGE; BIeRAGK, Cook & Cottins; Descote, O’DoNnELL, & Lourteic; ENGLER; ; G6 DE 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 541 LA Maza; Gray; Heviy, MEHRINGER, & Yocum; KaLicis-WALALANGI; KINGs- BURY; Lasser; MaArtTIN & Drew; Mauritzon, 1934b; PEREZ ARBELAEZ; Pirtier; PorTER; Roppins, BELLUE, & BALL; VAtL & RypBerc; and WELKIE & CALDWELL. At-Katis, Y. Comparative morphological study of the Arizona species of Kall- stroemia Scop. M.S. Thesis, Univ. Arizona, Tucson. 1958. [K. californica, K. grandiflora, K. hirsutissima, K. parviflora. | BENNETT, S. S. R. Genus Kallstroemia Scop. (Zygophyllaceae)— new to Indian flora. Indian Forester 91: 281-283. 1965. [K. pubescens. ] Datta, N. Cytological investigations in Kallstroemia pubescens (Don) Dandy. Curr. Sci. Bangalore 37: 477. 1968. Hevty, R. H., & P. S. Martin. Geochronology of pluvial Lake Cochise, southern Arizona. I. Pollen analysis of shore deposits. Jour. Arizona Acad. Sci. 2: 24-31. 1961. Matuews, F. P. The toxicity of Kallstroemia hirsutissima (carpet weed) for cattle, sheep, and goats. Jour. Am. Vet. Med. Assoc. 105: 152-155. 1944.* Prirrier, H. Ensayo sobre plantas usuales de Costa Rica. ed. 2. 264 pp. San José. 1957. [Crushed leaves of K. maxima used to bring boils and other abscesses to a head. ] Porter, D. M. The genus Kallstroemia (Zygophyllaceae). Contr. Gray Herb. 198: 41-153. 1969a. [17 species, all native to New World, recognized. . Kallstroemia (Zygophyllaceae) in Missouri. Ann. Missouri Bot. Gard. 56: 290. 1969b. [K. parviflora the correct name for the Missouri species. ] _ Kallstroemia in the Middle Atlantic States. Rhodora 72: 397, 398. 1970. [K. maxima in Pennsylvania, K. parviflora in Maryland. | Proctor, G. R. Additions to the flora of Jamaica. Bull. Inst. Jamaica Sci. 16: 1-84. 1967. [K. pubescens, 16.] Ruiz Leat, A. Dos zigofilaceas nuevas para la flora cuyana. Haumania 1: 11- 14, 1947. [New localities in Argentina for K. tribuloides and K. tucuman- ensis. | —. Kallstroemia tucumanensis Desc., O’Don. et Lourt. (Zygophyllaceae) en las inmediaciones del Cerro de la Gloria. Revista Fac. Ci. Agrar. Univ. Nac. Cuyo 3: 27-29. 1951. [Spread in Argentina by man. | 2. Tribulus Linnaeus, Sp. Pl. 1: 386. 1753; Gen. Pl. ed. 5, 183. 1754. Annual or occasionally perennial herbs [very rarely shrubby]; stems herbaceous to suffrutescent, diffusely branched, prostrate to decumbent or ascending, terete, somewhat succulent, becoming striate on drying, densely pubescent to glabrate, spreading radially from a central tap root to 1[-3] m. long. Leaves opposite, elliptical in outline, abruptly even-pinnate, one of each pair alternately smaller or sometimes aborting; leaflets 3—7[-10] pairs, opposite, entire, sessile to very shortly petiolulate, oblong to ovate or elliptical, terminal pair pointed forward, pubescent; stipules foliaceous. Flowers solitary, pseudaxillary, regular; peduncles emerging from axils of alternately smaller leaves. Sepals 5, free, pubescent, caducous. Petals 5, bright yellow or rarely white, darker basally, free, spreading (the corolla cuplike), deciduous, imbricate. Disc fleshy, 10-lobed, annular. Stamens 10, the outer whorl of 5 opposite petals, somewhat longer, and adnate [ VoL. 53 Ficure 1. Tribulus. a—h, T. stasis a, tip of Sheba ets with flower and ievelooius fruits, X 1/2: b, flower, X 1 1/2; c, gynoecium and staminal glands, the intrastaminal ones united around base of o male sen pretreat ones subtending stamens abit sepals, the positions of stamens, oon! and sepals ded, hairs on ovary not shown, X 6; d, flower in semidiagrammatic vertical en to show siecle and positions of stamens, glands, ssa ts nd petals — base of Sepal and stamen opposite it to left, base of petal and stamen oppo- ti of fruit in median vertical section showing four seeds and bia stony part of mature mericarp, softer outer tissues prin be x 2; h, seed, 5 "3 i, embryo, basally to petals, the inner 5 opposite sepals and subtended both ad- and abaxially by nectariferous glands, the intrastaminal glands free or connate Style simple, stout, cylindrical, 5-ridged, deciduous; stigma terminal, pyramidal or globose to inequilateral, 5-lobed, papillose; ovary sessile, 5- lobed and -loculate, ovoid or globose, densely pubescent; ovules 3-5 per locule, pendulous, superposed in 2 vertical rows on placenta, placentation axile. Fruit 5-angled, usually horizontally depressed, pubescent, at ma- turity dividing septicidally and separating into 5 or rarely fewer mericarps 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 543 and leaving no central axis; mericarps broadly triangular, each divided internally by oblique transverse septa into 2—5 one-seeded compartments, spiny [or winged or rarely only tuberculate abaxially]. Seeds oblong-ovoid, obliquely pendulous, horizontally arranged one above the other; seed coat membranaceous; embryo straight; endosperm absent; germination hypo- geal. Lectotype species: Tribulus terrestris L.; see Vail & Rydberg, N. Am. FI. 25: 109. 1910. (Name from Greek tribolos, a kind of caltrop, an iron instrument with four spines arranged so that one always projects up- ward, used to impede cavalry.) — PUNCTURE WEED, CALTROP, BURR-NUT. An Old World genus of several dozen species, commonest in Africa and the Near East. Most are weedy occupants of dry, disturbed habitats, either natural or man induced. The spiny mericarps of many species pro- vide an ideal mechanism for dissemination, and man has long been involved in their spread. Three species (two represented in our area) have been introduced into the New World. Tribulus cistoides L., burr-nut, 2n = 12, is common in southern Florida, and it is also known from northwestern Florida, Georgia, and Louisiana. Native to tropical and subtropical southern Africa, it is now a weed throughout the drier tropics, mainly in maritime habitats. It is an attrac- tive plant with showy yellow flowers two to four centimeters in diameter that have led to the common name “Jamaica buttercup.” The plant is occasionally grown in sandy soils as a garden ornamental or along roads to stabilize shifting soils. Tribulus terrestris L., puncture vine, is a noxious roadside weed most common in the West and Midwest. In our area it has been collected sporadically in Florida, Louisiana, South Carolina, and Tennessee, and more commonly in Arkansas. It is native to the Mediterranean region, probably to northern Africa. A widespread weed in the warm-temperate areas of the world, it occurs on all continents but Antarctica but has been found only rarely in the tropics. The flowers are one centimeter or less in diameter. there (Porter, 1971a & b). After trying unsuccessfully for some 50 years to control Tribulus terrestris by chemical means, the California Department of Agriculture in 1961 imported two species of weevil from India for biological control. The larvae of both feed selectively on T. terrestris, Microlarinus lareynii (Jacquelin du Val) on the seeds and M. /ypriformis (Wollaston) on the 544 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 stems. These weevils appear to be well on the way to controlling T. ter- restris in the Southwest. In Africa, Australia, and the United States ingestion of the plant by sheep leads to “geeldikkop” or “bighead,” a fatal disease involving hepato- genic photosensitization. Both nitrate and selenium poisoning may be in- volved (see Kingsbury, above). Diploid (2n = 12), tetraploid (20 = 24), hexaploid (2n = 36), and octoploid (2m = 48) races have been reported for Tribulus terrestris, but only diploids have been found in T. cistoides. According to Malik (1966) the races of T. terrestris can be distinguished by pollen and flower size. Chromosome numbers are not known for any other species of this taxonom- ically difficult and poorly-known genus. REFERENCES: Under family references see AGABABIAN; APLIN & CANNON; ASPREY & THOR- INGTON; Bosrov; BuRBIDGE; BURLAGE; CHopRA, et al.; SOR & COLLINS; ni ape O’DonELL, & Lourretc: ENGLER; ERDTMAN - GOMEZ DE LA Maz y & BENNETTS; INAMDAR; KALIGIs-WALALANGI; KING; seen? oa MartTIN; Marrinez, 1959b: Mauritzon; METCALFE & CHALK; Nar & NaTHAwar: NEAL; Necopr; OzENDA & QUEZEL; PEREZ ARBELAEZ: PITTIER; PorTER; RAu; RIDLEY; Rozsins, BELLUE & BALL; Ror; SABET; SABNIS: SAUNDERS; " SCHREIBER; SHAW; SMILEY; VAN STEENIS; VAIL & Rypeer G; VASILYENN; Warr & BREYER-BRANDWIJK; WEBERLING; WELKIE & Spang ’ WILLIAMS & WILLIAMS; and WYDLER ALLEN, E. K., &O. N. Atten. The anatomy of the nodular growths on the roots of Tribulus cistoides L. Soil Sci. Am. Proc. 14: 179-183. 1950. [No evi- dence that nodules are caused by microorganisms. Anores, L. A., & G. W. ANGALET. Notes on the ecology and host specificity of Microlarinus lareynii and M. lypriformis (Coleoptera: Curculionidae) and the biological control of puncture vine, Tribulus terrestris. Jour. Econ. Entomol. 56: 333-340. 1963. Austin, D. F. Interactions between Apis mellifera (Hymenoptera: Apidae) and Tribulus cistoides (Zygophyllaceae). Rhodora 74: 117-123. 1972. [Nectar and pollen 3 ering. | BorkowskI, B., & J. Luromsxr. Chromatographic examination of the alkaloid fraction from the herb and seeds of Tribulus terrestris. Biul. Inst. Roslin Leczniczych 6: 220-227. 1960.* Brown, W. H. The bearing of nectaries on the phylogeny of flowering plants. Proc. Am. Philos. Soc. 79: 549-595. 1938. [T. cistoides, 570, fig. 83.] Browne, E. M., & E. T. Browne, Jr. Flowering plants new to or rare in Ken- tucky. Rhodora 67: 180, 181. 1965. [T. terrestris, first report.] Crookston, R. K., & D. N. Moss. The relation of carbon dioxide compensation and chlorenchymatous vascular bundle sheaths in leaves of dicots. Pl. Physiol. 46: 564-567. 1970. [Low CO, compensation point and chloren- chymatous vascular bundle sheaths in T. terrestris, C, system traits. ] Davipson, A. New plant records for Los Angeles County, Part II. Bull. So. Calif. Acad. Sci. 2: 43. 1903. [T. terrestris, first report from California.] 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 545 DeceEttus, G. Plants from Turkey introduced with manganese ore into a place in Vastergotland (Sweden). Bot. Not. 112: 121-133. 1959. [T. terrestris.] Dittmer, H. J. A comparative study of the number and length of roots pro- duced in nineteen angiosperm species. Bot. Gaz. 109: 354-358. 1948. [In- cludes 7. terrestris. | Forepar, J. L., & S. K. Roy. Tetraploid Tribulus Linn. Sci. Cult. 35: 25, 26. 1969. [T. terrestris, 2n = 48. peapEMe si G. W., & A. L. Harenricuter. Anthokinetics. The — and cology of flower movements. Carnegie Inst. Publ. 420: 1-198. 1932. [T. pvsiapnen 129, pl. 15; flower opens about 9 A.M., closes in early afternoon. ] GRUENBERG-FERTIG, [., & M. Zowary. Nedssaclatand remarks on some paws of Palestine. Part Q Israel Jour. Bot. 19: 293-304. 1970. [T. alatus Del. T. longipetalus _ introduced into Peru, see PorTER (1967) ile Gurtiarmop, A. J. Ac ixibution towards the economic botany of Basutoland. Bot. Not. 119: nts 1966. [T. terrestris leaves used as a potherb, 211.] Guppy, H. B. Observations of a naturalist in the Pacific between 1896 and 1899, vol. 2. Plant dispersal. xxviii + 627 pp. 1906. [Tribulus cistoides, 552.] Henrict, M. Further nutrition studies on Tribulus terrestris. Sci. Bull. Dep. : Agr. S. Afr. 348: 1-60. 1953.* Howe tt, J. T. Errata. Leafl. West. Bot. 10: 363. 1966. [A presumed Cali- fornia collection of T. cistoides actually from the Dominican Republic; see TWISSELMANN (1963) below. ] Hsu, C. C., G. Opett, & T. WILLIAMs. Characterization of the saponin fraction of Tribulus terrestris. Proc. Okla. Acad. Sci. 47: 21-24. Hurraker, C. B., D. W. Riker, & C. E. KENNETT. Biological control of punc- ture vine with imported weevils. Calif. Age, 15(i2)t,.11, 12,196), [7. terrestris control by Microlarinus spp.] Hvanc, T. C. Pollen grains of Formosan plants (2). Taiwania 13: 15-110. 1967. [T. terrestris, 109.] Hurst, E. The poison plants of New South Wales. xiv +498 pp. Sydney. 1942. [T. terrestris, 206-209. ] Ixus, M. Pollen grains of Japan. 303 pp. pls. 1-76. T okyo. 1956. [T. terrestris, 97, spi a 2e B. L. Uber die Wurzelkndllchen bei Tribulus naib gh (In : German summary.) Bull. Jard. Bot. St. -Pétersb. 13: 23-31. 1913. [Root hogies perhaps caused by mycorrhiza; but see ALLEN : ALLEN above. JoHaNsEN, D. A. Plant embryology. Embryogeny of the spermatophyta. xvi + 305 pp. Waltham, Mass. 1950. [T. terrestris, 164. Jounson, E. The spread of the puncture vine in California. Mon. Bull. Calif. Dep. Agr. 9: 330-332. 1920.* [See also ibid. 14: 191. 1925; 16: 354. 1927.* The puncture vine in California. Univ. Calif. Coll. Agr. Agr. Exp. Sta. Bull. 528: 1-42. 19 a <6 . Puncture vine. Jn: W. S. BALL, ef ie Weed control. Calif. Agr. Ext. Serv. Circ. 97: meres 1936. [History, germination, control. | Josui, M. C., S. Kanate, & S. BISHNOT. alscoieteal studies on Rajasthan Desert plants. Il. Tribulus terrestris Linn, Jour. Indian Bot. Soc. 46: 169-184. 1967. , C. P. Corrected basic chromosome number and intraspecific polyploidy in ’ Tribulus terrestris Linn. CIS. Chromosome Inf. Serv. 7: 7, 8. 1966. 546 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 [Races of different ploidy levels can be distinguished by differences in pollen and flower sizes; x = MILtspauGH, C. F. Vegetation of Alacran Reef. Publ. Field Mus. Bot. 2: 421- 431. 1916. [T. cistoides, as T. alacranensis Millsp., 428.] Morton, J. F. A survey of medicinal plants in Curacao. Econ. Bot. 22: 87-102. 1968. [T. cistoides, root decoctions given to children when teeth hurt, 94.] PaTuak, P. S. Effect of sub-freezing temperatures (—15° C) on survival and subsequent growth of Tribulus terrestris Linn. Pp. 66-69 in R. Y. Roy & B. GopaL, eds., Professor R. Misra Commemoration Volume. Varanasi, India. 1967.* . Biological control of Tribulus terrestris by an insect of Hemiptera. Pp. 697-701 in R. Misra & B. Gopat, eds., Proceedings of the Symposium on Recent advances in tropical ecology. Varanasi, India. 1968. . Growth of Tribulus terrestris Linn. at reduced light intensities. Trop. Ecol. India 10: 240-255. 9. ——. Factors affecting seed production of Tribulus terrestris Linn, Ibid. 12: 228-236. 1971. [Inhibited by low temperature, low light intensity, wet soil. Porter, D. M. Another Tribulus adventive in the New World. Rhodora 69: 455, 456. 1967. [T. alatus = T. longipetalus (see GRUENBERG-FeERTIG & ZOHARY above) introduced into Peru.] . The basic chromosome number in Tribulus (Zygophyllaceae). Wasmann Jour. Biol. 26: 5, 6. 1968. [x = 6, based on T. cistoides; see MALIK above. ] ———. Notes on the floral gland morphology in Tribulus srdecramnipeon Ann. Missouri Bot. Gard. 58: 1-5. 1971. [Possible hybridization betwe : cistoides and T. terrestris. See also abstract, Am Jour. Bot. 58: 166. 1971. PoweLL, R. A. Harmful plant species entering New Zealand 1963-1967. New Zealand Jour. Bot. 6: 395-401. 1968. [T. terrestris.] RAMACHANDRAN, K., & P. I. nae ree: “e hexaploid Tribulus terrestris Linn. Sci. Cult. 36: 421, 422. 1970. [2n = Rosertson, E. T., & E. B. Gooptnc. ele for the Caribbean. x + 246 pp. London. 1963. [T. cistoides, 173-176, fig. 84.] ScumuTz, E. M., B. N. Freeman, & R. E. Reep. Livestock-poisoning plants of Arizona. 176 pp. Tucson, Arizona. 1968. [T. terrestris, 118; good illustra- tion of plant, flower, and fruit, 119.] SCHNETTER, R. Blattemperatur und Transpirationsmessungen an Tribulus cis- toides L. in Nord-Kolumbien. Ber. Deutsch. Bot. Ges. 82: 283-286. 1969. . Untersuchungen zum Wirme- und Wasseraushalt ausgewahlter Pflanzen- arten des Trockengebietes von Santa Marta (Kolumbien). Beitr. Biol. Pflanzen 47; 155-213. 1971. [Includes T. cistoides.] Scutruorr, P. N. Zytologische Untersuchungen in der Reihe der Geraniales. Jahrb. Wiss. Bot. 63: 707-759. 1924. [T. terrestris, 728. . Die Zytologie der Blutenpflanzen. Stuttgart. 1926. [T. terrestris, 585.] Scurwernexo, H. G. An account of the South African species of Tribulus . ex Linn. Bothalia 3: 159-178. 1937. [8 species recognized in cid study. | SovEces, R. Embryogenie des Zygophyllacées. Développement de l’embryon chez le Tribulus terrestris L. Compt. Rend. Acad. Sci. Paris 234: 1817- 1820. 1952. 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 547 Squires, V. R. A note on Aristotelia sp. (Lep., Gelechiidae) attacking Tribulus terrestris L. Jour. Entomol. Soc. Austral. (N. S. Wales) 2: 43, 44. 1965.* . Autecological studies on Tribulus terrestris L. (Zygophyllaceae). B. Litt. Thesis, Univ. New England, Armidale. —. Ecological factors contributing to the success of Tribulus terrestris L. as a weed in a winter rainfall environment in southern Australia. Proc. Ecol. Soc. Austral. 4: 55-66. 1969a. Distribution and polymorphism of Tribulus terrestris sens. lat. in Australia. Victorian Nat. 86: 328-334. 1969. SreEeNis, C. G. G. J. vAN. Contributions a l’étude de la flore des Indes Néer- landaises. 24. Zygophyllaceae. Bull. Jard. Bot. Buitenzorg. III. 13: 104, 105. 1933. [T. cistoides. | —. Pacific plant areas. vol. 1. iii + 297 pp. Manila. 1963. [Tribulus, 246.] TwISsELMANN, E. C. Plant records in Kern County, California. Leafl. West. Bot. 10: 46-64. 1963. [T. cistoides, 55; but see Howett above. | Verpcourt, B., & E. C. Trump. Common poisonous plants of East Africa. 254 pp. London. 1969. [T. terrestris, 31, 33, 34.] Wiurams, R. O. The useful and ornamental plants in Zanzibar and Pemba. ix + 497 pp. Zanzibar. 1949. [T. cistoides, 476; the plant can be trained as a climber. | Subfam. ZYGOPHYLLOIDEAE 3. Guaiacum Linnaeus, Sp. Pl. 1: 381. 1753, ‘Guajacum’; Gen. Fi, ed. 3: 179. 1754. Evergreen [deciduous] trees [or shrubs], [1-]2.5-10[-20] m. high, with a dense rounded crown; trunk short, stout, occasionally 1[-2] m. in diameter, the bark thin, light gray [light brown or olive-green], fissured vertically [smooth], separating on the surface into thin whitish [olive- green] scales, the wood hard and resinous, the heartwood dark green or yellow-brown, the sapwood thin and yellow; branches stout, slightly angled, pubescent, becoming glabrate, spreading, the branchlets stout, slightly angled, with conspicuously swollen nodes, green, becoming light gray, fissured, pubescent, becoming glabrate. Leaves opposite [sometimes crowded on short lateral branchlets], even-pinnate, the petiole and rachis pubescent [glabrous]; leaflets [1—]3—5[-7] pairs, opposite, entire, sub- sessile, [narrowly] elliptic to obliquely oblong or obovate [ovate], rounded [acute to obtuse or retuse] and mucronate apically, inequilateral basally, [membranaceous to] subcoriaceous or coriaceous, pubescent to glabrate, somewhat unequal in size, basal [apical] pair largest; stipules acuminate, usually mucronulate, usually deciduous, sometimes persistent. _ Flowers showy, solitary to several [or many], slightly irregular by twisting of petals; peduncles borne in axils of minute axillary bracts between stipules [sometimes crowded on short lateral branchlets], shorter than leaves. Sepals 4 or 5, free to slightly connate basally, obovate, pubescent, smaller than petals, deciduous. Petals 4 or 5, blue [violet or rarely white] drying yellow, free, obovate, rounded to lobed apically, clawed and twisted basally, spreading, glabrous or pubescent, imbricate. Disc annular, inconspicuous. 548 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Stamens 8-10, free; filaments subulate or slightly winged basally, un- appendaged, inserted on disc; anthers versatile, sagittate or cordate. Style simple, slender, subulate, persisting and forming a beak on fruit; stigma minute, entire or with as many lobes as ovary; ovary on a short gynophore, [2—] 4- or 5-lobed and -loculate, obovoid [obcordate and flattened], glab- rous [pubescent]; ovules 8-10 per locule, pendulous, the placentation axile. Fruit a capsule, [2-] 4- or 5-lobed and -winged, obovoid [obcor- date and flattened], broadest apically, smooth, glabrous [pubescent], coria- ceous, [green] orange to yellowish[-brown], narrowed basally into a short gynophore, septicidally dehiscent. Seeds ovoid, 1 per locule, [1 or 2] 4 or 5 maturing per fruit, surrounded by a thick fleshy red aril; seed coat membranaceous, brown or black; embryo straight, cotyledons foliaceous; endosperm present. LecToTyPE sPECIES: G. officinale L.; see Vail & Ryd- berg, N. Am. Fl. 25: 105. 1910. (Name from guayacan, a Caribbean Indian name for species of Guaiacum and Tabebuia Gomes ex DC. [Big- noniaceae| with extremely hard wood.) — LicNuM VITAE. A New World genus of four or five species occurring from Sonora, Mexico, south to Costa Rica and throughout the Caribbean basin north to the Florida Keys and the Bahamas. Guaiacum sanctum L., a low gnarled tree rarely up to 10 m. tall, reaches our area on the Florida Keys, where it was formerly of rather common occurrence. With the destruction of the tropical hardwood forests there, the only remaining trees of any size appear to be on Lignumvitae Key. It also occurs in the Bahamas, Cuba, Hispaniola, Puerto Rico, Yucatan, Guatemala, Honduras, Nicaragua, and Costa Rica. The related but quite distinct G. officinale L., 2n = ca. 26, is indigenous from the Bahamas and Greater Antilles, south through the Lesser Antilles to Tobago, and in Guyana, Venezuela, and Colombia Guaiacum Coulteri Gray occurs from northwestern Mexico to Central America, and the closely related G. unijugum T. S. Brandegee is endemic to the southern tip of Baja California. Both G. sanctum and G. officinale are cultivated in southern Florida (and elsewhere in tropical America) for their handsome blue flowers, colorful orange or yellow fruits, and attractive dark green persistent foliage. The violet-flowered G. C oulteri is grown to a much lesser exten The leaves of Galera sanctum have three or four pairs of leaflets, the lowermost and middle pairs being largest, and the fruits are pointed at the apex, are four- or five-lobed and -winged, and have four or five seeds. The leaves of G. officinale have two or three pairs of obovate leaflets, with the terminal pair the largest, and the fruits are emarginate or heart shaped at the apex and are two-lobed and -winged with one or two seeds. Both are evergreen plants with flowers 2.5 cm. or less in diameter that are produced as the new foliage is expanding. Guaiacum Coulteri has leaves with three to seven pairs of narrowly elliptic leaflets, the middle pairs being largest, and pointed four- or five-lobed and -winged fruits with four or five seeds. The flowers are larger than 2.5 cm. in diameter, and the petals are more dis- 1972] PORTER, GENERA OF ZYGOPHYLLACEAE . aae tinctly clawed than in the two preceding species. F spt usually begins while the plant is leafless before the new leaves unfold. Guaiacum sanctum and G. officinale are cui in the Greater Antil- les, and G. sanctum and G. Coulteri are sympatric in Central America. The name G. guatemalense Planchon ex Rydb. has been applied to collec- tions from Guatemala, Honduras, Nicaragua, and Costa Rica. However, the range in variation exhibited by these specimens, plus the observation that some have much-reduced seed-set, suggests that G. Coulteri and G. sanctum are hybridizing and introgressing in Central America. Guaiacum is closely related to Porlieria Ruiz & Pavon, which has about four species, one in Texas and northeastern Mexico and several in southern South America. Porlieria differs from Guaiacum in having persistent, commonly spinescent stipules; more, smaller, and narrower leaflets; stam- inal appendages; and two to four ovules per locule. Further study may prove them to be congeneric. The wood and resin of Guaiacum species, especially of G. officinale and G. sanctum, have long been the most important economic products of the family. Soon after Columbus’ voyage to the New World, large quantities of wood_were being shipped to Europe because of the supposed medicinal value of an extract of the heartwood. Sold by the pound, the wood was quite expensive, considering that its specific gravity is about 1.2. So many trees and populations have been destroyed that these plants have become extinct or very rare on many of the Caribbean islands. They are very slow growing and are not quick to recolonize areas from which they have been removed. The name lignum vitae (wood of life) derives from the belief that the wood exhibited miraculous medicinal properties. For over 200 years it was considered a specific remedy for venereal disease. With the invention of steam-powered ships, the wood became important for bearings or bushing blocks to line the stern tubes of propeller shafts. The wood is one of the strongest known, and its high resin content and concomitant self-lubricat- ing properties, combined with its strength, make it ideal for uses under conditions of stress and sea water. Today the wood is used mainly for the manufacture of such turned objects as mallets, pulley sheaves, caster wheels, bowling balls, stencil and chisel blocks, mortars and pestles, brush backs, planes, and the like. For good discussions of the sources of supply, marketing, and uses of lignum vitae, see Record (1921) and Longwood (1962) REFERENCES: Under — re ickeonyhs see AGABABIAN; AsPREY & THORINGTON; BARNER; Bartey & Tupper; BAILLON; BATE-SMITH; BRITTON & SHAFER; CHATTAWAY; Cook & iia Davis; Guilin & RaAvEN; ENGLER; ErptMAN; GOMEZ DE LA Maza; Gray; Hermscu; InmapAR; KALicIs-WALALANGI; LAssER; Mar- TINEZ; Metcatre & CHALK; NeaL; Perez ARBELAEZ; PITTIER; PorRTER, 1963, 1969; Recorp; Recorp & HEss; cae SAUNDERS; SMILEY; TORTORELLI; VaIL & RYDBERG; and WILLIAMs & Wits 550 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 BarrETT, M. F. Common exotic trees of _ Florida (Dicotyledons). xii + 414 pp. Gainesville. 1956. [Guaiacum, 8. BATERDEN, J. R. Timber. London. 1908. Sitka 130-132. | BayLey, I. The bush-teas of Barbados. Jour. Barbados Mus. Hist. Soc. 16: 103- 112. 1949. [G. officinale, 107.] Bere, O. C., & C. F. Scomipr. Atlas der officinellen Pflanzen. Vol. 3. ed. 2. iv + 102 pp. pls. 95-132. Leipzig. 1899. [Guaiacum, 8-10, Boutcer, G. S. Wood, a manual of the natural history and industrial applications of the timbers of commerce. 348 pp. London. 1908. [Guaiacum, 202.] Boyce, J. S. Forest Pathology. ed. 3. New York. 1961. [Guaiacum, 358.] Brusu, W. D. Lignumvitae. U.S. Dep. Agr. Forest Serv. 113 pp. 1938. Buswe.t, W. M. Native trees and palms of South Florida. Special Serv. Bull. Univ. Miami 19(6): 1-54. 1945. [G. sanctum, 21.] Corpus, V. Historiae plantarum libri. IV. lvs. 86-212 im C. GEsNer, ed., Annotationes in pedacii Dioseosides anazarbei de medica materica. Soar Dawson, E. Y. Some ethnobotanical notes on the Seri Indians. Desert Pl. Life 16: 133-138. 1944. [G. Coulteri, 136.] Duke, J. A. Keys for the identification of seedlings of some prominent woody species in eight forest types in Puerto Rico. Ann. Missouri Bot. Gard. 52: 314-350. 1965. [G. officinale, G. sanctum, 327, fig. 70 ———. On tropical tree seedlings. I. Seeds, seedlings, systems, and systematics. Ibid. 56: 125-161. 1969. [Guaiacum, 151.] Fatkx, E. Zur Kenntniss der Derivate des Guajacols. Ph.D. Thesis, Berlin. 1900. FERNANDEZ DE Oviepbo, G. Natural oe of the West Indies. (Transl. and ed. A. STOUDEMIRE) xvii+ 140 pp. Chapel Hill, North Carolina. 1959. [Translation of the Spanish wt Ea of 1526; Guaiacum, 88-90. FRANKLIN, G. L. A rapid method of softening “aud for microtome sectioning. Trop. Woods 88: 35, 36. 1946. [Lignum-vitae and other hard woods treated with glacial acetic acid and hydrogen peroxide. | GonzALez OrtEcA, J. Guayacan: Guaiacum Coulteri A. Gray y Guaiacum Palmeri Vail. Méx. Forestal 5: 139-141. 1927.* [Reviewed in Trop. Woo 14: 38, 39. 1928.] . Ebano, arellano o palo colorado, ampas, tecomate, ayele o guaje cirial, guayacan, haba. Bol. Pro-Cultura Reg. Mazatlan 1(33): 3-12. 1934.* Gray, A. Plantae Wrightianae texano-neo-mexicanae. Part I. Smithson. Contr. Knowl. 3(5): 1-146. 1852. [Guaiacum, 29.] Harris, D. R. Plants, animals, and man in the outer Leeward Islands, West Indies. An ecological study of Antigua, Barbuda, and Anguilla. Univ. Calif. Stud. Geogr. 18: 1-164. 1965. [Includes G. Oficina le.] Hess, R. W. Keys to American woods (continued). Trop. Woods 85: 11-19. 1946. [Guaiacum, 14.] Irvin, H. C. Marine bearings. Modern Plastics 19(8): 33, 34, 120. 1942.* [See Trop. Woods 70: 36. 1942.] EHIRA, R. Anatomical notes on Indian woods. Dep. Forestry Gov. Res. Inst. Bull. 4. 40 pp. 1 pl. Taihoku, Formosa. 1924. [Guaiacum, 34.] Kisser, J. Die Dampfmethode, ein neues Verfahren zum Schneiden hartester 1972] PORTER, GENERA OF ZYGOPHYLLACEAE 551 pflanzlicher Objekte. Zeitschr. Wiss. Mikroskop. Tech. 43: 346-354. 1926.* [See Trop. Woods 11: 27. 1927 Lirtte, E. L., Jr., & F. H. WapswortH. Common trees of Puerto Rico and the Virgin Islands. U.S. Dep. Agr. Forest Serv. Agr. Handb. 249. x + 548 pp. 1964. [G. officinale, G. sanctum, 212-215, figs. 93, 94.] Loncwoop, F. R. Present and potential commercial timbers of the Caribbean, with special reference to the West Indies, the Guianas, and British Hon- duras. U.S. Dep. Agr. Forest Serv. Handb. 207. iii + 167 pp. 1962. [Guaia- cum, 73-75; additional references. | Miranpa, F. La vegetacién de Chiapas. Vol. 1. Tuxtla Gutierrez, Mexico. 1952; Vol. 2. 1953. [G. Coulteri as G. sanctum, 1: 111; 2: 47, 48.] Monarpves, N. Joyfull newes out of the Newe Founde Worlde. (Transl. J. Frampton.) Vol. 1. xxvii + 177 pp. Vol. 2. 188 pp. London. 1925. [Originally published in 1568; Guaiacum, 1: 28-33. ] Nimz, H. Isolierung von Guaiacylglycerol-b-coniferylather aus Fichtenholz. Chem. Ber. 98: 533-537. 1965.* NuNez MEtenpez, E. Plantas medicinales de Puerto Rico. Univ. Puerto Rico Estac. Exper. Agr. Bol. 176: 1-245. 1964. [G. officinale, 96, 97.] Pertcuik, B., & H. Pertcuix. Flowering trees of the Caribbean. xii + 125 pp. New York. 1951. [G. officinale, 107-109. ] Recorp, S. J. Lignum-vitae, the vital wood. Sci. Am. Suppl. 20: 4. 5, 15. 1919. . Lignum-vitae blocks for stern bearings on steamships. Wood-worker p. 40. April, 1920.* . Lignum-vitae imports increase. Hardwood Rec. 50(9): 18, 19. 1921.* Rickert, H. W. The “original spelling” of botanical names. Taxon 4: 185-188. 1955. [Guaiacum, 187.] Roic y Mesa, J. T. Algunas propiedades medicinales de plantas sudamericanas que se conocen poco en la América del Sur. Lilloa 18: 191-201. 1949. [G. officinale, 193; diaphoretic. ] SarcENT, C. S. Guaiacum, Silva N, Am. 1: 59-64. pl. 28. 1891. SouTHWELL, C. R., & J. D. Butrman. Marine borer resistance of untreated woods over long periods of immersion in tropical waters. Biotropica 3: 81-107. 1971. [G. officinale, 104.] SpracuE, T. A. The correct spelling of certain generic names. Bull. Misc. Inf. Kew 1929: 38-52. 1929. [Guaiacum, 45, 46.] Storer, D. P. Familiar trees and cultivated plants of Jamaica. 81 pp. New York. 1958. [G. officinale, 29, 30.] Streets, R. J. Exotic forest trees in the British Commonwealth. Oxford. 1966. [Guaiacum, 403, 404.] Sturrock, D., & E. A. MENNINGER. Shade and ornamental trees for South Florida and Cuba. 172 pp. Stuart, Florida. 1946. [Guaiacum, 172; 1 Swasy, C. Some trees of Jamaica. Pp. 70-97 in C. B. Lewis, ed., Glimpses of Jamaican natural history. Vol. 2. Kingston. 1946. [G. officinale, 71, 72.] Tozzer, A. M., ed. Landa’s Relacién de las cosas de Yucatan. Pap. Peabody Mus. Am. Archeol. Ethnol. Harvard Univ. 18. xiii + 394 pp. 2 pls. 1941. [Guaiacum, 197, 198.] ne West, E., & L. E. ARNOLD. The native trees of Florida. xx + 212 pp. Gainesville, Florida. 1946. [G. sanctum, 94.] 552 JOURNAL OF THE ARNOLD ARBORETUM [von 53 Witson, E. O., & T. Etsner. Lignumvitae — relict island. Natural History 72(8): 52-57, Oct. 1968. [Estimates from tree-ring analysis place the age of the oldest individuals of G. sanctum on Lignumvitae Key at “over 1000 years.” Illustrations of fruit, on cover; of bark, p. 54; and of tree, flower, and fruit, p. 57.] Missourrt BoTANICAL GARDEN Sr. Louis, Missourtr Present address: DEPARTMENT OF BOTANY SMITHSONIAN INSTITUTION WasurncTon, D.C. 20560 1972] WEAVER, MACROCARPAEA IN COSTA RICA 553 TROP (WS \ new», fy ps » syn: g aisly On THE GENUS MACROCARPAEA (GENTIANACEAE) IN COSTA RICA = #4 90068 82 RicHArD E. WEAVER, JR.1 THE GENUS Macrocarpaea (Griseb.) Gilg consists of approximately 30 species of shrubby Gentianaceae with a distribution centered in the Andes of northern South America and extending south and east into the Amazon Basin and the mountains of southern Brazil, and north into the Greater Antilles and the mountains of Costa Rica. Three species are now known from Costa Rica. The most common of these, M. valerii Standl., is a familiar summer-blooming shrub of the Cordillera de Talamanca and the Cordillera Central. For nearly 70 years, the second species, M. subcaudata Ewan, was represented only by the type collection. A modern collection, from another locality, shows that this plant sometimes grows as an epiphyte. Recently, material of a third species has come to light which is described in this paper as Macrocarpaea acuminata. According to Ewan’s (1948) treatment of Macrocarpaea, the genus con- sists of two subgenera: PARANAGENES Ewan, with the herbaceous M. rubra Malme from southern Brazil as the sole species; and MACROCARPAEA (Eumacrocarpaea Ewan) with two sections based primarily on leaf char- acters, the typical one (Sect. Magnoliifoliae Ewan) and Sect. TABAct- oLIAE Ewan. Of the Costa Rican species, Macrocarpaea subcaudata belongs to the typical section while M. valerii belongs to Sect. TABACcT- FOLIAE. However, with its short internodes and relatively small but mem- branaceous leaves, M. acuminata appears to be intermediate between the sections. In many ways, Ewan’s infrageneric classification seems unnatural. Recent work by Nilsson (1968, 1970) has shown that three types of pollen are characteristic of the species included in Macrocarpaea by Ewan: a dis- tinctive one, having single grains with a coarse reticulum found in subgen. PARANAGENES as well as in the majority of the species in the typical sub- genus, including the three from Costa Rica; another single-grain type with verrucoid processes found in three species of sect. TABACIFOLIAE and char- acteristic of the monotypic Bolivian Rusbyanthus Gilg; and a final type, with pollen in tetrads, found in five species of the typical section and char- acteristic also of Chelonanthus (Griseb.) Gilg, another neotropical genus. Although those species with Rusbyanthus-type pollen appear to be refer- able to Macrocarpaea on morphological grounds, those with pollen in tetrads do not. These latter species differ from all the others in that their inflorescence consists of a single, terminal, compound dichasium, the divi- The author wishes to thank Dr. Robert L. Wilbur of Duke University for collect- ing and preserving flower buds of Macrocarpaea valerii, and the curators of the follow- ing herbaria for the loan of specimens: DUKE, F, GH, NY, US. 554 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 sions of which are not subtended by foliaceous bracts. The proper generic alignment of these species is unclear, although morphologically they appear most closely related to Chelonanthus. A chromosome number for Macrocarpaea valerii, 2n = 42, is reported here for the first time. The standard squash technique was used and the voucher specimen (Wilbur 14680) is deposited in the Duke University Herbarium. The only previous report of a chromosome number in the genus was for the Jamaican M. thamnoides (Griseb.) Gilg (Weaver, 1969), also 2n = 42 Key To THE Costa RIcAN SPECIES OF MACROCARPAEA A. Leaves subcoriaceous, only the midvein and the primary laterals evident; calyx lobes essentially alike and longer than the tube, minutely spiculate. . . M. subcaudata. A. Leaves membranaceous, the lesser veins evident, at least below; calyx lobes unequal, shorter than the tube or rarely equalling it, glabrous. B. Foliage leaves (those not subtending a flowering branch) 7 cm. broad or narrower, more than twice as long as broad; leaves glabrous or rarely the midvein very sparsely spiculate; corolla 4.4-4.7 cm. long; outer calyx lobes obtuse or acutish, the inner ones rounded, as long as broad or narrower; calyx less than 1/3 as long as the corolla tube; stems smooth; lateral flowering branches in 2 or rarely 3 pairs, unbranched or once branched below the dichasia; stalks of the lowermost flowering branches less than 10 cm. long. .................. 2. M. acuminata. Foliage leaves more than 9 cm. broad, less than twice as long as broad; foliage leaves and lower floral leaves spiculate on the veins beneath; corolla 3.0-4.0 cm. long; outer calyx lobes rounded, the inner ones truncate or broadly rounded and broader than long; calyx more than 1/3 as long as the corolla tube; stems papillose (under magnification) ; lateral flowering branches usually in more than 3 pairs, usually 2-3 times branched below the dichasia; stalks of the lowermost flowering Dranches more than 10 cm: bag = OP 1. M. valerii. w 1. Macrocarpaea valerii Standley, Publ. Field Mus. Bot. 18: 928. 1938. Type, Costa Rica: La Hondura de San José, Valerio 692 (F!). Sparsely branched shrubs or subshrubs to 4 m. tall. Stems to 3 cm. in diameter, the internodes hollow with age. Foliage leaves broad-elliptic to suborbicular and obovate, abruptly pape attenuate or cuneate- attenuate, 17.5-46 cm. long and 9.5-26 cm. broad, the venation very prominent, the 2 upper pairs of primary ‘cca veins arching into the tip of the leaf; petioles winged, 1.5-6.5 cm. long; leaves subtending the flower- ing branches smaller, the lowermost orbicular or broad-elliptic, the upper- most ovate and riakninerved: Inflorescences terminal and also axillary from the upper nodes, of long-stalked, branched systems of compound or ap- parently simple dichasia, Pedicels 5-20 mm. long, strongly recurved in fruit. Calyx campanulate, fleshy, 9-14 mm. long, the lobes hyaline-mar- gined and ciliolate, 3-6 mm. long and 3-7 mm. broad, strongly unequal, the 1972] WEAVER, MACROCARPAEA IN COSTA RICA 555 outer ones ovate, the inner ones transversely elliptic to broadly obovate and auriculate at the base. Corolla greenish-white or pale greenish-yellow, campanulate, somewhat fleshy, the lobes recurved, broadly triangular- ovate to nearly semicircular, obtuse to rounded and often notched, minutely ciliolate, 7-12 mm. long and 5-10 mm. broad, the tube 2.2—2.9 cm. long. Filaments 13-18 mm. long, surpassing the corolla tube or not; anthers 3-5 mm. long. Style 13-16 mm. long; stigma deeply bilobed, the lobes oblong-spathulate, ca. 2 mm. long. Capsules 15-21 mm. long, beaked. © CHROMOSOME NUMBER: 2n = 42. DiIsTRIBUTION: from 1000-1800 meters in the mountains of central Costa Rica (Map 1). REPRESENTATIVE COLLECTIONS Costa Rica. ALAJUELA: La Pena de Zarcero, Smith 1001 (¥, GH); ca. 19 km. n. of San Ramon, Wilbur & Stone 10684 (DUKE, US); ca. 4.5 mi. n. of Vara Blanca on Route 9 toward Puerto Viejo, Wilbur & Stone 10472 (DUKE, US). Cartaco: Tapanti, Jiménez M. 2023 (¥, GH, NY). HEREDIA: 1.5 mi. n. of Vara Blanca on road to San Miguel, Weaver 1405 (DUKE). SAN JosE: ca. 12 km. nne. of San Vicente de Moravia on Route 220 in the saddle between Irazi and Barba in the vicinity of La Palma, Wilbur 14680 (puKE); Cordillera de Talamanca, Cerro de la Muerte, 16 mi. (by road) n. of San Isidro General, Webster, Miller, & Miller 12329 (pUKE); 13 km. al nw. de San Isidro del General, Jiménez M. (F). As noted by Ewan (1948), this is the most common of the Costa Rican species of Macrocarpaea and is closely related to the Colombian M. macro- phylla (HBK.) Gilg. In fact, the two are virtually identical, except that the calyx of M. valerii is glabrous, while that of M. macrophylla is spicu- late. 2. Macrocarpaea acuminata Weaver, sp. nov. TYPE. Costa Rica: Cartago. Tapanti, Jiménez M. 2023. (holotype, F; isotypes, GH, NY). Frutex, ca. 3 m. altus. Folia membranacea anguste elliptica vel elliptico- oblanceolata, apice acuminata, basi attenuata, margine + revoluta, nervis -+ prominentibus, 7~20 cm. longa et 2.5—7 cm. lata, plus duplo longiora quam latiora, glabra vel raro costa sparsim spiculata; petioli ad 1 cm. longi. In- florescentiae terminales axillaresque, ramis floriferis lateralibus in 2 vel raro 3 paribus oppositis. Calyx 10-12 mm. longus, lobis inaequalibus 5-7 mm. longis et 4-6 mm. latis, longioribus quam latioribus, lobis exter- ioribus acutiusculis vel obtusis, lobis interioribus rotundatis. Corolla late infundibuliformis 4.4—4.7 cm. longa, tubo 3.2-3.6 cm. longo, calyce plus triplo longiore, lobis recurvatis triangulari-ovatis obtusis ciliolatis ca. 12 mm. longis et 10 mm. latis. Filamenta 20-25 mm. longa, inclusa vel leviter exserta; antherae ca. 6 mm. longae. Stylus 20-25 mm. longus; stigma 556 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 a M. acuminata A M. Subcaudata @ M. valerii Map. 1. Distribution of the Costa Rican species of Macrocarpaea. profunde bilobata, lobis ca. 3 mm. longis. Capsula 18-22 mm. longa apice rostrata. Macrocarpaea acuminata is known only from the type locality, where it grows in close proximity to M. valerii. Jiménez, who collected both species at Tapanti (M. acuminata, Jiménez 2023; M. valerii, Jiménez 2022), has noted on the label that what is here called M. acuminata may only represent older specimens of M. valerii. The two species are similar and are probably closely related, but they may easily be separated by the characters outlined in the key above. 3. Macrocarpaea subcaudata Ewan, Contr. U.S. Natl. Herb. 29: 224. 1948. Type. Costa Rica: [San José.] La Palma, Wercklé 16492 (holo- type, Us!; isotype, Ny! ). Low shrubs, sometimes epiphytic. Foliage leaves glabrous, narrow-ellip- tic to oblanceolate, cuspidate-acuminate, long-attenuate, the margins some- what revolute, 5.4-11.3 cm. long and 1.2-3.0 cm. broad; petioles to 1.5 cm. long. Inflorescences terminal or also axillary from the uppermost node, few-flowered, of long stalked groups of apparently simple dichasia (the dichasia sometimes reduced to single flowers). Pedicels to 7 mm. long or flowers subsessile. Calyx campanulate, minutely spiculate (rather dense in the type collection) or rarely glabrescent, 7-10 mm. long, fused for 1/4- 1972] WEAVER, MACROCARPAEA IN COSTA RICA 557 1/3 its length; calyx lobes essentially alike, ovate to oblong, obtuse or rounded, ciliolate, longer than broad, 5—7 mm. long and 3-4 mm. broad. Corolla cream-colored or somewhat greenish, narrowly campanulate, 3.2- 3.4 cm. long, the tube 2.4-2.6 cm. long, the lobes somewhat recurved, triangular-ovate, rounded to obtuse, minutely ciliolate, 8-10 mm. long and 6-8 mm. broad. Stamens inserted at about the middle of the corolla tube, the filaments 17-20 mm. long, slightly surpassing the corolla tube; an- thers 3-4 mm. long. Style 15-18 mm. long; stigma deeply bilobed, the lobes oblong-elliptic, ca. 3 mm. long. Capsule unknown. DIsTRIBUTION: at ca. 1500 meters in the mountains east of San José, Costa Rica (Map 1) SPECIMENS EXAMINED Costa Rica. Cartaco: Rio Grande de Orosi, 8 km. s. of Tapanti, Lent 934 (F, NY). Two species of Macrocarpaea are now known to be at least partially epiphytic, M. subcaudata and M. browallioides (Ewan) Robyns & Nils- son from the mountains of central Panama. Unlike the recent collection by Lent, the type collection of M. subcaudata bears no notation as to whether or not it was growing epiphytically. Ewan (1948) considered the closest relatives of Macrocarpaea sub- caudata to be M. cerronis Ewan and M. salicifolia from the Guiana High- lands. These species, however, are among those with pollen in tetrads, and, as mentioned earlier, should be excluded from Macrocarpaea. Rather, M. subcaudata is probably related to M. acuminata and to certain of the West Indian species such as M. pinetorum Alain and M. domingensis Urban & Ekman. LITERATURE CITED Ewan, J. A revision of Macrocarpaea, a neotropical genus of shrubby gentians. Contr. U.S. Natl. Herb. 29: 209-249. 1948. Nitsson, S$. Pollen morphology in the genus Macrocarpaea (Gentianaceae) and its taxonomical significance. Sv. Bot. Tidskr. 62: 338-364. 1968. . Pollen morphological contributions to the taxonomy of Lisianthus L. s. lat. (Gentianaceae). Sv. Bot. Tidskr. 64: 1-43. Weaver, R. E., Jr. Cytotaxonomic notes on some neotropical Gentianaceae. Ann. Missouri Bot. Gard. 56: 439-443. 1969. ARNOLD ARBORETUM HARVARD UNIVERSITY JAMAICA P LAIN MASSACHUSETTS 02130 558 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 CYTOLOGICAL STUDIES ON HIMALAYAN MELIACEAE P. N. MeuRA, T. S. SAREEN, AND P. K. KHOSLA THIS Woopy FAMILY is comprised of 52 genera and at least 1,000 species (Styles & Vosa, 1971) distributed throughout the tropics and subtropics of the world. Its importance to forestry rests on several valuable timbers such as the mahogany of the trade (Swietenia mahagoni and S. macro- phylla), mahogany-like African woods, especially of Khaya and Entan- drophragma, and toon of India. Meliaceous woods are prized for furniture and cabinet work. The family is represented in India by 50 species be- longing to 18 genera of which 13 species are of commercial importance. Chukrasia tabularis, Toona ciliata, T. serrata, Melia azedarach, M. com- posita, Aphanamixis polystachya and Dysoxylum binectariferum are some of the important timber species studied presently. So far about 90 species belonging to 37 genera of Meliaceae have been reported cytologically, the maximum contribution being by Styles and Vosa (I. c.). These authors made mitotic observations with the exception of a single meiotic count. The present work is an attempt to fill lacunae in the cytology of the Himalayan members, chromosome counts for most of which have already appeared in IOPB reports (Mehra & Khosla, 1969; Mehra & Sareen, 1969). Our aim has been not only to know their chromo- some counts but especially the process of meiosis, flowering season, and fruiting season, data which are prerequisite for any rational program for the improvement of trees. MATERIAL AND METHODS Material was collected from the wild in the forests of the Himalayas and the Khasia and Jaintia hills. Flower buds were fixed in Carnoy’s fluid. Squashing of anthers was accomplished in aceto-carmine. For mitotic studies, root-tips were pretreated with .003 M solution of 8-hy- droxiquinoline for four hours and squashed in aceto-lacmoid. Figures are at a uniform magnification of 1360. Voucher specimens have been de- posited in the Herbarium, Panjab University, Botany Department, Chandi- garh-14 (India). RESULTS AND DISCUSSION Chromosome numbers of 24 taxa comprising 18 species belonging to 10 genera are listed in TABLE 1, which also includes data on specific localities of collection, previous counts for the species, and flowering and fruiting sea- sons. The course of meiosis has been found to be normal in all the taxa. Ficures 1 and 2, habit Sston aphs. individual tree with straight bole. 1, Chukrasia tabularis, a tall tree with straight bole. 2, Dysoxylum binectariferum, an [2261 AVAOVITAW NVAVIVWIH “IV Ld VYHAW TABLE 1. List of species studied, supplementary information, and chromosome numbers obtained CHROMO- FLOWERING AND SOME Fic. Previous NAME Source FRUITING PERIOD * NUMBER No. REPORTS CHUKRASIA TABULARIS A. Juss. Darjeeling, Manjitar 5-6 9-12 n= 13 3 n = 13: Rao, 300 m. 1967 Toona crirata M. J. Roem. Darjeeling, Teesta 3-5 5-8 n= 28 + = 28: Singh, 300 m. 1951: 2n = 56: Styles & Vosa, 1971 Simla, Basantpur n = 26 5 900 m. T. CILIATA var, PILLISTAMINEA C. DC. Shillong, Barapani 1-2 3-4 n = 39 6 800 m. T. serrata (Royle) M. J. Roem. Mussoorie, 1950 m. 5-6 9-10 m= $2 7 T. microcarra (C. DC.) Harms Darjeeling, Mahanadi 11-12 2-3 #= 12 8 1500 m. APH ey a go iam Darjeeling, Sukna 9-10 2-3 "m= 18 9 2a = 72: (Wall.) P 150 m. Minfray, 1963; Styles & Vosa, 1971 MELIA AZEDARACH L. Darjeeling, Manjitar 3-4 11-2 n= 14 10 2n = 28: Bowden, 300 m. 1945; Pathak & Singh, 1949; Minfray, 1963; Nainital, 1950 m. n= 14 Gadella et al., 1966; Styles & Vosa, 1971 M. composita Willd. Khasia & Jaintia 3-4 winter “= 14 11 hills, Garampani, 800 m. M. TOOSENDAN Sieb. & Zucc. AZADIRACHTA INDICA A, Juss. DysoxyLUM PALLENS Hiern D. BrnectarireruM Hook. f. D. HAMILTontI Hiern D. procerum Hiern AGLAIA PeRviripis Hiern A. EDULIs King AmoorRA WALLICHII King HEYNEA TRIJUGA Roxs. CHISOCHETON PANICULATUS Hiern Khasia & Jaintia hills, Garampani 800 m. Darjeeling, Teesta 300 m. Haldwani, 400 m. Darjeeling, Sukna 150 m. Darjeeling, Sukna 150 m. Darjeeling, Manjitar 300 m. Darjeeling, Lebong 1800 m. Darjeeling, Teesta 300 m. Darjeeling, Mahanadi 1500 m. Darjeeling, Teesta 300 m. Darjeeling, Sukna 150 m.; Manjitar, 300 m. Nainital, Dogaon 1000 m. Darjeeling, Teesta 300 m. 4-8 3 onwards 1-2 8 onwards winter n= 14 n= 14 m= 14 n= 10 n = 40 n= 40 n= 40 n= 20 n= 40 n= 20 n= 14 n= 14 n= 23 2n = 28: Minfray, 1963 2n = 28: Pathak & Singh, 1949 n = 14: Mukherjee, 2n = 30: S. & G. Mangenot, 1958; Styles & Vosa, 1971 2n = 24: Nanda, 1962; m = 14: Rao, 1967 * Numbers in this column represent months, e.g., 1: January; 2: February; etc. 562 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Chukrasia A. Juss. includes one or two species distributed from South China to Indo-Malaya. The Indian species C. tabularis (FicuRE 1) is a much appreciated timber tree of the Sikkim Himalayas, Assam, the Anda- mans, Kanara hills of Deccan, Mysore and the western Peninsula. Dark pink blaze and a pink terminal pair of leaflets are diagnostic features for identification in the field. The chromosome count of m = 13 confirms Rao’s (1967) identical report for the species. Toona M. J. Roem. is a genus well known for its high grade timbers. T. ciliata and T. serrata are commercial timbers of India. The former is a morphologically variable tall tree of the tropical Himalayas, also found in central and southern India, which is confined mostly to mixed plain and sal forests. The species shows chromosomal polymorphism and exists in three cytotypes. The haploid count of = 28 in an east Himalayan popu- lation agrees with similar findings of Singh (1951) and Styles and Vosa (1. c.). The chromosome numbers of = 26 for a population from the western Himalayas and n = 39 for var. pillistaminea from the Khasia and Jaintia hills fall in the series of x = 13. T. serrata, a western Himalayan species with 2m = 52, is again based on x = 13. T. microcarpa occurs in the mid- dle hill forests of the Darjeeling hills. It is cytologically interesting in the sense that it adds a new chromosome count, » = 12, to the genus. Aphanamixis Blume embraces 25 species, mostly confined to Malaya. The Indian representative, A. polystachya, is a medium-sized tree and a commercial timber of the Darjeeling hills. The present taxon having = 18 is diploid, while Minfray (1963) and Styles and Vosa (I. c.) obtained tetraploid counts in this species where 2n = 72. Melia L. comprises 15 species. M. azedarach and M. composita, two Indian timbers, and M. toosendan uniformly reveal n = 14. Azadirachta A. Juss. is a small genus of two species. A. integrifolia is found in the Philippines, while A. indica is widely distributed in the Indo- Malayan region and displays two chromosome numbers, » = 14, and m= 15 (see Tape 1); the latter number might have arisen from the former by addition of one chromosome. Dysoxylum Blume is another Indo-Malayan genus of 200 species of which seven occur in India. Four species which have been studied here are common in the sal and mixed plain forests of the Darjeeling Himalayas. Of these D. binectariferum yields timber of economic importance. The counts of = 10 for D. pallens and n = 40 for D. binectariferum, D. hamiltonii, and D. procerum are the first cytological reports for these species. Other observations on the genus are D. pachyphyllum (2n = 82: Styles & Vosa, 1. c.), D. ramiflorum (n = 36-38: Paetow, 1931) and D. spectabilis (n = 42: Beuzenberg & Hair, 1963). Ten is the lowest gametic number recorded in the genus, and it is suggested as the primary base number. 1972 | MEHRA ET AL., HIMALAYAN MELIACEAE 563 Aglaia Lour. is unimportant from the point of view of forestry. This genus is studied cytologically for the first time here, showing ” = 20 in A. ayer Pad * ia ee e s dey 3-9, Chukrasia, Toona, and Aphanamixis. 3, wesiec a par M-I, 13. 4 and 5, Toona ciliata. 4, diakinesis, ” = 28; Tie i, 6. 6, i. ciliata, var. pillistaminea, diakinesis, » = 39. 7, T. serrat raltgllc el phase, 2% = 52. 8, T. microcarpa, akinesia, R= 1. 3, Tshaneeaieis poly- 18. stachya, M-I, ” = 564 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 yy, e ; FIGURES oly Melia, Azadirachta, and — 10, Melia azedarach, Seep m= 14. 11, M. composita, M-I, n = 14. 12, M. toosendan, diakine- sis, % = = 14, — nee rtastc indica, diakinesis, n= 14. 14, Dysoxylum pallens, diakiness m=10; 15, D. binectariferum, M-I, = 40; 16, D. hamiltoniit = 40. 1972] MEHRA ET AL., HIMALAYAN MELIACEAE f * ¢ Ficures 17-22. Dysoxylum, pons Amoora, Heynea, and Chisocheton. 17, Dysoxylum proce rum, M-I, n = 40. 18, Aglaia pervs, M-I, # = 20; 19, A. edulis, M-I, 40. 20, ora wallichii, M-I, fa 21, Heynea trijuga, M-I, n = Am 14. "02, Disscliton paniculatus, M-I, n= 566 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 perviridis and n = 40 in A. edulis. Twenty is suggested as its primary base number. Amoora Roxb. includes 25 species. A. wallichii, a commercial timber, is distributed in India in the moist evergreen forests of the eastern zone and in Andaman. A pink blaze with a few beads of milky juice is a useful field character for its identification. The only species of the genus known cytologically, it shows m = 20. It is presumed to be based on x = Heynea Roxb. is comprised of four species only. H. trijuga has been consistently observed by us, as well as by Rao (1. c.), to possess m = 14 in east and west Himalayan populations. An earlier report of 2m = 24 by Nanda (1962) appears to be erroneous. Chisocheton Blume is a genus with 100 species. The single Indian representative, C. paniculatus, is a small tree of the mixed plain forests, in Terai, Duars, and the lower hills of Bengal and Assam. This species pos- sesses 2 = 23, which is the only chromosome count for the genus. Twenty- three is the highest base number recorded in the present study. Unlike most of the other tropical hardwood families, the Meliaceae pre- sent a great heterogeneity of chromosome numbers with 2n — 16, 20, 22, 24, 26, 28, 30, 32, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 72, 78, 80, 84, 92, 100, ca. 280, and ca. 360 (cf. Styles & Vosa, |. c.). Evidently, the family has passed through an active phase of numerical chromosomal evolu- tion leading to the delimitation of genera and species. Most of the species possess high chromosome numbers and are in all probability palaeopoly- Ploids that have evolved from ancient extinct ancestors with low base numbers. The original base number of angiosperms is generally believed to lie between 6 and 8, 7 being the most widely presumed number for them, as well as for some primitive tropical woody families like Annonaceae (Raven & Kyhos, 1965) and Proteaceae (Johnson & Briggs, 1963). Ap- proximately 33 percent of the genera of Meliaceae investigated so far fall in the series x = 14, although the lowest gametic number is 8, reported only in one instance in Sandoricum koet jape. The other commonly repre- sented series seems to be x = 18. Both these series are high and could probably have arisen from 7 and 9 respectively. The rest of the numbers encountered in the family can be deduced from these by multiplication, stepwise reduction or addition, or by amphiploidy (Ficure 23). SUMMARY Cytological studies of 18 species belonging to ten genera from the Himalayas have been carried out. Of these, Chukrasia tabularis (n = 13), Toona ciliata (n = 26, 28, 39), T. serrata (2n = 52), Melia azedarach (n = 14), M. composita (n = 14), Aphanamixis polystachya (n = 18), and Dysoxylum binectariferum (m = 40) are some of the important timber species studied presently. The lowest gametic number in the present studies is reported in Dysoxylum pallens (n = 10). Normal meiotic be- MEHRA ET AL., HIMALAYAN MELIACEAE 567 1972] y Meli- havior is characteristic of all presently studied species. The famil aceae is known for the extreme polymorphism of chromosome num most of these are palaeopolyploids. bers and ACKNOWLEDGMENTS We are sincerely thankful to the sponsoring authorities of the U.S. government for financial help under PL—480, grant A7—FS-12 to the senior author. Ul qU2S2id U2|2q DADDY 0} pawnsaid‘ (6) wnisub 4 | sixiwpubydy Dapjd2q D2IDNS D1iAsD I apojD5 | Dwbosydospudzugy BI + 61 Dij!yr!43,Opidaq wnd!JopuDs 8 snyyuDiuoIndD> (2) DIL2W DIYDDIIpozy DINSIDM DauAay SIY>DYJOI23q Dssappdi> pUuoO]L Dj24p2> DIUDzIMS Dpiwios sndaps0)4x Dj24p25, DOAO> } DISDIXNYD D|2Ipa20pna2sq , v *S40}S29uUD Ajipa ID21}2yOdAY 21D Sasayquaiod uyzM UAB SJ2QWON :€z°9}4 DIOOWY Dipiby DI2Ip25 wnjAxoshq DIUZ}2IMS wWho1JopuDS DIUDWAN 6c L¢ Se e¢ DIUZIMS = =—»-»IFIYDIT].- UOYDOsSIYD Ddpip> D2DIIN]1 Dildos DIUOIUNYW DUIDSJaqopnasg DI24p2y—s«éDIBiaqany panBaqoan DIWDB2JDN BoAo7 DUOO] pAbyy DADYy 568 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 LITERATURE CITED BEUZENBERG, E. J., & J. B. Harr. 1963. Contributions to a chromosome atlas of New Zealand Flora—5. Miscellaneous families. New Zealand Jour. Bot. 1: 53-67. Bowpen, W. M. 1945. A list of chromosome numbers in higher plants. 1. Acan- thaceae to Myrtaceae. Am. Jour. Bot. 32: 82-92. DresHMUKH, N. Y. 1959. Chromosome number in Neem. Indian Oil Seeds Jour. (1): 45. GaveELLa, T. W. J., E. Kurpnuts, & E. A. MENNEGA. 1966. Chromosome num- bers of some flowering plants of Spain and S. France. Acta Bot. Neerl. 15: 9. 484-4 Jounson, L. A. S., & E. B. G. Briccs. 1963. Evolution in the Proteaceae. Austral. Jour. Bot. 11: 21-61. Mancenor, S., & G. MANGENOT. 1958. Deuxiéme liste de nombres chromo- somiques nouveaux chez diverses Dicotylédones et Monocotylédones d’Afrique occidentale. Bull. Jard. Bot. Etat, Bruxelles 28(4): 315-329. Menrpa, P.N., & P. K. Knosta. 1969. IOPB Chromosome Number Reports. Taxon 18: 217. & T. S. SarEEN. 1969. IOPB Chromosome Number Reports. Taxon 18: 440. Minrray, E. 1963. Contribution 4 étude caryo-taxinomique des Méliacées. Bull. Soc. Bot. France 110: 180-192. : MUKHERJEE, S. K. 1952. Meiosis in Azadirachta indica A. Juss. Curr. Sci. Bangalore 21: 287. Nanpa, P. C. 1962. Chromosome numbers of some trees and shrubs. Jour. Ind. Bot. Soc. 41: 271-277. Partow, W. 1931. Embryologische Untersuchungen an Taccaceen, Meliaceen, und Dilleniaceen. Planta 14: 441-470, PatHak, G. N., & B. SincH. 1949. Chromosome numbers in some angiosper- mous plants. Curr. Sci. Bangalore 18: 347. Rao, H. S. 1967. Chromosome counts of New Forest plants. Indian Forester 93: 243-254. Raven, P. H., & D. W. KyHos. 1965. New evidence concerning the original basic chromosome number of angiosperms. Evolution 19: 244-248. SINGH, B. 1951. Chromosome numbers in some flowering plants. Curr. Sci. Bangalore 20: 105. StyLEs, B. T., & C. G. Vosa. 1971. Chromosome numbers in the Meliaceae. Taxon 20: 485-492. ZERPA, D. M. ve. 1953. Los cromosomas de Melia azedarach. Agron. Trop. (Macaray) 2(4): 257. DEPARTMENT OF BoTANY ¢ PANJAB UNIVERSITY ¥ CHANDIGARH-14 (INpIA) ‘ 1972] INDEX 569 INDEX Acalypha deppeana, 321 —schiedeana, 321 Acraea, 515 — parvifolia, 515 Acrolophia, 520 iadencilistanpiinns 79, 81 ula, Aegiphila costaricensis, 320 Aeranthus porrectus, 522 20 (Agavaceae) II. Vegetative spec f Cordyline termin rpholo ete Studies in Cordyline, 1. Ageratum conyzoides, 321 Aglaia, 5 — edulis, 561 — perviridis, 561 Aizoaceae, 56 Alfaroa, 27, 28 ipsa, 515 Alnus nepalensis, 221 Alum-root, 4 bi Amoora, 566 — Walli Amphielottium, 516 — i DERSON, Loran C. Studies Bigelo owia \tiebdaaans IL. Xylary pesca ised Woodiness, and Paedo- morphos Nevethoen ee 320 Angraecum Lindenii, 526 ides Aphanomixis coe ao 558, 560 —— lemon 320 spatsiidese 202, 205, 212 Aretiys,> t fg hor a, Sou AxpeDiadase? — spicatm ) Artanthe seis ‘a, 3 Artocarpus, 2 —chaplasha 7, 218, 223 — gomezianus, 217, 218, 223 — heterophyll is, 217 Astereae, 499 sabe, 412, geal t. Asti mary shnpeiseeBae/3 24 e Genus Piper (Piper- w Guinea, Solomon _Is- acea w lands, and, Australian clis of Erythrina, II, . Notes on Asiatic-Polynesian-, 128 Astrocaryum mexicanum, 320 Auliza, 516 Aulizeum, 516 Azadirachta, 562 Balanitaceae, 532 ae et 516, 517 Beadlea Bean- caper Mey 531 450 Cara Smee 183 Bigelowia (Asteraceae), II. Xylary Comparisons, Woodiness, and co ie ig Studies on, 499 Bigelowia, 4 — nudata, es, <0, 510, 511, 513 —— su australis, — nuttallii, 500, 502, 503, a. 508 Bitternut hickory, 3 Black walnut, 47 —capitata, 390 Broughtonia, Broussonetia papyrifera, 216-218 Bulbophyllaria, 51 570 JOURNAL OF THE ARNOLD ARBORETUM Bulbophyllum, 213, 516 — . m, 5 516 rhea, big 105 Burlingtonia, 5 — candi ome Burr-nut, 543 Byrsonima, 104, 105, 109, 111 —subser. Glossolepis, 111 Calanthe, 517 Cala aay raakgrg KoA | ay ray — frigidus Calopogon, Me 523 Casta tice glauca, —sect. Apocarya, 34-36, 40, 41 — sect. Carya, 35, 36, 38, 40, 41 —sect. Rhamphocarya, 36 Caryocaraceae, 103 Casearia arguta, 320 Castilloa elastica, 216 — 517 — setaceum, 390 Conkieanences 183 Cestrum, 321 Chaetogastra chironioides, 400 — strigosa ray 320 a, 321 — seneitiote Charianthus (Melastomataceae) in the es, Notes on Tibouchina and, 3 Chartenthns, 399, 401 [voL. 53 Charianthus Fagg 401 — coccineus Chavica ambeinenss 3 — — latifolia, 3 = enti 7 — Imp — majuscula, 6, 12 — miniata, 10 a i, 6 Chekac ais. ie 81, 554 Chenopodiaceae, 56 CHEW, Wee-Loe. The Genus Piper (Plnéracese) in New Guinea, Solomon Chlorophora tinctoria, 216 Chrysosplenium, 413, 453-457 —sect, Alternifolia, 456 h ea | a (Verbenaceae), Dioecism n, 386 Cithareryumy, 3 86 — fruticosum, 386, 389 — staestkeet 320 Cleistes, 518 Clematis dioica, 321 Clibadium arboreum, 321 Climbing ae 497 Cochleanthes, 518 Coelia, 518 Comparettia, 518 — falcata, 518 Compositae, 49, 56 Corallodendrum divaricatum, 133 Cordyline ee II. Vegetative Morphology of a ot terminalis, Morphological Studies in Cordyline, 113, 115, 124 — stricta, 120 — terminalis, 113, 115, 120, 126 , 518 — corymbosa, 518 1972] Costa Rica, The cue Macrocarpaea ptpagearssoe ey in, 5 Costus spicatu Coutoubea Es 95 ill, 191 , 518 Croton macrodontus, 321 Cruciferae, 52, 56 cop eam 518 — lunata, 518 Cryptopiiorantins, 518 Cubeba canina, Cupania macrophylla, 320 68 Cycloptera, 519 Cymbidium Andersonii, 519 — luteum, 517 — triquetrum, 525 ; 530 Cytological Studies on Himalayan Meliaceae, 558 Cytomorphology of Arborescent Moraceae, 216 Decumaria, 413, 495-498 Dendrobium squalens, 530 Dendropanax arboreus, 320 Dendrophylax, 519, 526 ia, 485 — Baueri, 519 Diefenbach oerstedii, 321 0 Dienia agrees — congesta, 519 Dilomilis, 519, 525 INDEX Dinema, 519 ioecism " a ems (Verben- —procerum, 561 — pininana, 512 , THomas S. The Genera of ‘Tosseae in the Southeastern ey 519 Elm pat PB 517, 520 — calcarata, 390 Elvira biflora, 321 en kacad, 558 Enterolobium cyclocarpum exes | Ephedra, aporres Patterns in the Leaves of, 3 Ephedra, eeu, 373 — altissima, 364-370 — chilensis, 365-367, 370, 371, 374 — equisetiformis, 368 — foliata, 366-368 — fragilis var. campylopoda, 364, 374, 7 375 — gracilis, 366, 367 Epicladium, 520 Bpidendroideae 202, 205, 207, 208, 211- Bpidendrum, 516, 519, 520 ere 394, 395, 530 peony ae 394 — coc - cochleatum, oe Epos inoneagae 398 94 571 572 Epidendrum dendrobii, 526 —echinocarpum, 397, 398, 519 398 —tnpterum, 5 —uniflorum, 522 — vanilla, 5 1) — viridipurpureum, 394 Epistephium, 520 — elatum, 520 Epithecia, 520 Erodium, 183, ag “ig 193, 194, 197, 199 arbad 98 Erythrina, II, Notes on Asiatic-Polyne- sian-Australian species of, 128 Erythrina, 128, 136 — alba, 133 cies ciety, 129-131 — boni 6 — carnea, a — corallodendrum, 133 —— orientalis, 133 — crassifolia, 136 — divaricata, 133 Dereon geey 4 ig _ 137, 138 133 stalarle 128, 129, 138 JOURNAL OF THE ARNOLD ARBORETUM [von, 53 Erythrina livingstonia, 139 —merrilliana, 128, 129, 138, 139 — microcarpa, 128, 129, 132 1 — orientalis, 133 — parcellii, 133 a ese 133 —picta, 133 BE. no 128,129, 132 —spathacea, 133 — stipitata, 131 — stricta, 129, 132, va —suberosa, 129, 131, 132 — subumbrans, 129, ‘80 136, 137 —tahitensis, 130, 137, psx 128, 129, a 134-138 entalis Erythrodes, =i 520 — latifolia, 5 Seeks, 102 Eugenia, Floral Anatomy of Myrtaceae, Eu ugenia acapulcensis, 336, 358 58 1972] Eupatorium, 321 513 —leptocephala, 502, 503, 513 Fagaceae, 48 Fagales, 2 False-goatsbeard, 423 alse ogelehiie 431 FAWCET Fong 90 and P. B. em . Dioecism in Citharexylum ( Verbenacae), 386 Feather-Trees, Ficus, — altissima, 217 — benjamina var. comosa, 217, 223 22 — mysorensis, 217 —ne nme var, trilepis, 217, 223 — religi — roxbutht 217, 223 Filaree, 19 FIsHER, J. ‘" a . B. ToMLINson. Morphological Studies in Cordyline (Agavaceae) II. Vegetative Mor- phology of Cordyline terminalis, 113 Floral Anatomy of Myrtaceae, IT. S. Venation Patterns in the Leaves of Ephedra, 364 Fractiunguis, 521 Fuertesiella, 521 — pterichoides, 521 Galeandra, eg 521 — Baueri , pel Galphimia, 105 | ARAY, Lestig A. On the Origin of the Orchidaceae, t, 202 Garay, Leste A., and H R. Pec Notes on West Indian Orchids, Genera of Geraniaceae in the South- eastern United States, The, 182 Genera of Juglandaceae in the South- eastern United States, The, 26 INDEX 573 Genera of Saxifragaceae in me South- eastern United States, The, 4 Genera of Zygophyllaceae in is South- eastern United States, The, 531 » 93 — tribe Gentianess subtribe Tachiinae, 78 Gentianeae-Tachiinae, 78 Gentianella, 93 Genus ek (Gentianaceae), A Re- of — Neotropical, 76-100, 234— 272, enact Cite ee. (Gentianaceae) in Costa Rica, The, 553 Genus Piper (Piperaceae) in New Guinea, Solomon Islands, and Aus- tralia, 1 George B. Hinton, eet oa of Plants in outhwestern Mexi Geraniaceae, 102, sete te, 189, 532 183 Geranium, 183-185, 189-191, 193, 194, 200 —sect. Columbina, 192 —sect. Robertiana, 193 — sect. Sylvatica, we Geranium Family, Ghiesbreghtia qanade 521 Globulariaceae, 226 Gnetum, 375 — gnemon, 3 Golden saxifrage, 455 G6meEz-Pompa, ARTUR and SERGIO Grossulariaceae, 472 Guaiacum, 532, 539, 547-549 Guarea bijuga, 320 Guevara S., Sercio, and Arturo GOMEz- Pompa. Seeds from surface soils in a Tropical Region of Veracruz, Mexico, Gyrostachys, 521 Habenaria, 521 Haemodorales, 202 alenia, 93 574 Halictus purus, ois Hamelia axillaris, 3 Hans, A. S. Cs taulectionsis of Arbo- rescent Moraceae, 216 e, 78 Hass appendiculatus, 321 Helleborine purpurea ro ie 392 Hephestionia iggaaes Heron’s-bill, Heteropteris, Se: 104, 105 Heterotaxis, 522 — crassifolia, 522 Heuchera, 413, 435-439 ct. Heuchera, 436 Heruchea subsect. Pubescentes, — sect. Heucherella, 436 — sect. Holochloa, 4 — sect. Rhodoheuchera, 436 — subsect. Micranthae, 437 —subsect. Villosae, 437 Hibiscus lasiocarpus, 499 Hickory, 34 Himalayan ae Sas Cytological — on, ae Ss, and J. Rzepowsxk1, “esd BE. Hinton, Collector of Plants JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 idium, 522 Tlex serene 320 Inga spuri faetitatione: which duplicates of G. B. Hinton’s collections were distributed, 160 Ionops = es, 518, 522 — sect. Sempervirentes, 478 Jacobinia, 321 m Juglandaceae in the Southeastern United ates, The Genera of, 26 Juglandaceae, 26-29, 32 — sect. Trachycaryon, 47 Justicia, 320 Kallstroemia, 532, 539, 540 Khaya, 558 Kuosta, P. K., P. N. Meura, and T. S. wae, Cytological Studies on Hima- n Meliaceae, 558 Koellensteini a, 522 KRrUKOFF, B. on Asiatic-Poly- nesian- Australian Stcies of Erythrina, Labiatae, 49 Laelia, 515, 522 Laeliopsis, 522 Lagenanthus, 80 Ledocarpaceae, 183 Leianthus, 77, 234 —sect. Adenolisianthus, 78 — sect. Irlbachia, 78 — sect. Lisianthius, 77 — sect. Omphalostigma, 77.235 axill. — cordifolius, 293 — cuspidatus, 242 — latifolius, 279 1972] Leianthus longifolius, 288 —y cordifolius, 2 — seemanii, 284 — silenifolius, 250 — skinneri, 261 Lehmaniella, 79, 80 Lepanthopsis, 523 Leptorrhizae, 131 Lepuropetalon, 413, 458-460, 463 oT — nerv Esatawhin “(Gentianaces), A Revision of th — Genus, 76-100, 234-272, 2 tag Lisianthius, ees, 85, 86, 90-100, 234, 3, 290 — sect. Lisianthius, 235, 240 —sect. Omphalostigma, 82, 83, 85, 86, 91-93, 100, 235, 236 —subsect. Fruticosi, 82, 83, 85, 86, 88, 89, 99, 100, 241, 258, 283 — subsect. Herbacei, 81-83, 85, 86, 92, 96, 99, 100, 240, 241, 276 — acuminatus, 260, 263, 269, 270, 302 — adamsii, 92, 95, 259, 260, 267, 269 — alatus, 77 — auratus, 93, 95, 260, 270, 272 — axillaris, 93, 95, 259, 282-28 ol brevidentaites, 94, 95, 241, 255, 256, 258 — — brevidentatus, 93, 254-256 —— collinus, 94, 255 — calciphilus, 255, 256, 300, 301 — capitatus, 82, 83, 100, 269, 299, 300 — chelonoides, 77 —cordifolius, 82, 83, 92, 97, 261, 292, 294, 295 — cuspidatus, 241-243, 247, 248 _— ane 261, 280, 282 — elatus INDEX 575 marron erectus, 77 — 2, 92, 94, 99, 295, 296 — gla ne ; — glandulosus, 82,251, 261,)270,,274,.276, 278, Binbie 82, 260, 263, 265, 266, 267 — latifolius, 86, 92, 261, 274, 279, 280, 282 — laxiflorus, 93, 95, 260, 274, 276, 278 —longifolius, 82, "hag 97, 99, 261, 290, 292, 294, 295, 3 — iisianthus, 82, re 239, 2 —nigrescens, 81, 92, 94, a 245-248, 250 — — chiapensis, 93, 246 —— nigrescens, 93, 245, 246 — oreopolus, 93, 248-250, 253 — peduncularis, 93, 259, 263-267 — petenensis, 28 6 — quichensis, 93, 95, 252, 253, 255 d — saponarioides 86, 92; 236, 238, 240, 253, 255, 285, 286 — scopuli 264 ? — seemannii, 260, 284 — silenifolius, 95; 250, 2515-276 » 86, 94, 95, 259, 261, 263, 267 —stenophyllus, 275, 276 ae ase 95, 97, 261, 270, 272-274, Jean: 82, 83, 99, 297-300 — viscidiflorus, 93, 95, 241, 256-258 Lisianthus acuminatus, 238, 302 cutangulus, 302 — acutilobus, 302 —_ Soni 304 congestus, 237 bres scone 284, 287, 304 576 JOURNAL OF THE ARNOLD ARBORETUM isianthus crassicaulis, 304 — daturoides, 304 densi — Te 304 B ns, 307 — subcordatus, 308 Lisianthus Tatei, 308. —sect. Choriophyllum, 77 — sect. Helia, 77 — sect. Macrocarpaea, 77 — alatus, 302 — brevifolius, 303 — caerulescens, 303 — inflatus, 305 — loranthoides, 98 — Martii, 305 — Mathewsii, 305 — oblongifolius, 305 — obtusifolius, 305 — racemosus, 307 — tetragonus, 308 — uliginosus, 309 Litachne pauciflora, 320 Littorella, 226 Lysimnia, 523 Ma ea 523 — lutes Siaetenvuns 79, 80, 98, 553, 557 — subg. er tg — sect. Ma: gnaliifotes, 553 — sect. hanes 553 — domingensis, 557 o pinetoruth, 557 — quelchii, 81 [voL. 53 1972] Macrocarpaea salicifolia, 81, 557 — subcaudata, 553, 554, 556, 557 — thamnoides, 94, 554 assis om 554, 556 Malax ae ical ylis, 515 — * ah nag 515, 524 Malpighiaceae in ‘ne pepe United States, The, 10 Malpighiaceae, eke, —subfam. Gaudichaudioideae, 102 —subfam. Malpighioideae, 102, 109 — tribe Malpighieae, 109 Malpighiales, 103, 532 Malpighioideae-Galphimieae, 102 Mancoa, 57 Marchantia, 456 Masdevallia, 524 — decolor, 526 cary di Ta, —ramosa, 524 — superba, 521 EHRA, P. N.,, . SAREEN, and P, K Kyos . Cytological Studies on Him- alayan x Motiedsae, 558 Melastoma alpina, 401 562 — azedarach, 558, 560 — composita, 558, 560 — toosendan, 561 Meliaceae, Cytological Studies on Him- alayan, 558 ical pear "3 Veracruz, 312 Microstylis, 515, 5 Mitella, 413, pete Monsonia, 1 Monstera ‘abet, 320 Moraceae INDEX 577 in Cordyline ~— Morphological Studies (Agavaceae) II. Vegetative phology of Doategs terminalis, 1 Morus, 216, 2 — alba, 216, i 223 — australis, 217, 219, 223 — bombycis, 223 — laevigata, 217, 220, 222 Myricales, 27 Myriocarpa longipes, 320 Myrmecophila, 524 Myrobroma, 524 Myrtaceae, 49 Myrtus capuli, 345 Narica, 525 Nectandra salicifolia, 320 Neocogniauxia, 525 Neotropical Genus Lisianthius (Gen- tianaceae), A Revision of the, 76-100, 234-272, 273-311 Neottia acaulis, 525 — calcarata, 390, 391, 517, 520 — glandulosa 527 Neottoideae, 202, 205, 207, 211, 212 Neourbania, 525 Negeathals. 464 Nerissa, 52 “a Neuwiedia, New ine: oldehae Islands, and A heey 1, The Genus Piper Pipecicese Nidema, 525 Notes on Asi atic-Polynesian-Australian (Melastomataceae) in the Lesser An- tilles, Notes on West Indian Orchids, II, 390; TEL, 545 Nuciferoideae, 29 Nutmeg hickory, 35 Ocampoa, 525 — mexicana, 525 Ocotea dendrodaphne, 321 25 — serratifolia, aa 525 Odontoglossum, 525 — a eo pee 525 Olgasis, 525 Omphalea cardiophylla, 321 —Juridum, 394 578 Oncidium maculatum, 394 —variegatum, 525 eo — Origin of the Orchidaceae, II, peal corallorhiza, 518 ee 58S, S25 Piralis, 521 Orchidaceae, 211 subfam. Epi oreo nage Z214;213 —_ aie Neottioid tribe pee, 207 Orchids, 20 Or ie on West Indian, II, 390; Orchis monorrhiza, 521 9 Oreopanax capitatum, 321 Ornithidium, 526 Ornithocephalus, 526 Oxalidaceae, 183, 532 Pagaea, 79, 81 of Zan- Darién Based on the se tages of G. B. Hinton, 162 Passiflora Paullinia povioode 321 Ow | Piers 545 "52 6 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Phryma, 226-230 Phrymaceae in the Southeastern United States, The, 226 Piper (Piperaceae) New Guinea, Solomon Islands, wae Australia, 1, The ssp 1 Piper, 32 ~shreiatum, 13 — auritum, 320 — nu, 5 — bilobulatum, 13 — nei oe 1,2 — forstenii, 6, 7 — fragans, 1 — gibbilimbum, 8, 10 m, 5 seein, 320 — lauterbachii, 5 — macrocarpum, 5 — macropiper, 10, 11 — majusculum, 12 —martensianum, 320 — mestonii, 13, 17, 18, 20 —miniatum, 10 Bierce ton i 14, 16 — pubipes, — quinquenervium, 10, 11 — rhizocaule, 19, 20 — rhodo — ecteri, 20 — sclerophloeum, 16 1972] Piper squamuliferum, 10 2 Piratinera guianensis, 216 Pittosporales, 18. m magnum parvifolium, 2 Pleuranthodendron mexicana, 320 Pleurothallis, 526 —- opioids 391 — tuberculata, Polygalaceae, Polygalales, 102, 103 Polynesian-Australian species of Ery- Notes on Asiatic-, 128 Ponera, 526 9 ORTER, Duncan M. The Genera of UNC ophyllaceae in the Southeastern United States, 531 Porter, DUNCAN Three New Species ov phage from Darién Province, Pothomorphe umbels 321 Psilochilus, 52 » 52 Pte oideae is tt 5 530 INDEX 579 Puncture Weed, 543 Purdieanthus, 79, 80 Rapanea myricoides, 321 —— 521, 527 — modestus, 527 iboictateties gracilis, 320 Renealmia aromatica, 320 Restrepia parvifolia, 516 Revision of the Neotropical Genus Lis- ianthius (Gentianaceae), A, 76-100, 234-272, 273-311 Revision of Stenopetalum (Cruciferae), 52 Rhexia inconstans, 399 — longifolia, 401 3 Rhoiptelea, 27 ROBERTSON, KennetH R. The Genera of Geraniaceae in ae Southeastern United samara 182 ROBERTSON, Malpi- KennetH R. The aie in the Southeastern United States, 101 Robinsonella mirandae, 320 i=] re 8 8 i=} a i= 8 @ 103, 472 Rusbyanthas, 80, 553 Rutaceae, 532 Rutales, 102, 103 Sabatia, bat 93 Sacoila, Salacia methyl 322 Sapindales, 103, Sapium ae ean 321 Sarcocaulon, 183, se Se SAREEN, . Kosta, and P. N. ma Cytological Studies on Him- an Meliaceae, 558 Mocca ni om ea 515, 526 — repens, 521 Sauroglossum candidum, 521 Saxifraga, 412, 450 —sect. Aizoonia, 448, 449 580 pinta sect. Cymbalaria, 449 ermasea, . Discogyne, 447 — sect, omental 445, 446, 448, 449 —sect. Porphyrion, 44. — sect. Saxifr: — sect. eA gre 448 — sect. Xanthizoo Saxifragaceae in He rea iaaioe United States, The Genera of the, 409 Saxifragaceae, 409 — subfam. Baueroideae, 411 Z , 411 : Hydrangeoideae, 409, 480, 492 . Iteoideae, 411, 476 : Lepuropetaloideae, 411 m. Vahlioideae, 411 tribe Hydrangeeae, 487 —tribe Philadelpheae, 480 — subtribe Astilbinae, 422 my 52,57 Schomburgkia, 524, 528 —tibicinis, ees 5 ID, Rupo Floral Anatomy of Myrtaceae IT. tea 336 Schultesia, 85 ds from Surface Soils in ba Bicics egion of Veracruz, Mexico elenipedium, 2 Seraphyta, 528 Serapias polystachya, 517 Serjania, 321 Sevenbark, 488 WwW, ExizasetH A. Revision of Steno- petalum (Cruciferae), 52 Sicyos deppei, 321 Simaroubaceae, 49, 532 Sirium arborescens tertium, 10 — decumanum, 6, 12 Solanum, 320 — amazonium, 321 — nigrum, 321 JOURNAL OF THE ARNOLD ARBORETUM [voL. 53 Solomon Islands and Australia, I.. The ee Piper (Piperaceae) in New a, Southam United States, The Genera Geraniaceae in the Southeastern United States, The Genera Genera Genera ams in the, 101 Spigelia aitheliata, 321 Spiranthes, 515, 528 — esmeraldae, 524 — Galeottiana, 524 — lineata, bin, 321 SPONGBERG, STEPHEN A. The Genera of oo in the Southeastern 09 m, 52 Seuenieiie 52-54, 56-58, 61, 66, 69 Base a 53-55, 58, 61, 63, 6 — croce = deciens, uk 58, 70, 72, 73 — draboi wa — gracile, BE 730 sy 53-56, 58, 63, 65 — incisifolium, 7 — lineare, 52-61, ey 74 —— can — minus, 68 snipe Rana: 53, 54, 58, 67, 68 — pedicellare, 70 — procumbens, 74 — robustum, 52-54, 58, 69, 70 i nvinetpem 53-56, 58, 65, 66 — trisectum — velutinum, 53-55, 57, 58, 61, 72-74 Streblus asper, 217, 2 Studies on ie hy ogee Il. Xylary Comparisons, Woodiness, and Paedomorphosis, 499 Sturmia, 529 Sweet, Herman R., and Lest A. 1972] Garay. Notes on West Indian Orchids, II, 390; III, 515 oo ami 79, 80 Syngonium er 321 — podophyllum Synthlipsis, 57 Tabebuia, 548 78 Taraxacum officinale, 321 Tassel-white, 477 Tetradapa javanorum, 133 Telesonix, 441 Telipogon haa 529 bocbshccnggtnn —a s, 529 Ai Goteci ghiesbreghtii, 322 Thiebautia, 529 fb oHN W. The ee bnegnene in the Southeastern United States, 22 Three new species of Zanthoxylum ae seat from ‘Darién Province, Pana- Tiara, we 430-434 subg. Ant thonema, 432 subg. Tiarella, Tibouchina ad Cade (Melasto- mataceae) in the Lesser Antilles, Notes Oo Tibouchina, 399 — cham nd in Citharexylum ew benaceae), 386 ToMLINSO: B.. and: 3...5. R. ainbaaaa studies in Cordyline Agavaceae) II. Vegetative Morphology of Cordyline terminalis, 113 Trachelosiphon, 529 Tradescantia, 321 rem aceae, 103 Tribulus, 532, 539, 541-543 INDEX 581 Trichilia japurensis, 321 Trichopilia, 530 Trigoniaceae, 102, Trigonidium fe ae 525 Triorchos, 5 Triphora, 530 Tropaeolaceae, 183 Trophis mexicana, 322 — Tylochilus, 530 Ulantha, 530 Vanilla, 530 — mexicana, 530 Venation Patterns in the Leaves of o, Seeds from Surface Soils in a Tropical Region of, 312 Verbenaceae, 226 — tribe Lantaneae, 226 Verbesina greenmani, 321 Vernonia deppeana, 321 183 Vochysiaceae, 102 Walnut Family, 26 Warczewiczella, 530 ry, 3 WEAVER, aver E., Jr. A Revision of the Genus Lisianthius aes 76-100, 234-272, 273- Ric Ari) amet § Genus cece pte ena in Costa Rica, 553 gp Indian Orchids, Notes on, II, 390; pO ANE. | White liverwort, 461 Wild hen raniu mer 191 Wood-v shes sng 530 Xeilyathum, 530 Xylobium, 525, 530 Zanthoxylum (Rutaceae) from Darién Province, Panama, Three New Species 403 of, Zanthoxylum eliasii, 403 — jaimei — kellermanii, 403 582 JOURNAL OF THE ARNOLD ARBORETUM Zanthoxylum pucro, 406, 408 echelon — 103, 531, 538 ubfam. A geoideae, 532 — subfam. Chitontoidiean, 532 — subfam. Nitrarioideae, 532 —subfam. Peganoideae, 532 — subfam. Tetraiclidoiese, 532 —subfam. Tribuloideae, — subfam. pfasmemepet ay 532, 547 Zygophyllum, 531 Zygostigma, 95 [voL. 53