Annals of the Missouri Botanical Garden Volume XLIV 1957 Published Sec? at Galesburg, Illinois, by the Board of Trustees of the Missouri Botanical Garden, St. Louis, Mo. Entered as second-class matter at the post-office at Galesburg, Illinois, under the Act of March 3, 1879. Misecour: BOTA? "CAL A LIER Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the Missouri Botanical Garden and the Henry Shaw School of Botany of Washington University in affiliation with the Missouri Botanical Garden. Information e ANNALS OF THE Missouri BOTANICAL GARDEN appears four times during the calendar year: February, May, September, and November. Four numbers constitute a volume. Beginning with Volume 45, 1958: Subscription Price $12.00 per volume Single Numbers —--—---———— 3.00 each Contents of dem issues of the ANNALS OF THE Missouni BOTANICAL GARDEN are listed in the Arbeit Index, published by the H. W. Wilson Company. TABLE OF CONTENTS Some Lichens of Tropical Africa. IL Usnea (Continued ) Carroll W. Dodge Studies in the Capparidaceae. III. Evolution and Phylogeny of the Western North American Cleomoideae Hugh H. Iltis Stylites, A New Genus of Isoetaceae Erika Amstutz A Revision of the Fern Genus Pellaea section PELLAEA Alice F. Tryon A Revision of the North American Species of Gentianella Moench. John M. Gillett Morphological Complexes in Hops (Humulus lupulus L.) with Special Reference to the American Race... -Edward L. Davis Androcalymma, a New Genus of the Tribe Cassieae (Caesalpiniaceae) n D. Dwyer A Revision of the Genus Stalosantbe Robert H. Mohlenbrock General Index to Volume XLIV Errata in Volume XLIV:— Page 100, caption Fig. 13, read: L- 76 77-119 ‘121-123 125-193 195—269 271—294 295—297 299—355 357—360 Fig. 13. Geographic distribution and origin of genera. Arrows point from the less specialized to the more specialized of the two most closely related taxa, and in one genus and one in the other. Page 101, caption Fig. 13 (Continued), line 1—For “heavy” read "dotted' . MISSOURI BOTANICAL GARDEN STAFF Acting Director Hoen C. CUTLER EDGAR ANDERSON, GEorGE B. VAN SCHAACK, Curator of Useful Plants Acting Curator of Herbari CARROLL E Dopcz, JULIAN A. STEYERMARK, Mycologi Honorary Research Associate Rosert E. Woopson, JR., Joun D. Dwyer, Senior Taxonomist Research Associate Henry N. Aum, NELL C. Horner, Paleobotan Bibliographer and Editor TuIPOM VON SCHRENK, sistant Curator ng of Osc P. EX GLAESSNER, "ilconuniie Planks STEE BOARD OF TRUSTEES President Jonn S. LEHMANN Vice-President DANIEL K. CATLIN Second Vice-President EUGENE PETTUS LEICESTER B. FAUST RicHarD J. Lock woop DUDLEY FRENCH Henry B. PFLAGER Henry HitcHcock A. WESSEL SHAPLEIGH ROBERT BROOKINGS SMITH EX-OFFICIO MEMBERS ARTHUR C. LICHTENBERGER, STRATFORD LEE MORTON, Bishop of d the Diocese of ie och of the Ge of Science of St. Loui James J. McCarr ETHAN A. H. SHEPLEY, President of the ed of Chancellor of Washington Education of St. Louis University RaYMOND R. TUCKER, Mayor of the City of St. Louis Oscar E. GLAESSNER, Secretary | Volume XLIV Ahas o T of the p | dO Ut ur | Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the Missouri Botanical Garden and the Henry Shaw School of Botany of Washington University in affiliation with the Missouri Botancal Garden. | Informaton : ANNALs OF THE Missourr BOTANICAL GARDEN appears four times The A during the calendar year: February, May, September, and November. Four numbers constitute a volume. ` Subscription Price <3 $10.00 = volume «Single Numbers e 2.50 each = Contents of previous issues of the ANNALS OF THE ne ei GARDEN are listed in the Agricultural Index, published Lo the H. v. Wilson : Pp mk Annals of the Missouri Botanical Garden Vol. 44 FEBRUARY, 1957 No. 1 SOME LICHENS OF TROPICAL AFRICA. II. USNEA. (Continued) CARROLL W. DODGE* 1. UsNEA CERVICORNIs Dodge, Ann. Mo. Bot. Gard. 43:392. 1956. Type: Southern Rhodesia, Makoni Inyanga District, summit of Forrest-Hill Kop, 1700 m., on shaded rock face, Frederick Eyles 827, at Kew. Thallus erect, forming fragile tufts up to 8 cm. tall from a common holdfast, root-like fibers more than 2 mm. long (broken); unbranched or irregularly di- chotomous, 2 mm. in diameter, terete, somewhat flattened at the axils, citrine drab shading to olive buff above, not articulate nor inflated; surface areolate, sub- foveolate, minutely and irregularly papillate, papillae with black tips mostly broken off, leaving white punctiform scars, the younger portions smooth, irregularly white-reticulate; eramulose. Apothecia lateral, about 1 cm. below the recurved tip of a branch, 8 mm. in diameter, margin crenate, eciliate or with one to two cilia up to 8 mm. long, about 0.5 mm. in diameter, somewhat inflated in the middle, with obtuse tips; disc epruinose, ochraceous salmon or buffy brown; exciple surface similar to that of the thallus but cortex scarcely areolate and less white-punctate. Cortex fastigiate, 40—65 y thick, of slender hyphae in a gel, the outer 12 p deeply staining, heavily nubilated with minute brownish crystals, penetrated by occasional filaments of Trentepohlia, moniliform, cells nearly isodiametric, 7-8 p in diameter, outermost spherical protoplast not uniformly staining (cleavage of zoospores?). Medulla about 65 p thick, thinning somewhat toward the margin, of loosely woven, thick- walled hyphae about 4 p in diameter, heavily nubilated with minute hyaline crys- tals; algal layer under the parathecium about 25—30 p thick; parathecium hyaline, 40 p thick, hyphae perpendicular to the surface, thick-walled, slender, conglutinate; hypothecium of slender, somewhat loosely woven periclinal hyphae; thecium 75- ri Botanical Garden, St. Louis, Mo. en Mich 14, 1957 (1) I Vor. 44 2 ANNALS OF THE MISSOURI BOTANICAL GARDEN 80 » tall; paraphyses conglutinate, upper 15 p heavily nubilated with brown crys- tals, tips not thickened; asci clavate, 40—45 X 10-12 p, tips thickened when young; ascospores 8, monostichous, unicellular, about 8 X 4 u (in ascus probably not quite mature). Cortex of branch 190—200 y thick, consisting of an outer layer 40 y thick, of slender, predominantly radial hyphae so heavily nubilated by brown crystals that structure not clear even in thin sections, and a gelified layer 150—160 y thick, of hyaline, thick-walled, branched and interwoven hyphae predominantly radial; algal layer about 30 pu thick, of radial filaments of Trentepohlia, cylindric, 6-8 p in diameter, cells nearly isodiametric, sending broad columns of filaments up through the cortex, 25—135 y thick, forming fracture planes for the areolate cortex; medulla 80—95 p thick, of thick-walled, branched and interwoven hyphae, so densely nubilated with brown crystals that the structure is obscured; axis hollow, stuffed, 1500 p in diameter, chondroid cylinder shallowly lacerate on the outside, more so within, 250—300 p thick, of very thick-walled, interwoven, branched, predominantly longitudinal hyphae, lumen about 1 p in diameter; cavity 500— 750 pn in diameter, loosely stuffed with interwoven, predominantly radial hyphae 5-6 p in diameter, lumen 1—2 p, growing out as branches from the axis hyphae. This species is closest to U. Liechtensteinii, with a narrower cavity, ramuli and white-tipped papillae; the apothecia are terminal with a rugulose exciple. N . USNEA PERHISPIDELLA Steiner, Sitzungsber. K. Akad. Wiss. Wien, math.- naturw. Cl. 106:210. 1897. Type: Kenya, Machakos, Liechtenstein & Pospiscbill. Thallus about 6 cm. tall, fruticose, rather rigid, pale grayish green, drying raw umber to antique brown in the herbarium; base thick, slightly fuscous, often giving rise to several branches, rather closely dichotomous, internodes mostly 6-8 mm. long, axils acute to nearly right angles; branches about 2 mm. thick below, about 1.2 mm. thick above, continuous or distinctly annularly cracked, cortex sub- areolate, areoles plane, terete or only slightly deformed; papillae minute, elongating to short spinulose isidia which soon break away leaving minute pseudocyphelliform scars, often confluent in short rows, more rarely germinating in situ as slender pseudoramuli 2-3 mm. long, slightly curved upward; ultimate branches flexuous, slender, tips acute. Apothecia lateral, then appearing subterminal, sessile, 3-5 (—6) mm. in di- ameter, concave, cup-shaped, finally subplane; exciple rugose with short spines; marginal cilia rare, subdistant, mostly 2 mm. long, conic, a few up to 10 mm. long with spinules at their base; ascospores 10—13 7—8.5 p in clavate asci. Cortex up to 80 y thick, dilacerate, chondroid, fragile, of predominantly peri- clinal hyphae, the outer 20 p brownish, deeply staining, almost a thick-walled pseudoparenchyma; algal layer 25-50 y thick, of radial filaments of Trentepoblia, . 1957] DODGE-—LICHENS OF TROPICAL AFRICA. II. USNEA H cells 6-7 X 10—11 p; medulla 180, thick, closely woven to subarachnoid, K sulfur- yellow; axis solid, about 360 y thick. Only two young plants with a single apothecial initial were found in the material from S. Sudan. The material from Kenya is typical. S. SUDAN: Issore, Onejiro, 1775 m., corticole, 3° 50' N., 32° 50' E., T. F. Chipp 57, at Kew. KENYA: Endabarra-Mau Forest, 2355 m., on Acacia, P. R. O. Bally B4947, in E. African Herb. 3. UsNEA sonREDIOSULA Müll. Arg. in Motyka, Lich. Gen. Usnea Stud. Monog. 330. 193 Type: Tanganyika, Usambara, near Kumba marsh, C. Hols£ sub U. barbata v. sorediosula Müll. Arg. at Genéve. Thallus erect, rigid, fruticose, up to 8 cm. tall, pale ashy, drying deep olive buff from soredia, exposed portions of the axes Isabella color; holdfast conic, black, up to 4 mm. in diameter; base cylindric, 3 mm. tall, about 2 mm. in diameter, an- nularly cracked; repeatedly, often unequally, dichotomous at acute or right angles, the branches 1.6 mm. in diameter below, more slender at each dichotomy, variously curved and intricate, rarely concrescent, forming a dense tangle broader than tall; terete, indistinctly annularly cracked or subareolate below, smooth above between the low tubercles which are rather close, sometimes in rows or rarely concrescent in short low ridges, soon forming isidiose soredia and appearing farinose as the soredia are dispersed over the whole plant, except less dense over the basal internodes; eramulose; summits up to 20 mm. long, subcapillaceous, rather closely dichotomous but mostly broken off. Apothecia not seen. Cortex up to 160 p thick, very densely woven in upper 30 y, more loosely woven below, imbedded in a yellow gel; algal layer up to 130 p thick, of radial Trentepohlia filaments, cells up to 10 u long, 6-7 u in diameter, below forming a continuous layer about 50 p thick with projections into the cortex up to 80 p tall and some filaments growing in columns to the surface of the cortex to form the soredia; medulla 65-125 p thick, very densely woven under the algal layer and next the axis, loosely woven between, hyphae 4-5 p in diameter, lumen 1 p, nubi- lated with brownish crystals; axis 430 p thick, of interwoven hyphae 3 p in diameter, lumen 1 p, conglutinate, predominantly longitudinal. UGANDA: Kigezi, Kabale River, on trees, 1970 m., I. R. Dale L1; Mafuga, 2580 m., I. R. Dale 42; Ruizi River, 1390 m., T. Jarrett 402; all at Kew. 4. UsNEA NUTANS Dodge, Ann. Mo. Bot. Gard. 43:392. 1956. Type: Kenya, Northern Frontier Province, Mt. Kulal, 1935 m., on exposed rock, P. R. O. Bally 5685, at Kew. Thallus erect to decumbent, rigid especially below, about 8 cm. tall, cinnamon buff on the main axes shading to dark olive buff on the ultimate branches; holdfast black, hemispheric, 5 mm. in diameter, forming root-like structures below, giving [Vor. 44 E ANNALS OF THE MISSOURI BOTANICAL GARDEN rise to 10 or more primary axes above; bases below the first dichotomy varying from 1 to 10 mm. long, 1.5—2 mm. in diameter, blackened below; branching vari- able, dichotomous, at nearly right angles, the branches curving upward with short internodes below, then longer internodes and spreading branches in the mid-portion, more closely branched and the branches recurved above, terete, smooth below with the cortex annularly cracked and subareolate, minutely verrucose above, each ver- ruca bearing a single short isidium, black-tipped, decadent, leaving minute white scars; eramulose; the nodding ultimate branches smooth, minutely black-tipped. Apothecia not seen. Cortex up to 200 y thick, cartilaginous, outer 40 p fastigiate, of slender hyphae heavily nubilated with brownish crystals, the rest of loosely woven, branched and anastomosing hyphae in a gel, in some places penetrated by strands of somewhat disorganized algal filaments from below to a height of 100 y; algal layer variable in thickness, up to 50 u, of radial branched filaments of Trentepohlia 5—6 p in diameter, somewhat moniliform, cells 7-8 p long, apparently not helping to form the cortical cracks as in other species of this section; medulla about 100 p thick, of densely woven, slender hyphae 3 y in diameter, heavily nubilated with brownish crystals, tending to tear apart in the middle, allowing areoles to flake off and making sectioning difficult; axis 550—575 p in diameter, the inner half minutely lacerate, forming many small stuffed cavities, not a single cavity as in most other species. The habit of our species is reminiscent of a tuft of a species of Cladina with recurved tips. It resembles the smaller U. capensis somewhat in habit but differs in many macroscopic and microscopic details. UGANDA: Kigezi, Kabale River, 2260 m., on tree, I. R. Dale L43, at Kew KENYA: Northern Frontier Province, Mt. Kulal, 1935 m., on exposed ook. FR OUO. Bally 5686, type at Kew 5. Usnea DopoMar Dodge, Ann. Mo. Bot. Gard. 43:392. 1956. Type: Tanganyika, Dodoma Prov., Mt. Mpwapwa, 1610—1935 m., saxicole, P. J. Greenway 2432, in East African Herb. Thallus erect or decumbent in dense tufts, 10 cm. tall, rigid, from a common holdfast, up to 15 mm. in diameter, of dense yellowish, root-like processes at least 3 mm. long, penetrating between the rock crystals; branches about 1.5 mm. thick at the blackened base, expanding to 2 mm. below the first dichotomy, 10-15 mm. above the holdfast; further branching variable from dichotomous to almost sym- podial, ultimate branches up to 7 cm. long; flexuous, tapering gradually to the acute tip; eramulose (occasionally the ultimate branches once or twice unequally dichotomous 1-2 cm. from the tip, forming pseudoramuli), light brownish olive shading upward through Isabella color to almost honey yellow at the tips; not articulate nor inflated; cortex cracked and subareolate below, otherwise smooth, densely covered throughout most of the plant with short spinulose isidia which 1957) DODGE—LICHENS OF TROPICAL AFRICA. IJ. USNEA 5 soon break away leaving minute punctate scars (appearing as pseudocyphellae or when aggregated in short rows on indistinct ridges, giving a slightly foveolate appearance), the ultimate 2 mm. aude of spinules. Apothecia not seen. Cortex about 180 p thick, horny, fastigiate above (upper 15 p), more loosely woven below, all imbedded in a gel; algal layer about 50 p thick, of more or less radial filaments of Trentepohlia 4—5 p in diameter, cells varying from isodiametric to 8 » long; medulla 65—70 p thick, of very densely woven hyphae, nubilated with brown crystals, thick-walled, easily cracking away from the axis in sectioning; axis about 970 y in diameter, hollow, the walls about 360 un thick, of very densely woven, thick-walled hyphae in a gel; central cavity somewhat irregular in outline, 250—350 pn in diameter, stuffed with loosely woven brownish hyphae. TANGANYIKA: Dodoma Prov., Mt. Mpwapwa, 1619-1935 m., saxicole, P. J. Greenway 2432, in E. African Herb. 6. UsNEA PULVINULATA Dodge, Ann. Mo. Bot. Gard. 43:392. 1956. Type: Uganda, Kigezi, Kasatoro forest, 1935-2560 m., on trees, I. R. Dale L39, at Kew. Thallus fruticose, erect, up to 5 cm. tall, between wood brown and deep olive buff (seen from above), the main branches citrine drab, blackening below; hold- fast a black verrucose disc about 8 mm. in diameter, the base black, about 4 mm. tall, 2 mm. in diameter; branching dichotomous to sympodial, internodes short, producing a subhemispheric tuft, the ultimate branches about 2 cm. long, 1 mm. in diameter, annularly cracked and subareolate in the lower blackened portion, otherwise smooth but very densely ramulose, the ramuli rather fragile below and the broken bases suggesting papillae, some regenerating as several spinules which probably serve as isidia; the cortex cracking off in patches, exposing the light coral red lower medulla and surface of the axis; ramuli acuminate, 0.5-3 mm. long, straight, axils acute. Apothecia subterminal, mostly appearing terminal, up to 15 mm. in diameter, cupulate at first, becoming nearly plane; exciple shallowly scrobiculate, closely ramulose along the ridges; marginal cilia crowded, about 2 mm. long, tips acu- minate, rather fragile and many broken off in older apothecia; disc smooth to somewhat unequal, densely white-pruinose when young, becoming cinnamon rufous to apricot orange. Cortex 65—100 y thick, fastigiate, hyphae 6-7 um diameter, lumen 3 y, the outer 30 u nubilated with minute brown crystals, gelified, the hyphae less clearly radial in the inner portion; algal layer about 40 p thick, of predominantly radial filaments of Trentepohlia; medulla 30 p thick, of coarse periclinal hyphae 6-7 p in diameter, relatively thin-walled; axis 1550 & in diameter, hollow, the wall 270 y thick, of longitudinal hyphae 6 p in diameter, lumen 2-3 p, cavity 1100 p in diameter, stuffed with branched brown, thick-walled hyphae, 6-7 p in diameter, growing out of the wall. Vor. 44 6 ANNALS OF THE MISSOURI BOTANICAL GARDEN In habit, this species suggests the pENsIROsTRAE, the hollow axis and red lower medulla suggest the subgenus EuMrTRIA, although the presence of Trente- poblia algae clearly place it in the ROCCELLINAE. UGANDA: Kigezi, Kasatoro forest, 1935-2560 m., on trees, 1. R. Dale L30, at Kew. N . Usnea LIECHTENSTEINII Steiner, Sitzungsber. K. Akad. Wiss. Wien, math.- naturw. Cl. 106:211. 1897. Type: Kenya, Machakos, Liechtenstein & Poppiscbill, saxicole. Thallus 5-10 cm. tall, erect, caespitose, fruticose, rigid, citrine drab to deep olive buff, becoming brownish olive to Isabella color; holdfast blackened, sending root-like fibers about 0.5 mm. in diameter between the rock crystals, furnishing a common base for many primary axes; lower joint 5-10 mm. long, 1.5 mm. in diameter at the base, expanding upward to 2 mm. in diameter at the first dichot- omy, several times dichotomous with very short internodes, then less frequently and irregularly so above; terete, cortex areolate, the areoles slightly foveolate below, then closely verrucose and subpapillate above, the verrucae sometimes in close rows but never distinctly angled, tips white but not perforate nor sorediate; ramuli irregularly disposed, thick, curving upward and subappressed, mostly 4—5 mm. long, nearly smooth, rather fragile and the obtuse tips often broken off. Apothecia lateral, about 1 cm. back of the tips of the smaller branches, cupu- late, 5-7 (-10) mm. in diameter; exciple smooth at first, becoming reticulate- rugose in relatively large areoles at maturity; marginal cilia erect, not dense, about 3 mm. long, then more spreading and annularly cracked, especially at the base, and often broken off at maturity; disc epruinose, buffy brown. Cortex gelified, about 275 p thick, the outer 25 p obscured by brownish crys- tals, the hyphae thick-walled, branched and anastomosing but predominantly radial, 2.5—3 p in diameter; algal layer about 50 p thick, of Trentepohlia filaments, mostly radial, about 6 un in diameter, cells cylindric, somewhat longer than broad, pene- trating the cortex in bands and forming fracture planes between the areoles of the cortex; medulla 50-80 p thick, of slender, very densely woven hyphae, heavily nubilated with minute brownish crystals, axis 750 p in diameter, hollow, wall 250— 300 p thick, of very thick-walled, densely woven hyphae in a gel; cavity about 125—150 nm diameter, irregular in outline, more or less stuffed with medullary hyphae which grow through the axis wall in radial strands about 150 p» wide, analogous to the rays of woody stems of flowering plants and perhaps performing the same functions. ANGANYIKA: Dodoma Prov., Mt. Mpwapwa, 1610—1935 m., saxicole, P. J. Greenway m at Kew; SE jn’ E. Aízitin Herb: (separated from the collection of U. Dodomae Dodge). 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 7 8. USNEA sUBCRISTATA Dodge, Ann. Mo. Bot. Gard. 43:393. 1956. Type: Tanganyika, Kondoa District, Swaga-swaga Hill, summit, 1775 m., on shrubs, chiefly Tricalysia cacondensis, B. D. Burtt 891, at Kew Thallus erect, very rigid, fruticose, 7-8 cm. tall, light BEE olive below, shading through Isabella color and honey yellow to chamois on the slender branches and ramuli, the upper ends of branches between chamois and deep olive buff; base and first branchings not seen as the plant has been glued to the herbarium sheet, and are obscured by several layers of branches; branch 7 cm. long, 2 mm. in di- ameter at its base, expanding to 3 mm. at the first lateral branch and then tapering slowly to about 1 mm. in diameter below the apothecium, rather densely ramulose, nearly terete in the lower two internodes; cortex subareolate, then coarsely papillate above, the papillae in predominantly longitudinal rows, making the branch angled, some of the papillae growing out to form ramuli; larger lateral branches similar to the main axis but 1—1.5 mm. in diameter; ramuli 1—15 mm. long, thick, tips rounded, the larger foveolate, the smaller nearly smooth; some of the larger ramuli may develop into slender lateral branches 0.4—0.5 mm. in diameter, sparsely papillate below, foveolate above, tips rounded. Apothecia apparently terminal, nearly plane, up to 20 mm. in diameter; exciple smooth or papillate with radial ridges (as many as there are ridges at the top of the branch), smooth between; marginal cilia 2-ranked, subdistant, the outer rank coarse, often dichotomous above the base, tips acute, otherwise resembling larger ramuli, the inner rank slender, 1-2 mm. long, tips acute; disc vinaceous buff, or pale olive buff, pruinose. Cortex variable in thickness, 50—90 y, fastigiate, highly gelified, the outer 10 p deeply staining; algal layer dentate outward, up to 40 p thick, cells 6-7 nm diameter, closely packed, protococcoid; medulla K—, about 130 p thick, brownish, of densely woven hyphae with a narrow zone of loosely woven hyphae in the middle; axis 1400 p thick, hollow, wall 300 p thick, of longitudinal conglutinate thick-walled hyphae; cavity about 800 p in diameter, more or less lined with a layer of thin-walled hyphae 10 p in diameter, with deep brown contents; very loosely stuffed with hyaline thin-walled hyphae 5 p in diameter, nubilated with hyaline crystals and a few hyphae similar to those lining the cavity. This fertile highland species differs from the lowland, sorediose U. cristata in being a coarser plant with more but less sharply defined angles on the main axis and larger branches and in the more abundant, larger and subfoveolate ramuli. 9. USNEA ELATA Motyka, Lich. Gen. Usnea Stud. Monog. 57. 1936. Type: Congo; M'pala, west shore of Lake Tanganyika, Guillemé. Thallus 20 cm. long, probably erect, rather rigid, buffy olive to ecru olive; holdfast a coarsely verrucose disc 5—6 mm. in diameter, dark brown, giving rise to one or more main axes 2 mm. in diameter, with many small, short branches in the [VoL. 44 8 ANNALS OF THE MISSOURI BOTANICAL GARDEN lowest cm., then repeatedly dichotomous in the lower half, internodes about 1 cm. long, thinning at each dichotomy, axils complanate, nearly 90? angles, the branches soon erect before the next dichotomy, the internodes subterete to few-angled from low longitudinal or suboblique ridges, cracked and minutely isidiose, appearing farinose after the isidia are shed; in the upper two-thirds the ridges more irregular and anastomosing, giving a foveate appearance with minute isidiose verrucae between, the internodes more variable in length, some very short, giving a tri- chotomous appearance, others very long; summits long-capillaceous, rather fragile and broken off in our plants; eramulose (occasionally isidia grow out in situ form- ing groups of pseudoramuli 1-2 mm. long) or the small branches on the basal internode produce larger pseudoramuli from unequal dichotomies. Apothecia rather rare, about 5 mm. in diameter, cupulate, sessile on the larger axes; exciple smooth to slightly impressed; marginal cilia moderately close, erect, variable in length up to 4 mm. long, the longer dichotomous near the tips, sparsely and minutely verrucose, perhaps finally isidiose; disc flesh color with a white pruina. Cortex variable in thickness, 50-85 (—150) yu thick, gelified, loosely fastigiate, the outer 15 u very dark brown; algal layer 30-50 p thick, of discrete colonies of Trebouxia, cells 4-7 p in diameter, forming a nearly continuous layer; medulla pink to rose color, I—, about 240 y thick, quite closely woven next the algal layer, looser below (appearing dense under low magnification), hyphae 3—4 p in diameter, lumen 1 p; axis 840 pn in diameter, deep rose by reflected light, hyaline by trans- mitted light, slightly irregular in transverse section, wall about 200 p thick, but cavity very irregular in outline (radiating sinuses), from the coalescence of many small cavities, about 30 p in diameter, lined with larger thin-walled brownish hyphae, nearly hollow but traversed by an occasional hyaline hypha 6 y in diameter, lumen 3 A: wall hyphae longitudinal, very thick-walled, conglutinate, I slightly : blue, especially next the cavity. Jade Lake Ngunga, 1775 m., on trees, G. M. Allen 1782, in Howe herb. at Farlow erb. 10. UsNEA NANA Dodge, Ann. Mo. Bot. Gard. 43:393. 1956. Type: Uganda, Kigezi, Kasatoro, Kabale River, 2256 m., corticole, I. R. Dale L43 p. p. min. Thallus fruticose, erect, up to 4 cm. tall, deep olive to citrine drab; holdfast 2 mm. in diameter, conical, black, base 3—5 mm. tall, up to 1 mm. in diameter, black; once or twice dichotomous at the base, rarely dichotomous above; axes terete, up to 1 mm. in diameter below, tapering upward, smooth, densely ramulose, ramuli 1-2 mm. long, axils acute, tapering to a slender, obtuse, not blackened tip, not fragile, a few up to 4 mm. long, curved, rarely once-dichotomous; summits tapering to an acute tip, nude of ramuli in the upper 4 mm. Apothecia sessile on the lower portions, of small secondary branches, cupulate, 1 (-2) mm. in diameter; exciple smooth; marginal cilia close, radiating, 1 mm. long, 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 9 resembling small ramuli but more fragile; disc concave, flesh color, white-pruinose. Cortex 35-40 y thick, of coarse conglutinate thick-walled predominantly radial hyphae 8 p in diameter, lumen 3 p (in places appearing pseudoparenchyma- tous); algae Trebouxia, in colonies 20-30 in diameter, in a nearly continuous layer, cells 7-8 p in diameter; medulla up to 15 p thick, of compact, periclinal thick-walled hyphae 3 & in diameter; axis 800 p in diameter, hollow, wall 150 y thick, of slender thick-walled conglutinate hyphae; cavity 500 y in diameter, very loosely stuffed with very thick-walled hyphae 3 u in diameter. 11. UsNEA FIRMULA (Stirton) Motyka, Lich. Gen. Usnea Stud. Monog. 64. 1936. Eumitria firmula Stirton, Scottish Nat. 6:100. 1881. Type: Fernando Po, lava beds, G. Thomson, in the Art Galleries of the Glasgow Corporation. Thallus about 10 cm. tall, rigid, ashy green, sparingly branched, closely ramu- lose; branches straight to somewhat flexuous, terete, 1-2 mm. in diameter, tapering somewhat above and below, thus appearing somewhat inflated, verrucose-papillate between the ramuli (perhaps only very young ramuli or scars of broken ramuli regenerating) ; main axes and bases of slender lateral branches densely ramulose, ramuli 5-15 mm. long, perpendicular, straight or slightly flexuous and arcuate, mostly annularly cracked at the base showing the rose medulla, tips slender but obtuse, not very fragile, breaking off near the tips; summits of slender lateral branches nude with abundant low white tubercles in our plants. Apothecia terminal on lateral branches, 7-10 (—15) mm. in diameter, nearly plane; exciple convex, papillose, sometimes ciliate; marginal cilia rare, 5 mm. long, smooth; disc nearly plane, pale flesh color with a white pruina; epithecium dark; ascospores 8 X 10 m, in narrow asci. Migeod 419 is very old; the apical portions of most of the main axes are broken off, carrying away the apothecia. The color is between buffy brown and citrine drab rather than ashy green. The axis is black under the rose-colored medullar surface. It agrees in vegetative characters better with this species than with the sterile sorediose U. implicita (Stirton) Zahlbr. from the same region. The low white tubercles on the apical portions of the slender lateral branches may be initials of soralia but are not found on the main axes as are the soralia of U. implicita. Dale L43 seems to be the juvenile stage of this species: It is only 3 cm. tall, the axis still solid, the medulla only partly rose color. It is sorediose and begin- ning to form an apothecium less than a mm. in diameter. CAMEROONS: gully above Bamenda, 1930 m., F. W. H. Migeod 419; Cameroon Mt. 1610 m., Dunlap; above Buea, 1450 m., F. W. H. Migeod ss all at Kew; Segel misit Zahlbruckner sub U. Baileyi, i in Howe herb., at Farlow FERNANDO Po: lava beds, G. Thomson, type, in the Art pum of the Glasgow Cor- "ist Onyiro, Issore, 1775 m., 3? 50' N., 32? 2r E., T. F. Chipp 57, at Kew. Kigezi, Kasatoro, Seier River, 2250 on p ped young, I. R. Dale he Mods Pass, 2250 m., on trees, E. M. Sbillito "ar. "both st [Vor. 44 10 ANNALS OF THE MISSOURI BOTANICAL GARDEN 12. UsNEA MEDIO-AFRICANA Dodge, Ann. Mo. Bot. Gard. 43:393. 1956. Type: Uganda, Kigezi, Mafuga, 2400 m., on trees, I. R. Dale L41, at Kew. Thallus erect, subrigid, about 5 cm. tall; common holdfast about 7 mm. in diameter, flat, black, giving rise to 4—5 primary axes about 1.5 mm. in diameter; base very short, once or twice dichotomous with very short internodes and again dichotomous less than 1 cm. from the apothecia, mummy brown with dark olive buff ramuli; terete to slightly rugose, angular and subfoveolate; ramuli dense, 5—7 mm. long, very fragile and mostly broken off, the scars giving a coarsely verrucose to papillate appearance, smooth, annularly cracked, tips acute and minutely blackened. Apothecia subterminal up to 20 mm. in diameter, plane; exciple foveate, minutely verrucose, with radial ridges from concrescent ramuli; marginal cilia rather close, of variable length, up to 8 mm. long, the larger once dichotomous or sparsely ramulose; disc cinnamon, very slightly white-pruinose. Cortex 60 (—100) pn thick, of predominantly radial, loosely woven slender hyphae in a gel, the outer 15 a more densely woven and deeply staining; algal layer continuous, about 30 p thick, of closely packed protococcoid cells; medulla 40 (—180) p thick, white, pale rose next the axis, of very loosely woven, slender, thick-walled hyphae, more densely woven next the algal layer and the axis; axis 900 u in diameter, solid at first becoming hollow, wall 180 p thick, of slender conglutinate hyphae, cavity 550 p in diameter, stuffed at first then hollow. In sectioning the mature primary axes, the cortex and algal layer tear away from a thin layer of medulla next the axis, making it impossible to estimate the thickness of the mature medulla accurately. In Hendrickx 4330 the medulla is 30—35 un thick, compactly woven throughout. i This species is closest to U. trullifera Nyl. of Motyka's monograph and perhaps the atypical specimen cited by him from Tanganyika?. Mt. Mfimbwa, near Lake "Tanganyika-Nyassa", 2300 m., From, belongs here. Our plants differ from the Tahitian plants in being smaller with much larger apothecia and relatively thicker medulla. UGANDA: Kigezi, Mafuga, 2400 m., on trees, I. R. Dale L41,a CONGO: km. 30, route Kahusi, ramulicole, F. L. Hendrickx js xu in E. African Herb. 13. USNEA BRUNNESCENS Dodge, Ann. Mo. Bot. Gard. 43:393. 1956. Type: Uganda, Kigezi, Mafuga, 2400 m., on trees, I. R. Dale L41, at Kew. Thallus at least 10 cm. tall, fruticose, rather rigid, drying Mars brown to russet on the axes, younger ramuli cinnamon buff; holdfast a black disc about 7 mm. in diameter, giving rise to at least three branches, twice or thrice dichotomous with short internodes (1 cm. or less), then rarely dichotomous above, about 2 mm. in diameter in the lower half, then tapering to about 0.3 mm. at the densely ramulose blunt tip; cortex impressed and subareolate, nearly smooth in the lower internodes, 1957] " DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 11 then very densely verrucose-papillate and ramulose above; ramuli 1—3 mm. long, single or in short oblique rows, but the branches not clearly angled, annularly cracked at the base and often the tips broken off, but seldom breaking away at the base to leave scars, the larger ramuli sparsely verrucose, producing single isidia on the verrucae, but not sorediate. The tips of the branches end in a dense cluster of ramuli, perhaps a regeneration phenomenon following the breaking off of the end of the branch. Cortex about 50 y thick, of loosely woven, predominantly radial, slender hyphae in a gel, the outer half nubilated with minute hyaline crystals; algal layer about 40 pu thick, a nearly continuous layer of Trebouxia colonies, cells about 6 p in diameter; medulla about 30 (—40) p thick, of slender, very densely woven hyphae; axis about 1400 pu in diameter, hollow, wall 135 p thick, of longitudinal, thick- walled conglutinate hyphae, lumen about 2 un in diameter, cavity loosely stuffed with branched, anastomosing thick-walled hyphae up to 4 p» in diameter. UGANDA: Kigezi, Mafuga, 2400 m., on trees, I. R. Dale L41, at Kew. 14. USNEA VERSICOLOR Motyka, Lich. Gen. Usnea Stud. Monog. 74. 1936. Type: Cameroons, Cameroon Mt., 4040—4070 m., on volcanic ash, Weberbauer. Thallus prostrate or subpendent, 20—30 cm. long, larger joints clay color to cinnamon buff, others light brownish olive, slender branches deep olive buff, base not seen, subarticulate, branching subdichotomous, the larger branch continuous at the axil, the smaller articulate and constricted, cracks narrow, dividing into joints not constricted at the ends but slightly inflated in the middle, somewhat flexuous, surface very slightly and indefinitely foveolate, with small round pseudo- cyphellae, or short, oblong, scarcely elevated farinose longitudinal ridges; main axes up to 1.5 mm. in diameter below, tapering above to the capillaceous summits; eramulose but the more closely dichotomous, very slender branches of the summits may simulate ramuli. Our material from the same mountain at approximately the same elevation agrees well with Motyka's description except for the presence of inconspicuous pseudocyphellae and farinose ridges. As our plants are somewhat longer, perhaps they are older and these structures more developed. This species seems intermediate between the FOVEATAE and the ARTICULATAE. CAMEROONS: Cameroon Mt., 4030 m., on bare rocky ground, F. W. H. Migeod 196, at ew. 15. UsNEA COMPLANATA (Müll. Arg.) Motyka, Lich. Gen. Usnea Stud. Monog. 109. 1936. Usnea barbata v. strigosa f. complanata Müll. Arg., Hedwigia 31:276. 1892. Usnea florida v. strigosa f. complanata Zahlbr., Cat. Lich. Univ. 6:574. 1930. Usnea florida v. strigosa f. pygmaea Hue, Nouv. Arch. Mus. Paris IV. 1:35. 1899. Usnea Schim peri Lenormand, herb. nom. z [Vor. 44 12 ANNALS OF THE MISSOURI BOTANICAL GARDEN Type: Ethiopia, Debr; Eski, 2900 m., on trunks of Euphorbia Collqual. W. Schimper, 1852. The type of U. florida v. strigosa f. pygmaea Hue is: Ethiopia, Nahia in Machiha, 1930—2250 m., on Euphorbia Collqual, W. Schimper 1022, Nov. 1852. Thallus erect, up to 3 cm. tall (only 1 cm. tall in our plants), snuff brown to Sayal brown; holdfast hemispheric, verrucose, 5 mm. in diameter, giving rise to several axes varying from nearly terete and foveolate, 2 mm. in diameter, to quite flat, 5 mm. wide, and deeply scrobiculate (resembling a small Ramalina), with 1—3 short flat secondary branches, rather densely ramulose in places, nearly nude in others, ramuli infundibuliform at the base, inflated and abruptly narrowed to an acuminate tip, 1-2 mm. long. Apothecia up to 15 mm. in diameter, terminal, somewhat excentric; exciple foveate with sparse ramuli, some terete, some flattened; marginal cilia rather close, about 2 mm. long, resembling ramuli; disc somewhat concave, flesh color with a white pruina. ETHIOPIA: M. Bachit, 2400 m., W. Schimper 14, ex herb Sbarbaro sub U. Schimperi Lenormand, at Farlow Herb. (duplicate cited by Motyka). 16. USNEA CORRUGATA Motyka, Lich. Gen. Usnea Stud. Monog. 105. 1936. Type: Ethiopia, Podopolani, on Euphorbia, Schweinfurth, 1894. Thallus subpendulous, up to 12 cm. tall, subrigid, brownish olive with ramuli and more slender branches partly buffy olive, very rarely dark red-maculate; base about 5 mm. tall, 2 mm. in diameter, expanding below into an irregular disc; hold- fast 4 mm. in diameter, densely sympodially branched just above the base, axils only slightly flattened; branches perpendicular or nearly so, curving apically, secondary branches 10-15 mm. long, nearly 2 mm. in diameter, then trichotomous with greatly flattened axils, then more closely dichotomous or irregularly branched, the branches about 1 mm. in diameter, upper ones 0.5 mm. or less, deeply foveate throughout, the larger rugi longitudinal, acute, bearing dense pale greenish isidiose soredia; ramuli very irregular, often in dense groups on the smaller branches (per- haps from isidia germinating in situ), the longer ramuli branched, subarticulate and inflated; summits quite closely divergently branched, tips obtuse. Apothecia not seen. UGANDA: Kigezi, Kabale, 1960 m., corticole, I. R. Dale Lr, at Kew. 17. UsNEA PULVERULENTA (Müll. Arg.) Motyka, Lich. Gen. Usnea Stud. Monog. 107. 1956. Usnea barbata v. pulverulenta Müll. Arg., Flora 68:499. 1885. Usnea florida v. pulverulenta Stzbgr., Ber. Thátigk. St. Gall. Naturw. Ges. 1888/ 89: 1945. 1890. Type: Ethiopia, Bagla, 2410 m., Hildebrandt, 1885. 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 13 Thallus fruticose to subpendent, over 6 cm. long (only 4 cm. in our plants), rather flaccid, axis wood brown, ramuli deep olive buff; holdfast a verrucose brownish disc 2 mm. in diameter, base 1 mm. tall, 1 mm. in diameter; dichotomous with internodes 5 mm. long below, 1 mm. in diameter, terete, smooth between the rather dense ramuli, tuberculate to subpapillate above, thinning at each dichotomy, with granular farinose soralia, subfoveolate between, ultimate branches slender, acuminate but rather short; ramuli dense throughout, mostly about 2 mm. long, some 3—5 mm., probably small secondary branches, tuberculate and sparsely ramu- lose, slightly inflated, tips acute. Cortex 80 y thick, gelified, fastigiate, almost pseudoparenchymatous, cells about 4 » in diameter, 6-7 p long, rather thin-walled below, thicker-walled above with very narrow lumina; algal layer 20-35 p thick, of Trebouxia, cells 7-10 p in diameter, somewhat angular from mutual pressure; medulla about 225 y thick, the outer half of densely woven predominantly periclinal hyphae, the inner half of pre- dominantly radial, very loosely woven hyphae about 4 p in diameter, sparsely nubilated with hyaline granules, K— or faintly yellow (the axis tears loose from this inner zone in sectioning, making measurement of the thickness of the medulla inexact); axis 350 p in diameter, of conglutinate predominantly longitudinal thick- walled hyphae. Although placed by Motyka in the FovEATAE, it might equally well be placed in the GLABRATAE, PYCNOCLADAE, near U. ruwenzoriana Motyka, as the branches are not, or only very slightly, foveolate above, and the color becomes fuscous in the herbarium; the branches are more densely ramulose than is usual in the FOVEATAE. ETHIOPIA: Bagla, 2410 m., Hildebrandt, 1872, ex herb. Sbarbaro at Farlow Herb. 18. USNEA FLAVESCENS Motyka, Lich. Gen. Usnea Stud. Monog. 125. 1936. Type: Kenya, Ngongo Bagas, 2700 m., A. G. Allan. Thallus pendent, flaccid, up to 60 cm. long, pale olive buff (in our material), finally becoming pale stramineo-fuscous (fide Motyka); holdfast short-conic, black, base slender, about 10 mm. long, 0.8—1 mm. in diameter, 3- to 4-annulate, then more or less frequently dichotomous, often unequally so throughout, articu- late, the joints 6—30 mm. long, constricted at the ends, cracks narrow, partly slightly foveate, abundantly so on some joints, very rarely so on others, cortical cracks (pseudocyphellae) rare, inconspicuous, increasing rapidly to 2-2.5 mm. in diameter and decreasing slowly to the almost non-articulate, capillaceous summits; the larger branches somewhat flexuous, sparingly ramulose, sometimes several con- secutive joints eramulose; the smaller branches more closely ramulose (imitating a small U. longissima) ; the ramuli of the larger branches up to 1 cm. long, some- times branched and spinulose; the ramuli of the smaller branches 1-2 mm. long, usually articulate at the base, tips acute and sometimes blackened. Apothecia sessile on small secondary branches, 2-2.5 mm. in diameter, plane; exciple shallowly foveate; marginal cilia in two ranks, the inner rather close, 1—2 [Vor. 44 14 ANNALS OF THE MISSOURI BOTANICAL GARDEN mm. long, suberect, the outer rank subdistant, 3—4 mm. long; disc flesh color, slightly white-pruinose (based on Curtis 669 from Kenya). Greenway 3816 is a mixture of several species. On one of the two sheets at the E. African Herbarium is a single plant which seems to be this species, but the color is now (after 21 years) between snuff brown and Sayal brown throughout; some of the joints are very foveate; ramuli more abundant on the larger joints and less so on the capillaceous summits, most of which have been broken off in collecting. ERITREA: Habab, Tsertsera, F. Gallina, ex herb. Cufino at Farlow Herb. ETHIOPIA: Simien, Mindigabsa, Qut H. Scott 323, at Kew KEN Mt. Kenya, 2250—2575 m., on bamboo twigs, G. M. Allen 1 797, in Howe and Riddle ds trail from West SCH Forest Station to summit of Mt. Kenya, 3650 m., Edgar A. Mearns 1 522 (T. Roosevelt Exp.); 40 miles southeast of Narok, 2250 m., Anita teca: Curtis 669; all at Farlow Herb. ANGANYIKA: Kilimanjaro, Bismarck Hill, 3220 m., P. J. Greenway 3816 p. p. min., in E. reos Herb. 19. UsNEA PRAELONGA Stirton, Scottish Nat. 6:107. 1881. Type: South Africa, without locality, J. H. McLea, in Brit. Mus. Thallus pendent, flaccid, up to 60 cm. long, gray to yellowish gray, drying dark olive buff to deep olive buff or colonial buff in recent collections to chamois in older collections (Isabella color in the type), lighter on the capillaceous summits; densely dichotomous at the holdfast, branches about 0.5 mm. in diameter, scarcely inflated in the lowest cm., then articulate, joints 5 (—10) mm. long, 2 (-3) mm. in diameter, then sparingly and irregularly dichotomous for the next 30 cm., then repeatedly dichotomous, the joints abruptly more slender and almost cylindric, rather densely dichotomous at the summits; eramulose or very rarely ramulose in the capillaceous summits; surface smooth to somewhat impressed or foveate on the larger joints, pseudocyphellae very inconspicuous, scarcely visible under 10 X magnification, the slender branches minutely foveate with the cortex cracking along the ridges. No apothecia nor soredia seen. Associated on the sheet of Gillett 4708 are evidently juvenile plants, the joints not inflated, the surface more foveolate, resembling the smaller branches of the larger plants. Glover & Gilliland 476 are shorter plants, tawny olive to clay color (collected in 1944), but they seem to be this species rather than U. articulata v. Kotschyana Motyka from Syria, which they resemble in color. In the Eritrea to Somaliland region, this species seems confined to Juniperus, and the Somalis consider it a part of the juniper tree, it is so constantly associa The natives use it as a stuffing for pillows and beds. The Ishaak tribes call it sarokh or jeban. TREA: Filfa forest, P. E. Ellis 5, at Kew. SOMALIE FRANÇAISE: Dai, 1450 m., G. Popov 1296 (Desert Locust Survey). BRITISH SOMALILAND: without locality, R. S Taylor, Aug. 1927; Libah Mts., Wawa- nak Peak, iom m., 10° 20’ N., 43° E., J. B. Gillett 4708 (Abyssinia-Somaliland Bound. mm.) ; Surud Range, E Beris, 10° 45' N., 47° 12' E., Co e tte 402; Wagar qup ive mist belt, P. E. Glover e H. B. Gilliland 470: all at K AFRICA: without locality, J. H. McLea, type, in 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 15 20. USNEA PSEUDOCYPHELLATA Motyka, Lich. Gen. Usnea Stud. Monog. 125. 1936. ype: Cameroons, Buea (Bula in Motyka, lapsus?) , on old trees, Deistel. Thallus pendent, flaccid, up to 40 cm. long, tawny olive to pinkish buff, sum- mits deep olive buff; base about 5 mm. long, 0.6 mm. in diameter, narrowly blackened below, repeatedly dichotomous in the lower 2 cm., then less so until the capillaceous 7-8 cm., where the internodes are again short; articulate, flexuous, joints 0.3-20 mm. long, mostly about 10 mm. and up to 2 mm. in diameter throughout, tapering only slightly apically and narrowing abruptly to 0.5 mm. or less at the capillaceous summits, constricted at the ends; surface impressed or subfoveolate, pseudocyphellae rare, punctate or in short, very low ridges on the main axes, more abundant and elevated on the smaller secondary branches, farinose, white to pale yellow; ramuli rare, slender, up to 10 mm. long on the main axes, apparently from unequal dichotomies on the capillaceous summits. Apothecia lateral, 5 (-10) mm. in diameter, cupulate, becoming plane; exciple smooth or with a few periclinal cracks near the margin; marginal cilia few, up to 20 mm. long, slender, subtuberculate with pseudocyphellae; disc chalky white, pruinose, becoming carneous as the pruina weathers away. RA LEONE: without locality or collector at Kew. Evidently a very old specimen, Wee brown w ith tawny olive summits, sterile with somewhat shorter joints. In habit it resembles E Motyks from the Canary Islands, which is said to remain Zorten? the herba otyka cites a specimen from the Canary Islands collected by Humboldt sub Licben hirtus v. Pini which must have been over a century old. CAMEROONS: Mt. Cam , 1610 m., SE above Buea, 1290 m. and 1450 m., F. W. H. Migeod 53, 144 Die at Kew 21. UsNEa speciosa Motyka, Lich. Gen. Usnea Stud. Monog. 124. 1936. Type: Ilha São Thomé, 1300-2100 m., A. F. Möller. Thallus pendent, rather rigid, up to 30 cm. long, drying pale ashy fuscous (Motyka, in our material VanDyke brown to Sayal brown [Fernando Po], tawny olive [Cameroons]); base 3 mm. long, 0.5 mm. in diameter below, expanding upward; dichotomously or sympodially branched once or twice with short inter- nodes, then rarely dichotomous; articulate, joints up to 7 cm. long, 2-2.5 m. in diameter (only 1 mm. in diameter in specimen from 2900 m.), curved to falcate, sharply constricted at the ends, but cortex often continuous; terete, cortex with short, longitudinal or somewhat oblique cracks, farinose, mostly pale sulfur yellow, foveolate, with small punctate pseudocyphellae, not elevated; ramuli occasional on the secondary branches, very rare on the main axes (perhaps only an undeveloped secondary branch); summits abruptly capillaceous, branches 0.4—0.5 mm. in diameter, tapering distally, rather densely dichotomously branched. Melliss 21, from St. Helena, is apparently the basal portion of an old specimen, about 10 cm. long, hazel to russet with elevated farinose soredia. After years in [I Vor. 44 16 ANNALS OF THE MISSOURI BOTANICAL GARDEN the herbarium it has stained the paper to which it is glued, testaceous. SIERRA LEONE: Loma Mts., Galerie Forest, 1800 m., on Tridactyla, P. Jaeger 4283, at wW. CAMEROONS: Buea, Musaka swamp, 1610 m., T. M. 696; Mee e Meyer Crater, 2575 m., T. D. M > 7; both at Kew, ex Hb. gi e Vico ria, Camero FERNANDO Po: 640 m., 1610 m. and 2900 m., Gustavo Mann; all at Kew. ST. HELENA: without locality, J; C. Melliss 2 21, at Kew 22. USNEA VESICULATA Motyka, Lich. Gen. Usnea Stud. Monog. 114. 1936. Type: Tanganyika, R. Uusoro [Wusiro?], in swamp, 2600—3000 m., Stuhl- mann. Thallus pendent, up to 60 cm. long, flaccid, main axes drying buckthorn brown to yellow ochre, capillaceous summits between chamois and deep olive buff; hold- fast of rooting fibers, repeatedly dichotomous at the top, then sparingly dichot- omous on the main axes, thinning at each dichotomy and again closely dichotomous at the base of the capillaceous summits; articulate, joints inflated, 2 mm. in diameter, 10—20 mm. long below, shorter on the main axes above, the capillaceous summits nearly cylindric, but the cortex annularly cracked; surface smooth to impressed or foveate on the larger joints, pseudocyphellae punctate or oblong, farinose, very inconspicuous; eramulose on the main axes, sparingly and irregularly ramulose on the summits, ramuli slender, up to 2 mm. long, often black-tipped. Apothecia infrequent, not seen in our material, lateral on slender branches, rarely subterminal, 2-4 mm. in diameter, thin, plane; exciple lacunose, marginal cilia few but comparatively long, capillaceous; disc plane, flesh color, slightly pruinose, thecium 60 y tall; asci narrow; ascospores uniseriate, 10 X 11 p, with a distinct epispore. The material referred here is not uniform. Sometimes a collection, or plants separated from the same collector's number, are much darker, russet to Dresden brown, and the joints of the main axes only about 1 mm. in diameter, surface less impressed and foveate, pseudocyphellae in very low verrucae (visible with 12X magnification). More field work will be necessary to decide whether they may belong to a different species. In the following list of specimens, they are marked with an asterisk (*) preceding KENYA: Longondoti, crater forest, 2575 m., arboricole, R. A. Dummer 5308, at Kew. UGANDA; Bugishu, Butandiga, 2415 m., A. E Thomas 484; Elgon Madangi, 3550 m., A. S. Thomas i Kigezi, * (ee deg Kabale K ver, 2250 m., I. R. Dale L43; Mt. Mga Kew 3380 m., A. S. Thomas 2457 (both farita) Mafuga, 2575 m., I. R. Dale L42; Si at Kee o: Mt. Kahusi, 2700 m., F. L. Hendrickx 4300, 4305, 4316; km. 30, route Kahusi, E L. Hendrickx 4330, a fragment; all in E. African Herb. ANGOLA: Cuanza Sul, Amboina, Capir near the Carloaongo River, J. Gossweiler 9991 p. b. min., juvenile, at Kew NGANYIKA: Kilara near Bismarck Hill, m., in evergreen mist forest, Set paeas 3816 (both forms); * Dabaga, Dr. SE 1947, 6.1X.32; at Kew 2m LAND: Nyika Plateau, ramicole, 2340 m., L. J. Brass 17215, in N. Y. Bot Gard. er 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 17 23. USNEA DEMINUTA Motyka, Webbia 8:389. 1952. Type: Ethiopia, Alto Semién, Beleghes, Reg. Sancaber, 3300 m., Pichi-Sermolli 7 Thallus flaccid, pendent, 10-15 cm. long in type (more than 30 cm. long in our material), drying buckthorn brown on main axes, to yellow ochre or clay color on smaller branches; base absent in our plants, main axes 0.5—0.6 (7-1) mm. in diameter, tapering upward; terete, articulate, many of the larger joints im- pressed to foveolate or even perforate, epapillate, up to 10 mm. long, mostly shorter, constricted at the ends, cracks narrow but showing the medulla or axis, the slender branches usually continuous, only slightly articulate, irregularly dichoto- mous; ramuli rare on the larger joints, occasional on the more slender branches, perhaps only undeveloped branches from an unequal dichotomy, and rare spinules which break away leaving minute scars and probably serve as isidia; pseudocyphel- lae minute on the larger joints, also short longitudinal or oblique cracks in the cortex through which medullary hyphae protrude. Tangled with Mearns 1776 is a fertile terminal fragment which seems to be this species. The axis is orange cinnamon shading to light vinaceous cinnamon at the tips, and is regularly annularly cracked, exposing the white medulla; the joints cylindric, rarely very slightly inflated and faintly longitudinally rugose with rare pseudocyphellae; the smaller branches show more pseudocyphellae and longitudinal cracks. e apothecia are 2—4 mm. in diameter, cupulate, sessile; exciple smooth except for several very deep, rounded foveolae on either side of the concrescent branch; marginal cilia distant, of variable length, the shorter nearly erect, the longer spreading and curved, about 5 mm. long, annularly cracked at the base with a few very minute pseudocyphellae; disc very concave, white-pruinose Our specimens have been referred here although the dimensions are slightly larger and I have seen no material from Ethiopia. They have been found tangled with other much larger species of the section ARTICULATAE, but do not resemble the juvenile forms of those species. KENYA: west slope of Mt. Kenya, trail from West Kenya Forest Station to summit, in bamboo zone, ca. 3000 m., Edgar A. Mearns 1776 p. p. (T. Roosevelt Exp.), at Farlow po E Imatong Mts., Mt. Kineti, 2810 m., ramulicole, A. S. Tbomas 1873, at Kew. TANGANYIKA: S. Pare Mts., Mtonto, 2100 m., on bushes of Philippia, P. J. Greenway 6532a, in E. African Herb. 24. USNEA ARTICULATA (L.) Hoffm. subsp. AETHIOPICA Dodge, Ann. Mo. Bot. Gard. 43:394. 1956. Type: Ethiopia, below Debra Erki, on fir, Schimper 18, 21 Oct. 1850, det. U. articulata v. intestiniformis Ach. by Müller Argau. Thallus pendent, more than 25 cm. long, probably flaccid (now glued tightly to herbarium sheet), snuff brown; base about 1 mm. in diameter, 8 mm. long, an- nulate, not blackened, then about twice closely dichotomous and occasionally so [Vor. 44 18 ANNALS OF THE MISSOURI BOTANICAL GARDEN throughout; articulate, joints irregular in size in the same axis, straight or curved and cucumiform, the larger joints mostly basal but not always so, 25-30 mm. long, 4—5 mm. in diameter (crushed flat in drying and hence appearing up to 7 mm.), the smaller joints 5-10 mm. long, about 1-1.4 mm. in diameter, usually straight, occasional branches long and slender, scarcely articulate and occasionally quite closely ramulose; ramuli slender, about 5 mm. long; the larger joints often foveate to scrobiculate, with small slightly elevated pseudocyphellae, sometimes in rows, the smaller joints smooth with minute pseudocyphellae; the smallest almost non-articulate branches with small soralia of isidiose soredia. Superficially our pressed plant, glued tightly to the herbarium sheet, resembles subsp. intestiniformis (Ach.) Crombie, in habit and color. The latter grows on sandy soil in England, France, Belgium, and the Netherlands. The main axes are more uniform in diameter and larger with relatively shorter joints, pseudocyphel- lae quite rare, and no soralia have been observed. Our plant is apparently more closely related to subsp. Kotschyana Motyka from Syria, 40 cm. long, with joints smaller and more uniform in size, coarsely papillate and sorediose. I am unable to estimate the size of our plant as the upper portion of the main axes has been broken away; the portion present measures 25 cm. long. ETHIOPIA: below Debra Erki, on fir, Schimper 18, 21 Oct. 1850, det. as U. articulata y. pee des (Ach.) Crombie, at Kew KENYA: Gilgil, on trees, G. M. Allen 1 780, i in Howe herb. at Farlow Herb. 25. USNEA TERRESTRIS Motyka, Lich. Gen. Usnea Stud. Monog. 75. 1936. Type: nia Mt. Kilimanjaro, 4000 m., on rocks in grassland, above tree-line, H. Meyer. Thallus pred or pendent, pale gray green when fresh, drying tawny olive on larger branches to warm buff on the capillaceous summits, up to 20 cm. long, base very dark brown, about 1-3 mm. in diameter and 1 mm. tall, fraying out below into many root-like fibers enclosing soil particles and expanding above to 2 mm. in diameter at the first fork; dichotomous, constricted at the axil, articulate, joints 7-20 mm. long, constricted at the ends, lower joints inflated to 2 mm. in the middle, the upper ones to 1 mm., separate or cortex continuous, tapering with closer ramuli and repeatedly dichotomous branching to the capillaceous summits, tips of ultimate branches acute, slightly blackened; surface smooth to impressed and subfoveate, becoming nearly terete at the summits; ramuli rare on the larger branches, perpendicular, straight or slightly curved, articulate at the base but not inflated, tips acute, longer, and more slender, abundant on the summits but never crowded, mostly 5 mm. or less long, a few much longer, subtuberculate and occa- sionally pseudocyphellate; pseudocyphellae white, very small and inconspicuous, round or elongate; soredia clearly seen only on James’ specimen from Kenya, abundant on smaller branches, from rather large, farinose, pale sulfur soralia, hemispheric, rarely crateriform when eroded, very rarely substipitate. 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 19 The above description is based mostly on Turrall 71 from the type locality, as the specimen agrees well with Motyka's description although a little larger and with inconspicuous pseudocyphellae, easily overlooked, and with soralia. Burtt 4071 from Mt. Meru is a tangled mass of many plants with fragments of dead mosses glued to the herbarium sheet. The one plant I succeeded in disentangling appears smaller and younger, only 11 cm. long, but is morphologically similar. Burtt's specimen from Ruanda is a fragment, slightly darker, mostly terete, rarely impressed, with very rare pseudocyphellae. Apparently broken smaller branches and the ultimate branches are capable of forming a new holdfast and continuing to develop new secondary axes. Burtt's specimen seems to be a fragment of one of these secondary axes. James" specimen from Kenya is a mass of plants glued to the herbarium sheet, evidently terrestrial, as the mass contained a small pleurocarpous moss, bits of dead leaves, and very rotten twigs, dead grasses and scraps of Parmelia. The plants are tawny olive to Sayal brown, the capillaceous summits cinnamon buff to chamois, the joints on some of the principal axes 20 mm. long, 6.5 mm. in diameter (up to 10 mm. when pressed flat), thus suggesting U. articulata subsp. aethiopica Dodge, but these large joints are shorter and foveate and occasionally with a small soralium, or a short farinose ridge; soralia abundant and well developed on the smaller branches. KENYA: Aberdare Mts., Sir Evan James, 1905, at Kew; w. slope of Mt. Kenya, 4200 m., Edgar A. tb 1654 (T. Roosevelt Exp), in Mo. Bot. Gard. Herb., det. Motyka, and at Farlow H jaro, Bismarck-Peters, 3420 m P G. Ferrall 71, at Kew and E. Aeon Herb.; ad Waldrand, 3000 m., Dr. Geilinger, at Kew; Arusha District, Mt. Meru, 4680 m ., B. D. Burtt 4071; at Kew; all on i EA rocks in SE meadows above tree line. coNco: Ruanda, Mufumbiro Mts., Mt. Mikeno, among grasses, B. D. Burtt, at Kew. 26. USNEA RUGOSA Motyka, Lich. Gen. Usnea Stud. Monog. 115. 1956. Type: Ethiopia, Bagla, 2700 m., Hildebrandt. Thallus pendent, flaccid, up to 65 cm. long, olive buff on the main axes, pink- ish cinnamon on slender branches in recent specimens (2 years old), finally light brownish olive to Isabella color on the main axes, lighter on slender branches in older specimens (22 years old); holdfast inconspicuous, base up to 15 mm. long, 0.6-1 mm. in diameter, infrequently dichotomous throughout; articulate, joints mostly 2 mm. in diameter on main axes with an occasional joint up to 4 mm., varying in length from 2 to 10 mm.; surface variable, some joints smooth, terete, others with inconspicuous pseudocyphellae and low, narrow, farinose, more or less longitudinal ridges, the larger joints somewhat compressed and foveate with dense patches of dichotomous ramuli which may grow out to form the long secondary, capillaceous branches, up to 15 cm. long, surface asperate with punctiform soralia, somewhat larger and isidiose soralia on some of the basal joints; pseudoramuli aris- ing from unequal dichotomies in the slender portion of the thallus. [Vor. 44 20 ANNALS OF THE MISSOURI BOTANICAL GARDEN Only lateral apothecial initials seen in our material. Motyka describes the apothecia as lateral, about 5 mm. in diameter, exciple very thin, plane, smooth; marginal cilia small, variable in length, flexuous, articulate; disc nearly plane, rugose; thecium 60 y tall; asci ventricose; ascospores 7 X 9 p, with a thin epispore. Hildebrand 307 at Kew seems to be this species; the color is uniformly clay color, mostly with a white pruina from excess of mercuric chloride in poisoning; the main axes are rather more slender and smoother, sterile. It may be a duplicate of the type or perhaps of the specimen cited by Motyka as U. articulata subsp. mediterranea Mot. which, however, remains pale stramineous or very pale olive fuscous in the herbarium. There is a discrepancy in the elevation cited for the type; in the citation before the description, Motyka gives 2700 m., while in the citation of specimens seen at the end he gives 2420 m. The specimen at Kew is given as 7500 ft. (2420 m.). "e Galla Pass, 9° 28’ N., 42° 19' E., J. B. Gillett 5130, Abyssinia-Somaliland Bound. Comm.; Simien, Mindigabsa, H. Scott 323; Bagla, 2420 m., J. M. Hildebrandt 307; all : X ei 27. UsNEA VaiNIOANA! Zahlbr., Cat. Lich. Univ. 6:597. 1930. Usnea fragilis Vainio, Hedwigia 37:40. 1894, non Stirton. 1882. Type: Uganda, Mt. Ruwenzori, 0? 5' S., Scott-Elliott. Thallus up to 30 cm. long, pendent, very flaccid, rather fragile when dry, between colonial buff and olive buff, one group of plants cream buff after drying; base 5 cm. long, 0.6 mm. in diameter, annularly cracked; holdfast conic, 1 mm. in diameter, black; main axes 0.6 mm. in diameter, tapering gradually at each di- chotomy to long capillaceous summits; branching infrequently dichotomous, the larger branches articulate, joints up to 5 mm. long, not or very slightly inflated below, not constricted at the ends, terete or very slightly impressed to subfoveo- late, smooth, eramulose (less-developed branches may simulate ramuli); summits capillaceous, flexuous. Apothecia and soredia not seen in certain material. Turrall 55 in the E. African Herbarium and a fragmentary plant untangled from another species in specimens from Kew seem to be this species. The main axes are up to 1 mm. in diameter, the joints somewhat more inflated, some slightly more foveolate or very slightly rugose; medulla K—. Apothecia very rare, lateral on smaller branches, 0.3 mm. in diameter, exciple subimpressed, very slightly pruinose; disc plane, 2 mm. in diameter, pale carneous, densely white-pruinose with 6-8 marginal cilia of variable length. This plant seems to be intermediate between U. Vainioana Zahlbr. and U. exasperata (Müll. Arg.) Vainio. ETHIOPIA: Simien, Mindigabsa, H. Scott 32 y Geech, H. Scott 250; both at TANGANYIKA: Mt. Kilimanjaro, 2900-3030 m., on giant heather, R. G. soins 55, in E. Africa T Aa ent at Kew; S. Pare Mts. — Mtonto, 2100 m., P. J. Greenway 6532, in E. ipte NYASALAND: e anje District, Luchenya Plateau, Mt. Mlanje, 1890 m., ramicole, L. J. Brass 16846 (Vernay eat Exp.), in N. Y. Bot. Gard. Herb. Original spelling Wainioana corrected by Motyka (Lich. Gen. Usnea Stud. Monog. 437. 1937) to Ber to Vainio's preference for spelling his name after the Finnish revolution. 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 21 28. USNEA GRACILIS Ach., Lichenogr. Univ. 627. 1810. Type: Ile Réunion (Bourbon), com. Weber & Mohr. Thallus pendent, up to 30 cm. long, flaccid, gray to pale green, becoming deep olive buff in the herbarium; the main branches often Natal brown or darker in the lower portion; base hemispheric, attached to the bark; immediately dichotomous and repeatedly so throughout, often unequally so when one branch resembles a long, slender curved ramulus; lower portion of the main axes about 0.7 mm. in diameter, gradually tapering to the capillaceous summits; terete, continuous or annulate with very minute cracks, smooth, with rather indistinct paler lines; soredia unknown. Apothecia very rare, sessile on smaller branches, up to 4 mm. in diameter, cupu- late to subplane; exciple smooth; marginal cilia distant, 2-10 mm. long, annularly cracked at the base, tapering to acute tips; disc flesh color, faintly white-pruinose. Cortex 90 p thick, fastigiate, highly gelified; algal layer about 40 p thick, cells protococcoid, 6-7 y in diameter, dying in the older portions of the thallus and leaving cavities; medulla 225 p thick, of densely woven hyphae 3—4 y in diameter, lumen 1 p; axis 275 p in diameter, of conglutinate, thick-walled longitudinal hyphae. As the duplicate of A. G. Allan from Kenya in Herb. Boule de Lesdain is cited by Motyka, probably his statement "medulla tantum ca. 20 p crassa” is a misprint for 200 y. ETHIOPIA: Pichon: sili J. Omer-Cooper, at Kew KENYA: Ngongo Bagas, A. G. Allan, com. Rev. W. Lillie, in Howe herb. at Farlow gë eege E. Usambara, eae Maramba, 1030 m., P. J. Greenway 4151. NGO: Mt. Kahusi, ca. 2700 m., on Arundinaria alpina, F. L. Hendrickx 4300, 4302, 431 " sacingly fertile, in E. African n Herb. : Mt. Mlangi, 1450-2250 m., P. J. Greenway 6332. RITIUS: without locality or eld: ex herb. Macreight in Tuckerman Herb., sub U. blicata at Farlow Herb. REUNION: com. Miiller Argau, at Farlow Herb. 29. UsNEA PULCHELLA Motyka, Lich. Gen, Usnea Stud. Monog. 442. 1937. Type: St. Helena, Diana Peak, Dumont d’Urville, at Uppsala, not seen. Thallus pendent, very flaccid, 25 cm. long (45 cm. in our specimens), drying bister to snuff brown on the main axes, pinkish cinnamon to cinnamon buff on slender branches and ramuli; holdfast a small black disk under the outer layers of bark cells, the base growing outward through a crack, 10 mm. long, 1 mm. in diameter, curving downward, annularly cracked, the part within the bark black; repeatedly dichotomous throughout, internodes somewhat shorter near the base and in the capillaceous summits; larger branches articulate, joints of variable length and diameter, some slightly inflated, up to 1.3 mm. in diameter, mostly about 1 mm. in diameter, others on the same branch shorter, cylindric, about 0.8 mm. in diameter, smooth, or on the slenderer, less articulate branches, very mi- nutely white tuberculate (or pseudocyphellate?); the larger branches almost [Vor. 44 22 ANNALS OF THE MISSOURI BOTANICAL GARDEN eramulose, the slender branches quite irregularly ramulose; ramuli 10—40 mm. long, the shorter slender, straight or curved and smooth, the longer somewhat tuberculate and occasionally once or twice dichotomous (perhaps arrested develop- ment of a branch), usually flexuous; sometimes several growing out at a "node" forming a tuft (as in U. nidifica) in the type, absent in our specimens. It seems quite likely that U. monumenti Fée,! type from St. Helena, trees about Napoleon's tomb, without collector, is the same species, but I hesitate to reduce U. pulchella Motyka to synonymy without seeing the types. U. Lyngei Motyka, from the same locality as U. monumenti Fée, has minute white farinose tubercles and ramuli not in dense clusters. U. umbrata Motyka, also from St. Helena, is very rigid and less articulate. Our plants are longer and somewhat more slender than the type (fide Motyka) and have minute white tubercles which probably develop a few isidiose soredia. Perhaps the tufts of ramuli mentioned by Motyka are such isidiose soredia which have developed in situ. Otherwise our plants agree well with Motyka’s description. FERNANDO PO: Santa Isabel Peak, 2900 m., Gustavo Mann 682, probably arboricole, 2 sheets at Kew. 30. UsNEA MONILIFoRMIs Motyka, Lich. Gen. Usnea Stud. Monog. 447. 1937. Type: S. Africa, Natal, hills above Mauritzburg, Anton Rebmann. Thallus pendent, up to 30 cm. long, flaccid, lower portions of main branches sepia shading upward to cream buff at the summits, somewhat shining; holdfast a disc 2 mm. in diameter, base blackened, about 2 mm. long, 0.5—0.7 mm. in di- ameter; several times dichotomous near the base, some of the secondary branches forming holdfasts, then less closely dichotomous throughout; eramulose (occasion- ally one branch develops more slowly and resembles a ramulus); main branches 0.5—0.7 (—1) mm. in diameter, thinning at each dichotomy, summits capillaceous; terete, closely annulate, joints 1-3 (—10) mm. long, not constricted at the ends, cracks narrow, white-margined; very smooth, variegated with white points and lines (pseudocyphellae ?), very rarely the lines are slightly elevated. cially, this species somewhat resembles U. gracilis with somewhat the same geographic distribution at low elevations, but with a thicker, laxer medulla and darker fuscous color. If the white points and lines are considered as pseudo- cyphellae, it would suggest inclusion in the ARTICULATAE instead of the sTRAMI- NEAE-AMOENAE and a relationship to U. deminuta Motyka from Ethiopia at much higher elevations. KENYA: Ngong hills, 2370 m., P. R. O. Bally B94; Marsabit Mt., A. T. A. Ritchie, com. E A e. Peu B1416; at Kew aro, near e Bismarck Hill, 3220 m., in ien mist forest, P. J. Greenway 3816 Marangu, Bismarck, 2140 m., R. G. Tarral 37, at Kew . Kahusi, 2700 m., on Ericaceae, F. L. Hendrickx 3440, in E. African 1 In Dictionnaire Classique (ed. Bory de St. Vincent) 16:481. 1830. D 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 23 31. USNEA STRAMINEA Müll. Arg. emend. Motyka, Lich. Gen. Usnea Stud. Monog. 466. 1937; Müll. Arg., Flora 62:162. 1879; p. p. excl. syn. Type: Mauritius, Robillard at Genéve, a portion in Farlow Herb., selected by Motyka as type. The New Zealand specimen from Nelson, com. Ferd. v. Müller, also cited by Müller Argau, is referred by Motyka to his var. substraminea as well as the types of the synonyms cited by Motyka. Thallus, pendent, more than 20 cm. long, somewhat flaccid, pale stramineous when fresh, main axes drying hazel or lighter, ramuli warm buff, shining; holdfast and base broken off in material available; several times closely but somewhat un- equally dichotomous near the base, then rarely so above, axils about 90^ angles, the branches curving downward, nearly parallel, about 1 mm. in diameter, somewhat flexuous, annularly cracked and subarticulate, terete, some joints not constricted at the ends, some with surface slightly foveate, and some sparsely low- tuberculate (probably initials of ramuli); secondary branches 0.3—0.5 mm. in iameter, about 10 cm. long, foveolate, cortex more or less longitudinally cracked, forming elongate soralia, shorter and nearly circular above, soredia granular; ramuli 10-20 (—30) mm. long, very slender, infundibuliform at the base, annularly cracked just above, low-tuberculate and small-soraliate, tapering to an acute tip; summits capillaceous, distantly short-ramulose. Cortex very fragile when dry, 115 p thick, fastigiate, highly gelified; algal layer 65 p thick, of discrete flattened colonies of Trebouxia 20-25 p in diameter, separated by strands of closely woven medullary hyphae; medulla 300—350 y thick, the outer 100 » very compact, of predominantly periclinal hyphae, the inner portion of very loosely woven radial hyphae, occasionally dichotomous, 4-5 p in diameter, very thick-walled, heavily nubilated with minute hyaline granules, aris- ing as branches of dark brown periclinal hyphae, 6-7 u in diameter; axis 360— 400 p in diameter, hyaline, of thick-walled longitudinal conglutinate hyphae. ^ The Socotra plant; referred bere by Müller Argau, has a main axis about 2 mm. in diameter, is more clearly articulate with A aot somewhat constricted at the ends and deeply foveate. MAURITIUS:, Robillard, portion of type com, Miller Argau, at Farlow Herb. SOCOTRA ISLAND: Schweinfurth, com. Mëller Argau, at Farlow Her 32. USNEA EXASPERATA (Mall: Arg.) Motyka; Lich. Gen. Usnea Stud. Monog. 113440, 31932, | Usnea jerata v. Ste Müll. Arg., Flora 73:336. 1890. ius dasypoga v. exasperata Zahlbr:, Cat. Lich. Univ. 6:556. 1930. Type: Td gany yika, Kilimanjaro, 3000 m. . [3145 m., fide Stein in Meyer, Across É African Glaciers, 353. 1891], H. Meyer i in, ‘Genéve; portion of same collection in Berlin, typical; another. i in Breslau, up to 60 cm. long, branches papillose and verruculose, probably an unnamed species. [Vor. 44 24 ANNALS OF THE MISSOURI BOTANICAL GARDEN Thallus 30—35 cm. long, pendent, flaccid, pale yellow when fresh, fawn to wood brown below, shading to vinaceous buff above, the ramuli and long capilla- ceous summits colonial buff to olive buff; holdfast a black verrucose disc giving rise to several axes, base 3—4 (—10) mm. long, about 0.8 mm. in diameter below, increasing to about 1.2 mm. in diameter just below the first branch, the lower t black, annulate; closely and somewhat unequally dichotomous in the lower 4 cm., with very short internodes, then very rarely dichotomous in the next 10—12 cm., about 1 mm. in diameter, tapering upward, then once or twice dichotomous at the bases of the very long, capillaceous summits; the larger branches subfoveo- late, closely articulate, joints not or only very slightly inflated, axils somewhat complanate, very rarely ramulose; the upper portions of the main branches slightly obliquely rugose, the joints longer and the axils not complanate, the cortex some- times subareolately cracked, the cracks white-margined, not verrucose nor papillate; ramuli 10-30 mm. long on the larger branches, perpendicular to the axis but vari- ously curved and twisted, only 4—5 mm. long on the slender capillaceous portion of the thallus, with small low tubercules which probably develop minute isidiose soredia on most of the ramuli, although only a few have developed so far in our plants. Medulla K— or slowly yellow Apothecia (in Hendrickx 4330) 2-3 mm. in diameter, lateral on slender branches or large ramuli; exciple nearly plane, smooth or nearly so; marginal cilia subdistant, up to 10 mm. long, curved, annularly cracked at the base and sparingly so above, slender; disc carneous, white-pruinose. Turrall 37 is sterile and quite sorediose, but apparently belongs here. The material from Kikuyu, Kenya, is only a fragment, the joints somewhat longer and more inflated, the tubercles on the smaller branches forming slightly elevated isidi- ose soralia. Some of the broken branches have regenerated dense clusters of short ramuli, about 3 mm. long. ENYA: Kikuyu, gir — no. 35 Elgon, Miss R. J. Busb, in E. Africa Herb. 10, os, a YED NG 203, at TANGANYIKA: Kiliman se DAMM SE 2140 m., R. G. Turrall 37; Bismarck Hill, 3220 m. L J Greener 3816, a CONGO: route Kahusi, 2450 m., on a, F. L. Hendrickx 3679; Mt. Kahusi, ca. 2700 m., F. E Baies iw: 4305, od, E 4316, 4330; all in E. African Herb. 33. UsNra PINKERTONI Stirton, Scottish Nat. 6:294. 1882. Type: Egypt, Petrified Forest near Grand Cairo, Pinkerton, in the Art Galleries of the Glasgow Corporation. Thallus 18 cm. long in type, over 25 cm. in our plants, pendent, flaccid, buffy brown on main axes to citrine drab on smaller branches; base blackened and rather hard in type, absent in our specimens, infrequently dichotomously or sympodially branched on the main axes, more closely so on the smaller branches; main axes up to 1 mm. in diameter, closely articulate, joints very slightly inflated, slightly rugu- lose and subfoveolate, some very minutely and acutely papillate especially in the 1957] DODGE—LICHENS OF TROPICAL AFRICA. IIl. USNEA 25 mid-portion of the thallus, others epapillate, backs of ridges often white-punctate (pseudocyphellae ?) ; ramuli irregularly disposed, sometimes quite close, sometimes absent for some distance on the main axes and secondary branches, usually close on the smaller branches, very slender, mostly 2-3 mm. long, some up to 7 mm.; summits capillaceous, very densely branched. Apothecia lateral on secondary branches, up to 10 mm. in diameter, cupulate becoming nearly plane; exciple smooth in smaller apothecia, slightly rugulose and subpapillate on the larger ones; marginal cilia quite close, rather long, the longer often forked; disc, dark carneous, white-pruinose at first, becoming nearly nude in age. Although placed by Motyka in the BARBATAE, subsect. PENDULINAE, the olive fuscous color, the very thin cortex and thick lax medulla, the subarticulate thallus with rugose to subfoveolate joints all point to a close relationship to the sTRAMI- NEAE subsect. SULCATAE, although papillae are very rare in the sTRAMINEAE except in U. straminea Müll. Arg. Bally 5611 seems to belong here. The specimen con- sists of the upper portion of a main axis and six secondary branches. No papillate joints were seen. GYPT: Safe Forest near Grand Cairo, Pinkerton, in Art Galleries of the Glasgow Conon ETHIOPIA: Galla Pass, 9° 28’ N., 42° 19’ E., 2520 m., J. B. Gillett 5130, Gare ppa Bound. Comm.; Simien Mindi gabsa, H. Scott 3235, sterile; both a : Northern Frontier Dresses 1560—1610 m., on Cineraria eat, P. RO, Bally E at Kew. 34. UsNEA CAMEROONENSIS Dodge, Ann. Mo. Bot. Gard. 43:394. 1956. Type: Cameroons, Miss Cheeseman L3, at Kew. Thallus pendent, rather flaccid, more than 55 cm. long, main axes drying snuff brown, Sayal brown to cinnamon, ramuli deep olive buff, base broken away; articu- late, joints about 1.5 mm. in diameter below, mostly 1 mm. above, thinning grad- ually; closely several times dichotomous below, then unbranched for 10-20 cm., occasionally dichotomous above; axils subcomplanate, joints variable in length, many about 10 mm. long, constricted at the ends, usually showing the axis in the cracks, not or only very slightly inflated, longitudinally rugose, the ridges low, cracked at the top and farinose, flattened but not sulcate between; ramuli irregu- larly disposed throughout, up to 30 mm. long on the main axes, base subarticulate and inflated, tapering to a slender acute tip, about 10 mm. long on the secondary branches, surface low-tuberculate and deformed, forming isidiose soredia which soon break away causing the tops of the tubercles to appear farinose, somewhat in rows in the lower portions. Apothecia not seen. Cortex 55 p thick, fastigiate in a gel, outer 8-10 p nubilated with minute brownish crystals; algal layer 30 (—40) un thick, of densely packed protococcoid cells 6—8 u in diameter; medulla 185-195 un thick, of loosely woven thick-walled hyphae 3 p in diameter, somewhat more compact next the algal layer; axis 250 p [Vor. 44 26 ANNALS OF THE MISSOURI BOTANICAL GARDEN in diameter, of conglutinate, thick-walled hyphae 5—6 p in diameter, lumen 3 p and in the central portion a few large brownish hyphae 6—8 p in diameter, nubilated with minute brownish crystals. , CAMEROONS: Miss Cheeseman L3, at Kew. 35. Usnea Fusca Motyka, Lich. Gen. Usnea Stud. Monog. 473. 1937. Type: S. Africa, Transvaal, Lydenburg, Wilms. Thallus pendent, up to 20 cm. long (30 cm. in our material), green becoming buffy brown on the main axes, shading to honey yellow or chamois on the smaller branches; holdfast thin, circular, 273 mm. in diameter, darker but not black, base 1-2 mm. long, about 0.5 mm. in diameter, then dichotomous at right angles and the branches pendent from opposite sides of the branch of the tree, rarely dichoto- mous in the next 4—10 cm., then more frequently and more closely so near the summits, flexuous to nearly straight; main axes up to 1.5 mm. in diameter, articu- late, joints variable in length, cylindric or very slightly inflated, not or only slightly constricted at the ends, nearly terete but somewhat deformed to obtusely angled, rugulose, the ridges oblong and mostly oblique, backs ecorticate and farinose, cortex sometimes cracked and subareolate; axils of larger branches flat- tened; ramuli almost absent below, more frequent above but perhaps only members of unequal dichotomies; summits long-capillaceous... Apothecia not seen; . Soredia isidiose, becoming farinose from tubercles of the main axes, very ge Age not elevated on the smaller branches. In the northern portion of its range, whence our. material came, the species assumes more the habit of U. exasperata; cf. specimen from Kilimanjaro, Ublig, cited by Motyka, loc. cit. 474. KENYA: Chyulu Hills, H. D. van RER (2 sion in E RE Herb. y. UGANDA: Kigezi, Mafuga, 2580 m., I. R. Dale L41, fra gnon; Toro, Fort Portal, 1625 m., I. R. Dale L40, fragments tangled E: other Spp. Lot ac at TANGANYIKA: S. Pare Mts., Mt: Mtonto; 2100 mi; P. J. Seat 6532; a conco: Mt. Kahusi, 2700 om. Fs, L. indria 4314 inaina: name: Geer mel in E. geg Herb.. é 36. USNEA DECIPIENS T Lich. Gen. Usnea Stud. Tm 7 19577 Type: piter Mt. Meimbwa, : near Lake Tanganyika, 23 00 m., ‘corticole, Miinzner. SS Thallus pendent, somewhat ced over 40. cm, , long, Jebiaceaus-Euscous, (snuff ies on larger branches, Isabella color on small branches i in our material), smooth, almost shining; base indistinct, 3 mm. long, L mm. in diameter, expanding . to 2 mm. at first branch, not blackened i in our material; main branches 2 mm, in di- ameter below, tapering. very slowly, rarely and irregularly, articulate, only. heich constricted, fractures, indistinct, acutely angled or Subterete. with distinct. longi- tudinal (rarely somewhat oblique) ridges, backs pale, very. deeply foveate or impressed, with very rare pseudocyphellae between; ramuli close, large and small 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 27 intermingled, the smaller 5 mm. long, the larger almost as large as the branches, thick at the base, tapering above to obtuse tips, continuous or rarely slightly articulate, often arcuate and contorted ridges, closer and less regular; summits in- distinct, of longer and denser ramuli. Our Uganda material is rather fragmentary, tangled with U. vesiculata, both species glued to the herbarium sheet. The above description is taken in part from Motyka’s monograph. Our material differs in having the main axes almost eramu- lose, the ramuli irregular on the secondary branches, sometimes quite close, in others rather scattered. Some of the ramuli have tubercles developing hemispheric isidiose soralia which appear farinose after the isidia are shed. UGANDA: E Madangi, 3550 m., on heath trees in alpine meadow, A. S. Thomas 607 p. b., KENYA: Ets ta nya, Kinangop, Wieder 2840—2900 m., on trees, P. Chandler 2278, ex hb. Botanist Dept. Agr. Uganda, at 36a. Var. RHoDEsiANA Dodge, Ann. Mo. Bot. Gard. 43:394. 1956. Type: Southern Rhodesia, Matopos District, top of high granite hill, 1610 m., on tree, Frederick Eyles 1023, at Kew. allus pendent, subflaccid, over 30 cm. long (incomplete), Isabella color, ramuli somewhat lighter; base not seen; infrequently dichotomous, not articulate nor annulate, longitudinally and subreticulately rugose, ridges subacute, cortex often cracking along the ridges, subareolate and somewhat scaling, deeply foveolate or sulcate between the 6—10 ridges per circumference of transverse section, about 1 mm. in diameter, thinning to about 0.5 mm. in the upper 15 cm., but summits not capillaceous; ramuli close throughout, 3-15 mm. long, the longer perhaps arrested branches as they are sparsely and minutely ramulose or bear low tuberculi- orm, farinose soredia. Cortex 40 y thick, fastigiate, pseudoparenchyma cells 5—6 u in diameter, nubi- lated with brownish crystals in the outer 15 p; algal layer 30 p thick, nearly con- . tinuous except in cortical cracks, cells protococcoid; medulla 30-40 , thick, of densely woven mostly periclinal slender hyphae; axis 600—625 p in diameter, slightly angled, of uniformly longitudinal, conglutinate thick-walled hyphae. The cortex and medulla are proportionally much thinner than in U. decipiens and the ramuli closer and more sorediose. 37. UsNEA AEQUATORIANA Motyka, Lich. Gen. Usnea Stud. Monog. 476. 1937. Type: Uganda, Mt. Ruwenzori, 2900—3220 m., Scott-Elliott. Thallus pendent, up to 60 cm. long, flaccid, main axes bister to snuff brown, ramuli and capillaceous summits buffy olive or lighter; holdfast black, about 3 mm. in diameter, an interwoven mass of root-like fibers penetrating the bark; base articulate, about 0.8 mm. in diameter below, rapidly expanding to 1 mm. or more, about 2 mm. long, then several times dichotomous forming several main axes, then very rarely dichotomous, about 1 mm. in diameter, slowly tapering to the long [Vor. 44 28 ANNALS OF THE MISSOURI BOTANICAL GARDEN capillaceous summits; articulate, joints not or very slightly inflated, slightly con- stricted at the ends in the larger branches, cylindric in the smaller branches; larger branches slightly but distinctly angled, ridges longitudinal, few, backs white- pruinose, rarely only slightly sulcate or foveolate, smaller branches smooth, terete, mostly ramulose; ramuli 4—5 mm. long, curved, bases thick, tips acute, rather dense but with sections of the principal axes almost eramulose. Soredia not seen. Apothecia common, up to 10 mm. in diameter, mostly smaller, subterminal on lateral branches, exciple smooth except for 3-5 low, rounded, radiating ridges, a few of the larger apothecia subfoveate or impressed; marginal cilia dense, up to 10 mm. long, patent or recurved, thick, annulate; disc subconcave to almost plane or somewhat deformed, white-pruinose. ETHIOPIA: without locality, Schimper, ex Herb. Sbarbaro sub U. florida v. strigosa, at Du) doge Mines Mau forest, 2350 m., in Acacia grove at forest edge, P. R. O. Bally B4047; uplands, 2250 m., E. R. V., ex Corydon ua Kenya 5161, at Kew; Mt. Elgon, gard 8 Mrs. Cyril Lugard 490, at A: zez Kasatoro, Kabale ns i m., I. R. Dale L. 43; Imatong Mts., Mt. Kineti, A. s T bomas 1893, both at Kew 38. USNEA FLAccIDA (Müll. Arg.) Motyka, Lich. Gen. Usnea Stud. Monog. 475. 1937. Usnea angulata v. flaccida Müll. Arg., Flora 73:337. 1890. Type: Kenya, Leikipia at foot of Aberdare Mts., corticole, L. von Hoebnel 176. Thallus pendent, flaccid, ca. 13 cm. long (Motyka), up to 30 cm. long in our material, becoming buffy brown on main axes, ramuli somewhat lighter; base about 1 cm. long, 0.5 mm. in diameter; articulate, joints cylindric, with short ramuli, not blackened; then closely dichotomous, axils flattened, then less frequently di- chotomous throughout; flexuous, articulate, joints not inflated, up to 1 mm. in diameter below, thinning at each dichotomy but mostly about 0.5 mm. with long capillaceous summits; few elevated, acute ridges on the lower foveate joints, less ridged above, the upper joints almost cylindric, otherwise smooth; ramuli rather close, evenly distributed throughout, slender, 0.3-10 mm. long, somewhat curved. Apothecia absent in the type and in our material. Motyka reports a specimen from Kilimanjaro, 3300 m., Ublig, with apothecia up to 10 mm. in diameter, sub- lateral, thin, exciple plane or slightly concave, distinctly foveolate; marginal cilia few, slender, variable in length, flaccid; disc plane or slighly convex, often tuber- culate-deformed; thecium 50 y tall; ascospores 7 X 10 m, ellipsoid, in narrow asci. DAN: lmatong Mts., Mt. Kippia, pendent from twigs of trees in grassland, 2580 m., A. S. Thomas 1823A KENYA: Lakia plateau and Aberdare Range, district around Nyeri, hanging from twigs, W. Scoresby Routledge, at Kew UGANDA: Kigezi, Kasatoro, Kalie River, a m., on ge I. R. Dale L SE OUTH AFRICA: Adelaide, Miss D. K. Kingon, com. Rev. W. Lillié, in Howe Herb. sub U. florida v. perplexans (Stirton) Zahlbr. Nc M. Bouly de Ledái, at Farlow Herb. 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 29 39. USNEA TERRICOLA Dodge, Ann. Mo. Bot. Gard. 43:394. 1956. Type: Cameroons, Mamoquilles, 2450 m., terricole among mosses and Stereo- caulon sp., Miss Cbeeseman 25, at Kew. Thallus fruticose, subrigid, 8—10 cm. tall, cinnamon buff to buffy brown, irregularly black-maculate on the main axes, summits dark olive buff; base black, densely dichotomous at soil level, probably sending coarse root-like structures into the soil (broken off in collecting), rapidly expanding from 1 mm. to 2-3 mm. in diameter in the lower and middle joints, then tapering to 0.5 mm. in the ultimate joints; articulate, joints 5-20 (—30) mm. long, constricted at the ends, cracks varying from very narrow to 1 mm. wide, dichotomous throughout; basal joints nearly smooth with punctiform pseudocyphellae, and cortex cracked, forming low short ridges, sulfur-farinose, which may develop minute isidiose soredia; inter- mediate and upper joints more or less minutely verrucose with more abundant and elevated ridges, sometimes becoming reticulate and giving a foveate appearance, on the upper joint occasionally developing an apotheciform soralium covered with minute isidiose soredia, or part of the verrucae may form small elevated soralia. The soredia sometimes develop in situ producing dense clusters of pseudoramuli or becoming minute dichotomous thalli, joints only 0.3—0.5 (—1) mm. in diameter; summits obtuse. . Cortex 50 y thick, loosely fastigiate in a gel, outer 20 u nubilated with minute brown crystals; algal layer 35 p thick, a continuous layer of densely packed proto- coccoid cells 8 u in diameter; medulla 375 p or more thick, of thick-walled hyphae 5—6 y in diameter, very densely woven in the 50 p next the algal layer, rest very lax, of predominantly radial hyphae; axis 300 mw in diameter, of conglutinate, longitudinal thick-walled hyphae, 3—4 pu in diameter. CAMEROONS: Mamoquilles, 2450 m., terricole, Miss Cheeseman 25, at Kew. 40. UsNEA HIsPIDULA Motyka, Lich. Gen. Usnea Stud. Monogr. 488. 1937. Usnea barbata v. bispidula Müll. Arg., Bot Jahrb. [Engler] 20:245. 1894. Type: Tanganyika, Usambara, Holst. Thallus about 15 cm. long, subpendent, quite rigid, fuscidulous in the her- barium, larger branches ochraceous tawny, ramuli dark olive buff, smooth; base absent in type, of rooting fibers penetrating bark at least 2 mm., not expanded into a holdfast, not blackened; primary branches 1.5—1.8 mm. in diameter, secondary branches thinner, constricted at the base, then slightly inflated and tapering grad- ually to the tips, continuous, or annularly cracked where the secondary branches leave the primary, smooth, with cylindric papillae (probably arrested ramuli); ramuli slender, 1-2 (—4) mm. long, minutely tuberculate and soraliate, forming all gradations between spinules and small tertiary branches; summits capillaceous, about 1 cm. long, tips acute, smooth or nearly so; soralia small, isidiose at first [Vor. 44 30 ANNALS OF THE MISSOURI BOTANICAL GARDEN becoming farinose, abundant on the smaller branches and the longer ramuli, rather rare on the larger branches. UGANDA: Bugishu, Butandiga, 2130 m., arboricole, A. S. Thomas 2565, at K RHODESIA: East Isoka District, uplands, 1480 m., arboricole, C. G. Trapnell, a Firm lacking both base and summits, at Kew 41. USNEA USAMBARENSIS Motyka, Lich. Gen. Usnea Stud. Monog. 489. 1937. Type: Tanganyika, Usambara, Nguelo, Kummer. Thallus up to 15 cm. long, pendent but quite rigid for its size, pale green when fresh, drying between deep olive buff and colonial buff; holdfast convex, about 2 mm. in diameter, not blackened; closely branched just above the holdfast, then less frequently dichotomous, branches nearly perpendicular, internodes variable in ength, flexuous, about 1 mm. in diameter below to about 0.5 mm. above, summits mostly broken off, the few remaining about 2 cm. long, flexuous, tips obtuse, rather rigid, about 0.2 mm. in diameter; primary axis smooth, irregularly annulate and areolate, the areoles with raised margins; the middle portion rugose and low-angled, resembling U. africana; the upper portion, especially the smaller branches, low- tuberculate and deformed, the tubercles bearing a few rather coarse isidioid soredia, appearing farinose as the isidia break away, sometimes almost ecorticate in patches; ramuli subdistant, 2-3 mm. long, some longer, dichotomous, and resembling small lateral branches, flexuous, tapering from base to rather blunt tip, the shorter smooth, the larger minutely annulate and tuberculose, forming small soralia an eventually appearing decorticate Our plants from the Sudan are small, only about 5 cm. tall and only beginning to form soredia, probably juvenile. The plants from Madagascar are also small but quite punctate-sorediose above. SUDAN: sade -— 1710 m., 32° 50’ E., 3° 50' N., T. F. Chipp 57, at Im a. Andcsiolonks Hildebrandt, ex Herb. Sbarbaro dog: U. dasypo- goides v. Sorel Mall. d ., at Farlow H 42. USNEA CARTILAGINEA Laurer in Motyka, Lich. Gen. Usnea Stud. Monog. 493. 1937. Type: Cape of Good Hope, Table Mt., J. F. Drége. Thallus 7-10 cm long, subpendent, quite flaccid, pale green, drying fuscous (Sayal brown to cinnamon), slender branches lighter; basal disc 2 mm. in diameter, base 1 mm. in diameter below, expanding rapidly to 2 mm. at the first branch, branching Page branches articulate, joints inflated in the middle and con- stricted at the ends, not tapering upward, terete, smooth, with scattered minute tubercles below, closer above where they produce single small isidia; ramuli irregu- larly disposed and rare on the lower joints, more regularly subdistant above, straight, articulate, constricted at the base, inflated then tapering to a slender acute tip, indistinctly tuberculate, paler than the axes; summits more slender, less 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 31 conspicuously articulate, tips obtuse, more closely tuberculate, each tubercle pro- ducing an isidium, very rarely more. en the isidia are shed the remaining scars resemble minute white pseudocyphellae. t. Kahusi, 2700 m., F. L. Hendrickx 4316; km. 30, route Kahusi, F. L. Hendricks E both in E. African Herb. ME Elgon, 1290 m., arboricole, W. Small 218, at Kew so CA: without locality. Drége, ex herb. W. Sonder, in 5 Tuckerman herb. sub U. L4. at Bartow Herb. 43. USNEA TRANSVAALENSIS Vainio, Ann. Univ. Fenn. Aboens. A2?:1. 1928. Usnea pectinata Stirton, Scottish Nat. 7:77. 1883, non Taylor. Usnea tropica Zahlbr., Cat. Lich. Univ. 6:596. 1930. Usnea florida v. densirostra Stzbgr., Ber. Thatigk. St. Gall. Naturw. Ges. 1888/89: 145. 1890, excl. syn. Type: Transvaal, Barbeton, saxicole, P. A. van der Bijl 229. The type of U. bectinata Stirton and U. tropica Zahlbr. is from S. Africa, Somerset East, P. Mac- Owan, in the Art Galleries of the Glasgow Corporation. Thallus erect, fruticose, rigid, 3.5—5 cm. tall, pale green, becoming old gold to Isabella color; holdfast up to 4 mm. in diameter, of black, branched and tangled Es formed by a coalescence of several plants or branching at the very base; rimary axes 1—1.5 mm. in diameter, 2-6 mm. long to the first branch, not Hen, then closely dichotomous, almost at right angles, the branches curving upward; branches rare in the upper half, slightly constricted and annularly cracked at the base, about 0.8 mm. in diameter below, tapering upward to blunt tips; sparingly annularly cracked throughout, terete, glabrous below, appearing minutely verrucose above from the initials of soralia; ramuli very dense, mostly about 1 mm. long, a few up to 2 mm., tips blunt, very minutely verrucose but mostly covered with minute soralia of isidiose soredia which often germinate in situ, forming small isidia 0.1—0.2 mm. long, with blackened tips, and thus appearing secondarily ramulose; apothecia not seen. Cortex 80 p thick, fastigiate, highly gelified, the outer 10 p with minute brownish crystals; algal layer 30 (—40) y thick, continuous, cells protococcoid, 11-12 p in diameter, somewhat polyhedral from mutual pressure; medulla 270 p thick, of very loosely woven hyphae 3—4 pn in diameter, lumen about 1 y, somewhat nubilated with hyaline granules; axis solid, 150 pn in diameter, of slender thick- walled, predominantly longitudinal hyphae. The relationship of this species to U. pulvinata Fr., collected by Preiss in West- ern Australia, is not clear. Motyka's description is based on the whole complex from South Africa northward to Kenya. He states that the medulla is evanescently K yellow, while in the type of U. transvaalensis and in our material, the medulla is K yellow becoming red and finally reddish fuscous. He states that the ramuli are intermixed, short and long, up to 10 mm., not cracked at the base, and tips [Vor. 44 52 ANNALS OF THE MISSOURI BOTANICAL GARDEN subulate, often blackened; while in our material the ramuli are nearly uniformly 1 mm. long, with very few reaching 2 mm. In our plants the cortex is thicker and the algal layer and medulla are twice the thickness of the axis, while in U. pulvinata (fide Motyka) the cortex is thinner and the medulla only three fourths as thick as the axis. NDA: Karamoja, Napak, 2250 m., on rocks in mountain grassland, A. S. Thomas 3631; ; Tmatong Mts. 2100 m., A. e T bows 1678, both at Kew CA: Soinbriet East, P. MacOwan, type of U. pectinata Stirton non Tayl. and v. fies nri in the Art ‘Galena of the Glasgow Corporation. Imerina, Andrangoloaka, saxicole, Hildebrandt 2172; beoe. Betafo, ies albei. ex Herb. E. G. Paris; both at Farlow Herb. 44. UsNEA RUWENZORIANA Motyka in Zahlbr. & Hauman, Mém. Inst. Roy. Colon. Belg. 5:27. 1936. Type: Congo, Mt. Ruwenzori, 4000 m., ramicole, L. Hauman 929, 947. Thallus erect, 4 cm. tall, cespitose, pale ashy green, drying dirty fuscous (Sac- cardo’s umber to light brownish olive in our plants); base hemispheric, 1.5 mm. in diameter, narrowing to 0.3—0.6 mm., 1-3 mm. to the first dichotomy, lower portion blackened; closely several times dichotomous near the base, axils 45-90°, main branches 1—1.2 mm. in diameter below, tapering gradually to short abrupt summits, terete, smooth, dull, rarely somewhat foveolate; ramuli quite dense, mostly about 2 mm. long, quite slender, slightly constricted at the base and somewhat inflated in the middle, some growing out as secondary branches; tubercles on summits lg upper ramuli forming soralia, isidiose at first, soon eroded and farinose. : Kigezi, Kasatoro, Kabale River, 2250 m., arboricole, I. R. Dale L43; pecu aere 2410 m , A. S. Thomas 484 (slightly lighter i in vem Ie both a MODE Kills 2900—29 : Mt. Kahusi, 2700 m., F. L. R pec in E. prises Herb. 45. UsNEA HAUMANI Motyka in Zahlbr. & Hauman, Mém. Inst. Roy. Colon. Belge 5:25. 1936. : Congo, Mt. Ruwenzori, southwest slope, 4200 m., on Senecio Friesorum and H robe biomate Stublmannii, L. Hauman 921. Thallus pendent, flaccid, up to 10 cm. long, pale ashy green, (larger branches tawny olive in our much older material), constricted at the base, then expanding rapidly to 1.5 (—2) mm. in diameter, irregularly dichotomous throughout, axils varying from 45 to 90°, irregularly curved, branches somewhat constricted at the base, rather thick throughout, irregularly annularly cracked, nearly terete, surface smooth or slightly rough from low tubercles (initials of small soralia?) ; ramuli very rare or absent, the few present suggesting that they are young branches arising from unequal dichotomy; summits indistinct. Apothecia not seen. Soralia abundant on the slenderer branches, up to 0.2 mm. in diameter, sometimes crowded but rarely confluent, much smaller and more 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 33 scattered on the larger branches, subhemispheric, probably isidiose at first but mostly appearing farinose, some isidia germinating in situ and forming nest of propagula which early break away. Our material from Cameroons consists of somewhat smaller plants, base broken away in collecting, color of the main branches darker (as they have been 23 years in the herbarium instead of 4 years at the time Motyka studied the type). Our plants were growing on stones, as were those from Kilimanjaro, 3500—4000 m., Fr. & R. Wettstein, cited by Motyka in his original description. The Uganda material is more typical. The Angola material is just beginning to form apothecia, sessile on the main axes for some distance back of the ultimate branches. quoq: Mt. Cameroon, 4030 m., on bare rocky ground, F. W. H. Migeod 196, at Kew. UGANDA: Kigezi, Muhavura, 3610 m., saxicole, I. R. Dale L18, at Kew. ANGOLA: Cuanza Sul, Amboim, Capir near the Carloaongo-Cuvo River, 1000 m., on dead trees, J. Gossweiler 99Q1a, at Kew. 46. USNEA sIMPLicissiIMA Motyka in Zahlbr. & Hauman, Mém. Inst. Roy. Colon. Belge 5:29. 1936. Type: Congo, Mt. Ruwenzori, 4000 m., on Hypericum, L. Hauman 952. Thallus 5 cm. or less tall, simple or sparingly branched, subfruticose, pale green when fresh, drying between buckthorn brown and Dresden brown, opaque; base gradually but distinctly attenuate and blackened, several plants from a common holdfast giving a branched appearance; branches usually simple, somewhat curved, thickened to 1.3 mm. in diameter, subinflated, tapering from the middle upward, continuous or rarely articulate, constricted in the cracks, terete to slightly com- pressed and deformed, quite closely ramulose; ramuli 5-15 mm. long, variously curved, slightly tuberculate, tips rather obtuse. Apothecia apical, up to 5 mm. in diameter, often subirregular; exciple smooth, subconvex; marginal cilia relatively abundant, erect at first, finally spreading, of variable length; disc concave at first, becoming nearly plane, carneous, white- pruinose; ascospores 9—11 p in narrow asci. Cortex about 30 pn thick, papery; medulla about 200 p thick, very lax, white; axis 180 p in diameter. Our material from the Buddu District is evidently quite young, 2 (-3) cm. tall, the taller mostly once-dichotomous near the middle; apothecia cupulate with erect marginal cilia, only about 1 mm. in diameter, lateral near the tip, not strictly apical. Except for smaller dimensions, our plants agree quite well with Motyka's description translated above. UGANDA: Buddu District, on twigs in sandy plain, 3950 [m. or ft.?], H. U. Baul [surname rather illegible], ex Bot. Gard. Entebbe, 1905; Kigezi, Kasatoro, in forest, 1930— 2580 m., I. R. Dale L39 p. p. min.; both at Kew. [Vor. 44 34 ANNALS OF THE MISSOURI BOTANICAL GARDEN 47. UsNEA FUSCORUBENS Motyka, Lich. Gen. Usnea Stud. Monog. 546. 1937. Type: Mauritius, without locality, Simony. Thallus 6 cm. long, evidently subpendent, main axes Mars brown, secondary branches and ramuli russet; irregularly dichotomous throughout, axils about 90° angles; branches 1—1.5 mm. in diameter, irregularly curved, almost terete, some- times deeply foveate, continuous or sometimes annularly cracked, verruculose- papillate, papillae short, cylindric, esorediose; ramuli distant, variously curved and ranched, very slightly attenuate at the base, subinflated, almost smooth, the larger subtuberculate and subfoveolate, tips acute. Apothecia about 6 mm. in diameter; exciple smooth, shining, the larger ones minutely papillate and foveolate; marginal cilia up to 10 mm. long, often branched, similar to the ramuli; disc slightly concave, densely white-pruinose. The specimens from the Cape of Good Hope are somewhat larger, one about 7 cm. long, the other 12 cm. long, the lower joints about 2 mm. in diameter, and the apothecia of the shorter plant 8-9 mm. in diameter; in other characters, both macroscopic and microscopic, they agree well with Motyka's description translated above. TH AFRICA: Cape of Good Hope, without locality or collector, ex herb. Menzies in eier herb. sub U. florida, at Farlow Herb. 48. USNEA sUBMOLLIs Steiner, Verhandl. Zool. Bot. Ges. Wien 53:229. 1903. Type: Cameroons, Fako, Bornmüller. Thallus up to 5 cm. tall, simple or branched below, olive green (olive lake to ecru olive in our plants) ; holdfast indistinct, base not blackened, somewhat attenu- ate below, expanding to 0.6 mm. at the first branch; branching sympodial just above the base, branches somewhat constricted and annularly cracked at the base, inflated to 0.6 mm. in diameter in the lower third, then tapering to the tip; terete, not articulate, smooth or minutely and sparsely verruculose; ramuli not dense, about 1 mm. long, nearly perpendicular to the axis, bases articulate and slightly constricted, straight or slightly curved upward, slightly inflated, a few longer, up to 5 mm. long, with an occasional ramulus, perhaps aborted secondary branches; summits subulate. Apothecia subterminal or lateral, sessile, 1.5 mm. in diameter, concave, exciple smooth; marginal cilia few (about 8), distant, suberect, resembling the small ramuli in size and shape; disc pale flesh color, white-pruinose. Our Cameroons plants may be still immature as the dimensions are somewhat smaller than given by Motyka; the disc is covered by a white pruina. Our Nyasa- land plants are a little larger than the type, the ramuli denser, and the apothecia up to 3 mm. in diameter. CAMEROONS: Buea, 1290 m., on sticks on the ground, F. W. H. Migeod 22, at Kew. Se ne Mt. Minagongo, D. H. Linder 2172, at the Farlow Herb. and in Mo. Bot. ard. He 1957] DODGE—LICHENS OF TROPICAL AFRICA. IIl. USNEA 35 NYASALAND: Mlange District, Luchenya Plateau, Mt. Mlange, 1890 m., at forest "m L. J. Brass 16828, 16847 (Vernay Nyasaland Exp.), at N. Y. Bot. Gard. and Mo. Bot. Ga nu erbs. AFRICA: Cape of Good Hope, summit of mountain above Simon's Town on Simon" s Bay, Charles Wright (Rogers & Ringgold N. Pacific Expl. Exp.), det. U. barbata v. florida by Tuckerman ex Gray Herb., at Farlow Herb. 49. USNEA BLEPHAROIDES Dodge, Ann. Mo. Bot. Gard. 43:395. 1956. Type: Tanganyika, Kilimanjaro, at upper limit of forest, 3000 m., G. Geilinger 4417, at Kew. Thallus fruticose, little-branched, prostrate or suberect, about 9 cm. long, axes drying tawny olive, ramuli and cilia cinnamon buff; holdfast broken off, sub- articulate, lowest joint slender, tapering upward, more than 3 mm. long, 1 mm. in diameter below, expanding to 2 mm. above; joints constricted but cortex mostly continuous, with usually two lateral branches on the outer side of the curve of each joint, producing flexuous main branches, the last 3 mm. of the axes and all lateral branches unbranched, or with 1-2 secondary branches scarcely larger than ramuli but bearing minute apothecia; main branches 2 mm. in diameter below, tapering to 1 mm. under the apothecia, terete (somewhat crushed in the type), slightly verrucose (visible under 12 magnification), becoming minutely papillate on the joints just below the apothecia and at the bases of the larger ramuli; ramuli subdistant, 4-5 (—10) mm. long, curving upward, articulate and subinflated at the base, tips acuminate. Apothecia 11-12 mm. in diameter, sessile on axes and slender branches, about 1.5 mm. from the tip, appearing terminal when mature, with concrescent ramuli forming low radial ridges, minutely verrucose (under 12X magnification); marginal cilia distant, up to 10 mm. long, curved, tips acuminate; disc flat, flesh color with a thin white pruina. Cortex 85-125 p, very variable in thickness, forming verrucae outside up to 60 p tall and protruding as teeth between the algal colonies, 60 toward the medulla, outer 55 p fastigiate, of conglutinate thick-walled hyphae 8 y in diameter, lumen 2 p, the rest of densely woven, predominantly periclinal thick-walled con- glutinate hyphae, but the teeth between the algal colonies of radial hyphae, the outermost 15 y nubilated with minute brown crystals; ra layer discontinuous, colonies about 30 u in diameter, cells of Trebouxia 6-7 p in diameter, closely packed; medulla about 375 y thick, the outer portion (eur?! 80 ») densely woven, the rest of very loosely woven, predominantly radial hyphae 8 » in diameter, lumen 3 m, dichotomously branched; axis 250 p in diameter, of longitudinal conglutinate thick-walled hyphae 3 p in diameter This species differs from U. blepharea Motyka in having a larger axis, more gradually attenuated upper portions of the branches, fewer ramuli, an exciple without spinules, and a more regular disc. Perhaps the specimen from Kilimanjaro cited by Motyka belongs here rather than in U. blepharea, which is otherwise a Javan and Bornean species. Perhaps a single plant sorted out of R. G. Turrall 55 [Vor. 44 36 ANNALS OF THE MISSOURI BOTANICAL GARDEN may also be this species. It is only 4 cm. tall, about 1 mm. in diameter below, surface not quite as verrucose and apothecium only 3.5 mm. in diameter. TANGANYIKA: Kilimanjaro, at upper limit of forest, 3000 m., G. Geilinger 4417, type; 2930-3045 m., on twigs of giant heather in clearing in forest, R. G. Turrall 55 p. p. min.; both at Kew 50. USNEA MOLLIUSCULA Stirton, Scottish Nat. 7:77. 1883. Type: Australia, Victoria, subalpine country in upper Ovans River, Mrs. Mc- Cann I, in Art Galleries of the Glasgow Corporation. Thallus 6-7 cm. tall, fruticose, subflaccid, ochraceous tawny to buckthorn brown; holdfast about 3 mm. in diameter, of rootlike strands, dark, giving rise to 2—3 axes up to 2.3 mm. in diameter, somewhat attenuate but not blackened at the base; branching unequally dichotomous to sympodial, articulate, lowest joint nearly cylindric, about 5 mm. long, smooth, upper joints inflated, tuberculate to sub- papillate, many papillae growing out as short, rather coarse spinules, sometimes in rows in the upper joints just under the apothecia, giving a somewhat foveolate appearance; ramuli sparse in the type, mostly 1-3 mm. long, a few up to 10 mm., annularly cracked and slightly inflated at the base. Apothecia sessile, about 8 mm. in diameter, margins somewhat inrolled when dry; exciple smooth or somewhat reticulate-foveate; marginal cilia distant to ir- regularly close, very variable in length, up to 5 mm. long, tips acute, mostly broken off; disc white-pruinose; ascospores 9—11 X 7-8 p In our South African material, the ramuli are irregularly dense, 2—3 (-10) mm. long, the longer annularly cracked in the middle and sparingly again ramulose. The apothecia are subterminal on short lateral branches, the exciple is smooth, the marginal cilia closer, shorter, and more inflated. AUSTRALIA: Victoria, subalpine country of upper Ovans River, Mrs. McCann 1, in the Art Galleries of the Glasgow Corporation, a fragment in the British Museum conco: Mt. Kahusi, 2700 m., ramulicole, F. L. Hendrickx 4308, 4314, both in E. African Herb. SOUTH AFRICA: Hermansdorp, Blue Lilies Bush, corticole, A. Woodcock, at Kew. 51. USNEA oBTUsATA Motyka, Lich. Gen. Usnea Stud. Monog. 622. 1938. Type: Ethiopia, north slope of Mt. Kubbi, Schimper. Thallus up to 8 cm. tall, fruticose, subrigid, dirty ashy greenish in herbarium (Motyka) but sepia shading to tawny olive in our material; base black, about 1 mm. in diameter, main axis rapidly expanding to 2 mm. and tapering upward to 1.5 mm. just below the terminal apothecium; lateral branches about 1 mm. in diameter, nearly perpendicular to the main axis but curving upward, subarticulate, slightly inflated, smooth or irregularly papillate (evidently very young ramuli) ; ramuli dense, perpendicular, mostly 1-2 mm. long, sometimes in longitudinal rows in the upper portion, giving a subrugose appearance, tapering from the base to a blunt tip, not articulate nor inflated, slightly curved upward. 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 27 Apothecia terminal (really subterminal, the tip of the branch concrescent with the exciple and protruding about 3 mm.) ; exciple smooth, with spines about 1 mm. long; marginal cilia close, about 3 mm. long, straight or slightly incurved, tapering to an acute tip, rather fragile; disc concave to somewhat deformed and coarsely rugose, flesh color (Motyka), tawny olive or darker in our material, white-pruinose. Cortex 40—50 y thick, of fastigiate pseudoparenchyma, thick-walled conglu- tinate cells, lumina 5—6 pu in diameter; algal layer 100 p thick, cells densely packed under the cortex and next the medulla, more scattered between, protococcoid; medulla 300—360 pu thick, of very loosely woven, thin-walled hyphae 3—5 p in diameter; axis solid, 160 y in diameter, of vertical, thick-walled hyphae 2.5—3 p in diameter. 5la. f. PERsTRIGOSA Motyka, Webbia 8:391. 1952. Our material probably should be referred to f. perstrigosa Motyka, differing in strict, more acute ramuli, often very dense and paler fuscous The species was originally placed in the ALBOMACULATAE, but transferred by Motyka (loc. cit.) to the sETULOSAE, DENSIROSTRES. Although the original description calls for "sor- dide cinereo-viridulus," the description of the variety implies that the species is fuscous. Pseudocyphellae are not mentioned in the original description, hence the species does not belong in the ALBoMACULATAE. The thick, lax medulla, fuscous color in the herbarium, lack of papillae, and denser ramuli indicate a place in the GLABRATAE, CILUFERAE, perhaps near U. simplicissima, rather than in sETULOSAE, DENSIROSTRES, with their thinner, very dense medulla and grayish green colors. Unfortunately I have not seen Motyka's types. The axis is relatively thinner than the dimension given by Motyka. ETHIOPIA: Mulu, 20 miles north of Addis Ababa, 2580 m., P. R. O. Bally 2078, at Kew; Bagla, 2410 m., Hildebrandt, 1872, ex Herb. Sbarbaro, at Farlow Herb. 51b. f. PERFLAvA Motyka, Webbia 8:392. 1952. Type: Ethiopia, Bacino del Tana, 2100—2200 m., Pichi-Sermolli. Differs from the type in its Perna fuscous color, especially in the upper part of the branches with golden yellow ram Chipp 57, from southern Sudan, seems to belong here. The main axis is tawny olive, the lateral branches shading to chamois and the ramuli colonial buff. The main axis is articulate, the joints more constricted at the ends; the apothecia are somewhat smaller, the marginal cilia less dense, and some of the longer ones branched. SUDAN: Onyiro, Issare, 1770 m., 32° 50’ E., 3° 50' N., T. F. Chipp 57, at Kew. 52. UsNEA UNDULATA Stirton, Scottish Nat. 6:104. 1881, non alio loco. Usnea foveolata Stirton, Scottish Nat. 7:77. 1883. Usnea barbata v. farinosa Müll. Arg., Flora 74:375. 1891. Usnea farinosa Zahlbr., Cat. Lich. Univ. 6:564. 1930. Usnea Steineri v. tincta £. sorediosa Zahlbr., Bot. Jahrb. [Engler] 60:541. 1926. [Vor. 44 38 ANNALS OF THE MISSOURI BOTANICAL GARDEN Type: Bechuanaland, West Griqualand, Kimberley, J. Shaw, in Art Galleries of the Glasgow Corporation. Type of U. foveolata from same locality and collector in British Museum (Natural History). Type of U. barbata v. farinosa and U. farinosa from South Africa, Peter MacOwan. allus 9 cm. or more long (up to 30 cm., fide Stirton), fruticose, pale ashy green, becoming buffy olive in the herbarium; holdfast of discrete, branched, flattened rootlets, not united into a disc, about 3 mm. long, rose color below where torn from the bark, buffy olive above; several times dichotomous just above the base, axils nearly 90? angles, the branches 1.5 mm. in diameter, curving and soon . nearly parallel, flexuous and undulate, internodes longer above, rather densely covered with short ramuli, about 1 mm. long below, shorter and more slender spinules above; the more slender branches somewhat foveolate from minute isidiose soredia in rows, forming very slight anastomosing ridges (perhaps only scars of broken-off spinules) ; summits indistinct, mostly broken off. Apothecia (absent in our plants) about 7 mm. in diameter, subterminal; exciple smooth or foveolate, with short spinules; marginal cilia very few, short, subulate; disc plane, white-pruinose; ascospores 4 X 8 p, in narrow asci (fide Motyka). A somewhat variable species, extending from Kenya to South Africa. Material from Kenya seems to be shorter with more spreading branches, some discolored reddish (about cameo brown), probably from too slow drying in the plant press. Ridges and foveolation are very indistinct in most of the plants, but are much more distinct in the type of U. foveolata and in Dale 43 from Uganda. KENYA: Chyulu in H. D. van Someren (2 sheets in E. African Herb.) ; Lake Ngu nga, 1775 m., on trees, G. M. Allen 1838, in Howe Herb. sub U. florida v. pU rigosa at Farlow Herb.; Nairobi, on trees, G. M. Allen 1784, in Howe Herb. sub U. florida v. strigosa at Farlow arhi Mt. Kenya, UG m., Edgar A. Mearns 1776 (T. Roosevelt Exp.), ex mp? = Se € at Farlow H ISL ee blind, Pomoni cde tad on Cocos trunks, Hildebrandt, ex herb. ‘Sbarbaro Sc: V. barbata f. sorediata, at Farlow H : Kigezi, Kasatoro, Kabale R , 2250 m., srboricole, I. R. Dale L43. NGO: km. 30, route Kahusi, F. L. Hendricks 4330; Mt. Kahusi, ca. 2700 m., F. L. Hendrickx 4142, 4303, 4307, PET all ia E Alsicin Herb. NORTHERN RHODESIA: uplands of E. Isoka District, 1480 m., on tree in Brachystegia mimosaef olia scrub — d, C. G. Trapnell, in E. African Herb. BECHUANALAND: W. Griqualand, Kimberley, J. Shaw, in Art Galleries of the Glasgow Corporation, type St U. dene same locality and collector, in British Mus. (Nat. Hist.). Hope, Uitenhage, forest near Zunday's River, Zeyber, in Taylor herb. sub U. florida, AER a state?, at Farlow Herb. 53. USNEA STRIGOSELLA Steiner, Bull. Herb. Boissier, II. 7:637. 1907. Type: South Africa, Transvaal, Letaba District, Mt. Mamotsuri, Henri A. Junod; terminal fragment in Howe herb. at Farlow Herb. Thallus pendent, about 30 cm. long, 0.6—1 mm. in diameter, slender, greenish stramineous, drying mikado brown or lighter on the main branches, ramuli and smaller branches cream buff; base short (absent in our plants), several times closely dichotomous below, axils acute, and again several times dichotomous with longer 1957] DODGE—LICHENS OF TROPICAL AFRICA. IIl. USNEA 39 internodes above, axils nearly right angles; spinulose and closely ramulose, terete, annularly cracked, sometimes slightly constricted at the cracks, sometimes not; smooth at first, often becoming spinulose and the surface uneven from the scars of the fallen spinules, sometimes slightly deformed where the spinules and small ramuli are arranged in close rows but not angular; ramuli close to dense, mostly about 2 mm. long, a few longer; summits short and indistinct. Apothecia rare (absent in our plants), about 6 mm. in diameter, sublateral; exciple nearly plane, smooth; marginal cilia dense, radiating, of variable length, up to 12 mm. long, often annularly cracked; disc nearly plane, pale sulfur pruinose; ascospores 4 X 8 yp in narrow asci. Cortex 65—100 (-115) p thick, fastigiate, gelified, yellowish brown from minute crystals in the outer 65 p, more hyaline within; algae in discrete gen about 30 u in diameter, with smaller colonies scattered in the medulla; med variable, 65—190 , thick, thicker where the cortex is thinner, rather dci except in a narrow zone next the axis where it mostly tears apart in sectioning, hyphae 6-7 yw in diameter, nubilated with granules, hyaline in outer portion, brownish in inner 30 p next the axis; axis about 420 p in diameter, transverse section somewhat irregular, of conglutinate, longitudinal hyphae 3 p in diameter, with scattered cavities 7-8 p in diameter, filled with brownish crystals and sur- rounded by a layer of thin-walled hyphae (reminding one of the resin canals of conifers). NYASALAND: Mlanje District, puro T Mt. Mlanje, 1890 m., L. J. Brass T Biwer Nyasaland Exp.) in N. Y ard. Herb. TH AFRICA: Transvaal, Letaba pue Mt. Mamotsuri, Henri A. Junod, type ape com. Zahlbruckner, in e herb. at Farlow Herb. 54. USNEA PERsPINOSA Motyka, Lich. Gen. Usnea Stud. Monog. 2:533. 1937. Type: Cameroons, mountain forest, 2200 m., C. Ledermann. Thallus about 5 cm. tall, fruticose, spreading, rigid, green when fresh, drying fuscous (Isabella color to light brownish olive in our plants), ramuli lighter; hold- fast low, conic, verrucose, not blackened, 4 mm. in diameter, base about 2 mm. long and wide and densely sympodially branched at the top, then less closely, but internodes short throughout; axils nearly right angles, main axis 2 mm. in diameter below, tapering to about 1 mm. below the apothecia; branches constricted at the base and often articulate, 1—1.5 mm. in diameter, terete, smooth to somewhat rugose from rows of spinules, in places appearing short-papillate from healing scars of the caducous spinules which are about 0.2 mm. long; cortex areolate, diffract in some places; ramuli up to 3 mm. long, annularly cracked at the base, slightly v: and surface uneven, tips acute, often broken off. pothecia abundant, subterminal, sessile, cupulate, 2-3 (-10) mm. in diameter; De subfoveolate, densely spinulose; marginal cilia very dense, erect, inflated, up to 3 mm. long, tips forked, acute; disc very concave, white-pruinose. The Uganda plants are somewhat variable; Dale Loo is larger, nearly 7 cm. tall, [Vor. 44 40 ANNALS OF THE MISSOURI BOTANICAL GARDEN less densely branched, with one apothecium about 6 mm. in diameter; Dale L. 85 has a few apothecia up to 10 mm. in diameter; Dale L38 consists of plants smaller in all dimensions; Dale L41 p. p. has large apothecia but is shorter. The Kenya plants are paler and less densely branched, and the marginal cilia of the apothecia are shorter and more spreading. However, they seem closer to this species than to any other. Mearns 508 has apothecia up to 18 mm. in diameter. t. above Buea, 2500 m., on dead decorticate shrub in grass- land, E Milbraed KR 28, de and juven : Bunyoro, Busingoro, on biko f Jacaranda tree, I. R. e Loo: Kigezi, Mar. 2410 m., on eg , I. R. Dale L41 p. p.; Kasatoro, Tahal River, néi m., on trees, I. R. Dale L43; Kabale, 2000 m., on bash of Hibiscus rosa-sinensis, I. R. Dale L55; Fort iunii on ze scht 1620 m., I. R. Dale L38; Fort Portal District, Toro, on trees, I. R. Dal Mt. Elgon, "a m., W. Small 217; Bugishu, Bulambuli, on trees, A. S. Thomas 548, ed young, all at K ENYA: Ngong, H. D. " SR eren rte E. African Herb.; Guaso Nyiro, G. M. Allen 1792, in Howe herb. U. florida v. strigosa Ach.; Narok River, 2000 m., Edgar A. Mearns 508 (Theodore ds at Exp.); both at Farlow Herb. 55. UsNEA LUTEOLA Motyka, Frag. Florist. Geobot. 1:33. 1954. Type: Madagascar, Manakambahiny Est, Marovato between Alaatra and Maro- vato, 800 m., com. des Abbayes Thallus up to 8 cm. long (5.5 cm. in our plants), caespitose, fruticose, sub- flaccid, pale yellowish to yellowish stramineous (axis Sayal brown, slender branches and ramuli chamois in our plants) ; holdfast of radiating, rather large, dichotomous, white rootlets, adhering to the bark and penetrating it at their tips; base not black- ened, 1.2 mm. in diameter, 14 mm. long; about thrice dichotomous with relatively short internodes, axils nearly right angles but curving upward, divergent, not parallel; terete, rather densely covered with ramuli about 3 mm. long, and slender lateral branches up to 15 mm. long; cortex continuous or cracked but not articu- late, the fertile branches tapering to about 0.5 mm. in diameter below the apothecia, sterile branches tapering to long-acuminate summits; main axes smooth except for small verrucae which develop to ramuli, the more slender branches papillate and tuberculate, the tubercles producing soralia with very few isidiose soredia, becom- ing farinose after the soredia are shed; ramuli rather dense, about 3 mm. long, with acute tips on the main axes and lower portions of the slender branches, mixed with coarser ones up to 10 mm. long (perhaps better considered as young lateral branches as they are tuberculate-sorediose) ; summits about 1 cm. long, tuberculate-sorediose, usually eramulose, sometimes forked above, tips acute. Apothecia lateral about 1 cm. below the tips of the branches, bending them downward, so that the apothecia appear subterminal, 3-5 mm. in diameter, nearly plane; exciple smooth; marginal cilia 2-ranked, the outer row subdistant, up to 10 mm. long, tuberculate-sorediate, the inner row short, 0.5-1 mm. long, closer; disc flesh color, covered by a dense white pruina at first, somewhat rugose. MAURITIUS: Reduit, corticole, G. Orian 3, at Kew. ER eee ree Se ee a we Les 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 41 56. USNEA SULPHURASCENS Motyka, Frag. Florist. Geobot. 1:34. 1954. Type: Madagascar, Manakambahiny Est, between Alaatra and Tanatave, 800 m., com. H. des Abbayes. Thallus subdependent, flaccid, about 9 cm. long, sulfur yellow becoming ochra- ceous buff; base short, attenuate, indistinctly blackened; dichotomous to sym- podially branched, axils nearly right angles, branches slightly constricted at the base, articulate, arcuate, axes 1—2 mm. in diameter, joints inflated in the middle, gradually tapering to the next articulation, terete, minutely tuberculate, tubercles bearing single white isidia and appearing pseudocyphelliform after the isidia are shed; summits slender, acuminate; apothecia not seen. Cortex up to 40 y thick, of coarse radial hyphae in a gel; algal layer 25-35 p thick, cells Trebouxia ?, 6-7 p in diameter in discrete colonies, forming a nearly continuous layer; medulla 125—150 y thick, very compactly woven just under the algal layer, then very lax, of relatively thin-walled, branched hyphae, mostly 5—6 p in diameter. Our De is ochraceous buff instead of stramineo-ochraceous, perhaps from having been much longer in the herbarium; the branching is more dichotomous on the larger axes; otherwise it agrees well with Motyka's description. Our plant is only 5 cm. long, but the lower portion has been broken o MADAGASCAR: without locality, Dr. Moss, March 1895, at Kew. 57. USNEA syriaca Motyka, Lich. Gen. Usnea Stud. Monog. 278. 1936. Type: North Syria, Amanus near Beiland, on Abies at foot of Akma Dagh, 2600 m., Kotschy. Thallus about 10 cm. long, fruticose, rather slender, pale ashy green, slightly fuscent in the herbarium, darker on the primary branches (main axes mikado brown, slender branches and ramuli cinnamon buff in our plant), opaque, base short, rather rigid, dark, infrequently branched from the base, axils nearly 90* angles or more acute; branches somewhat flexuous, about 1 mm. in diameter near the base, tapering gradually upward, terete, continuous or rarely simply annulate, the thicker branches shortly and densely papillate, the more slender branches less so, forming very obtuse ridges, the backs decorticate and farinose, evidently producing soredia as well as the tips of some papillae; ramuli not close, mixed longer and shorter, very thick at the base, tapering gradually upward; summits rather long, farinose, so that they appear whitish. Cortex 60 p thick, rather hard and elastic, pale; medulla about 150 y thick, white, dense, K intensely reddening; axis 330 p in diameter with some dark hyphae. Our plant from Eritrea agrees more closely with Motyka's description translated above than with any other, but the medulla is K—. The base is absent and the axis blackened about 2 cm. and up the secondary branches, suggesting that it is prostrate and the axis covered with leaf mold, in which case it may be much longer and the portion collected only the terminal portion. The cortex regenerat- [Vor. 44 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN ing in the annular cracks shows as white rings about the lower portions of the secondary branches. The soralia are well developed, slightly elevated, producing granular soredia, in some places so crowded as to form almost continuous areas. The proportions of cortex, medulla and axis are nearly the same, but the absolute dimensions are nearly double. The structure of the cortex is unusual, being formed of interwoven and anastomosing hyphae with moniliform lumina, leaving entrapped minute air bubbles (seen in freshly mounted sections), suggesting the spongy cortex of the subgenera LETHARIELLA and cHLoREA. The algae occur as small colonies throughout the moderately lax medulla, not forming a separate layer be- tween the cortex and medulla. The algal cells seem to be short, discrete cylinders, although packed in rounded colonies, suggesting that they may have come from filaments of Trentepoblia with nearly isodiametric cells, occasionally more fila- mentous as they push out along cracks in the cortex. The axis is solid and quite uniform with an occasional small bundle of brownish hyphae in the outer 50 p. Motyka mentions a plant from Anatolia, Ibidh dere, H. Czeczott, at Cracow, which has a solid axis and a K— medulla, tentatively referred to U. syriaca. Our plant may be a distinct species, but I dislike to describe a new taxon on a single incomplete plant. a, without locality, F. Gallina, ex herb. Sbarbaro sub U. plicata, at Farlow erp. 58. USNEA SUBLEPROSA Motyka, Lich. Gen. Usnea Stud. Monog. 605. 1938 Type: South Africa, Transvaal, Lydenburg, Wilms. Thallus up to 20 cm. long, usually smaller (7 cm. in our plant), rigid, perhaps pendent, subglaucescent drying olive green (Motyka), (fuscous on larger branches shading to dark olive buff at the summits and ramuli in our plants) ; base penetrat- ing crevices in the rock (broken off in our plants); lower portion about 1.5 cm. long, 1.5 mm. in diameter, black, forming an arc of about 90°, sympodially branched with very short internodes; principal axes flexuous, articulate, joints constricted at the ends, slightly inflated in the middle, up to 2 mm. in diameter below, of variable length, rarely annularly cracked, densely verrucose to sub- papillate, the more slender lateral branches subarticulate, verrucae less dense and larger, forming slightly elevated soralia of isidiose soredia which appear farinose after the soredia are shed; ramuli subdistant, about 10 mm. long, curving toward the summits, verrucose and sorediose, slightly constricted at the base, annularly cracked, inflated in the lower portion, tapering to a short obtuse tip, often broken off; summits about 20 mm. long from dichotomy of the axis or lateral branch, about 0.5 mm. in diameter, resembling ramuli in structure. Apothecia rare (absent in our plants), lateral to subterminal, up to 7 mm. in diameter; exciple uneven, sublacunose, verruculose and sorediose, sometimes ciliate; marginal cilia few or none, about as long as the diameter of the disc, rather thick, 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 43 obtuse, subarticulate, with farinose soredia; disc plane or somewhat deformed, flesh color, slightly or not pruinose. NYA: Ge? ém G. M. Allen 1792 p. p., in Howe herb. sub U. florida v. strigosa Ax ae — Nco: Mon kt, about 3000 m., saxicole, F. L. Hendrickx 1273. P GESE MAIOMHE: Chiluango, J. Gossweiler (1919); rather young, papillae in dense longitudinal rows on some branches, not on others, giving an angular appearance. E) 59. UsNEA LrpreNa (Stein) Motyka, Lich. Gen. Usnea Stud. Monog. 593. 1938. Usnea strigosa v. Ledienii Stein, Jahresber. Schles. Ges. Vaterl. Cultur 66:139. 1888. Type: Cameroons, Buea near the caves, Preuss. Thallus fruticose, erect, rigid, about 7 cm. tall, drying citrine drab, more slender branches and ramuli dark olive buff; base 3 mm. in diameter, a dense mass of root-like fibers penetrating the bark, slightly darker; densely sympodially branched at the base and rarely dichotomous above, branches up to 1.5 mm. in diameter below, tapering gradually upward, remaining rather thick to the summits, somewhat flexuous, slightly attenuate and articulate at the base, not or only ve slightly articulate above, with simple annular cracks; terete, sparsely and irregu- larly tuberculate to subpapillate below, more closely so above, tubercles sometimes in rows but not close enough to form ridges, forming soralia of isidiose soredia, especially on the more slender branches and larger ramuli; ramuli quite close, some 1-3 mm. long, smooth, others up to 15 mm. long, tuberculate and sorediose, some- times so densely so that they appear furfuraceous, axils right angles but the longer ramuli curve upward, flexuous. Apothecia not abundant, subterminal on the main axes, up to 10 mm. in di- ameter, often deformed; exciple plane or deformed, smooth when young becoming uneven and foveolate or scrobiculate; marginal cilia not close, long and short inter- mixed, rarely longer than the diameter of the disc, relatively thick, the longer tuberculate and with secondary ramuli; disc nearly plane or deformed, chalky white pruinose; ascospores 7—10 u in narrow asci. The above description of apothecia is translated from Motyka, loc. cit., as only a single young apothecium is present in our materi oer — above Buea, 1450 m., F. W. H. Migeod 144 p. p. m GOLA: Cuanza Sul, Amboim, Capir, near the Carloaongo River, 1000 m., on dead eek E: — d Benguells; country of the Ganguellas and Ambuellas, J. Goss- < ANDA: Ki sie? Kasatoro, Babele Km 2250 m., I. R. Dale L43; Mt. Elgon, 1290 m., on trees, W. Small 218; both at 60. Usnea HisPIDA Motyka, Lich. Gen. Usnea Stud. Monog. 589. 1938. Type: Cameroons, “Huea” (lapsus calami for Buea?), corticole, Lem percb. Thallus probably pendent, rather rigid, up to 15 cm. long, drying sepia to bone brown, the ramuli lighter, cinnamon buff; holdfast black, thin, very rugose, 5—6 mm. in diameter, nearly surrounding a small twig; base very short, rigid, sym- [Vor. 44 EE ANNALS OF THE MISSOURI BOTANICAL GARDEN podially branching into 8—10 branches which bend around the twig and hang down on both sides; about thrice dichotomous in the basal portion with internodes about 10 mm. long, rarely dichotomous near the summits; branches straight or slightly curved, with annular cracks in the basal internodes, then articulate in the middle portion of the principal axes, joints constricted at the ends, slightly inflated, 1.5—2 (-3) mm. in diameter, thinning at each dichotomy, the more slender branches annularly cracked but not distinctly articulate; surface with very dense long papillae with a few areas sparsely so; ramuli subdistant on the larger joints, rather close on the more slender branches, 2-10 mm. long (some longer and themselves ramulose may be young lateral branches), the shorter smooth, the longer low- tuberculate, tips acute to acuminate; summits short, slender with subterminal apothecial initials. Apothecia abundant, mostly about 5 mm. in diameter, a few up to 10 (-15) mm., sessile about one third the way from the tip of stout ramuli or small lateral branches, cupuliform becoming plane or rarely revolute; exciple smooth or in very large apothecia with a few radial ridges and sparsely tuberculate; marginal cilia few, radiating, 3-4 mm. long, quite fragile, tips acute; disc concave to plane or somewhat convex, densely white-pruinose, becoming snuff brown in a In Hendrickx 3679, the apothecia are much larger than usual, the marginal cilia are closer and longer (up to 20 mm. long), exciple with occasional short ramuli or large spinules on some of the larger apothecia, not on others of equal size. There are suggestions that some of the papillae grow out to form small spinules which easily break off as propagula. The colors are lighter, perhaps because more recently collected. AMEROONS: Cameroon Mt., m., Collier 1169, ex herb. Bot. Gard. Victoria; in Wiyar Crater, 1930 m., on Ke, piso Ted, R. 91 o ex herb. Bot. Gard. Victoria; above Buea 1290 m., rauiülico le, F. W. H. M eg 53; all at K FERNANDO Po: [Santa Isa bel Peak], 1 .» Gustav ideal 682 (two plants, tangled ith summits of U. speciosa ys pronto Sa tae Ne glued to sheet, making it diffi- caf to trace branching) ,at THOMÉ: Gus v Mem at Kew. a nee Route Kshusi, 2450 m., on Hagenia, F. L. Hendrickx 3679, in E. African er 61. UsNEa ocHRoPHoRA (Stzbgr.) Motyka, Lich. Gen. Usnea Stud. Monog. 254. 1936. Usnea florida v. ochrophora Stzbgr. in herb. Type: South Africa, Cape of Good Hope, Talbe Mt., Peter MacOwan. Thallus erect, rather rigid, up to 10 cm. tall, ashy green, drying between Sac- cardo’s umber and buffy brown on main axes, lateral branches and ramuli wood brown to avellaneous; holdfast a dark disc about 2.5 mm. in diameter, base 1 mm. in diameter below expanding to 1.5 mm. above, 2 mm. tall, branching sympodial with short internodes in the lower half; axils of lower branches 90°, upper acute; 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 45 main axis 1.5 mm. in diameter below, tapering to 0.5 mm. below the subterminal apothecium; subarticulate, joints very slightly inflated, annular cracks wide, ex- posing the medulla; lateral branches similar but only 0.8 mm. in diameter, slightly constricted at the base, quite closely and minutely papillate throughout, a few papillae growing out as short spinulose ramuli; ramuli subdistant, about 5 mm. long, tips acuminate, paler, bases slightly constricted and articulate, slightly in- flated and minutely tuberculate, rarely with a secondary ramulus, rather fragile. Apothecia 4—5 mm. in diameter, sessile to subterminal on the main axis and lateral branches, cupulate becoming plane; exciple smooth to slightly rugulose with a few papillae near the point of attachment; marginal cilia subdistant, radiating, 2-7 (-10) mm. long, sometimes forked near the base, similar to ramuli; disc concave to plane, chalky-pruinose. S sch AFRICA: Cape of Good Hope, Hermansdorp District, Blue Lilies bush, A. d ND: — plateau, 2340 m., ramicole, L. J. Brass 17212, in N. Y. Bot. Gard. riesen Nyasaland Exp TANGA A S Par e Mts., Mt. Mtonto, 2100 m., ramulicole, in evergreen forest of steep p eae P. 3 Ü owe 6532 : Mt. Kahusi, ca. 2700 m., ramulicole, F. L. Hendrickx 4300, in E. African Herb. KENYA: Mt. Kenya, 2250-257 n bamboo twigs, G. M. Allen (only 4 cm. tall, cinnamon buff to pinkish buff, BERT wholly mee Kigezi, Mafuga, 2410 m., Dale L41 p. p. min. (only 4.5 cm. tall with GANDA: single lateral branch and immature apot d ERITREA: without locality, F. e i ex ud Sbarbaro sub U. plicata, at Farlow. 62. UsNEA suUBFOVEATA Dodge, Ann. Mo. Bot. Gard. 43: 395. 1956. Type: Uganda, Bunyoro, Busingoro, on bark of Jacaranda, I. R. Dale Loo, at Kew. Thallus SE erect, rigid, 5 cm. tall, between buffy olive and yellowish olive when dry; holdfast 3 mm. in diameter, not blackened, giving rise to three main axes, 1—1.5 mm. in ECH closely dichotomous, branches slightly curved, axils 60-90°; articulate, joints about 10 mm. long, constricted at the ends, often bearing one or more lateral branches to a joint, more slender than the main axis, terete, subfoveolate from ridges of small papillae, many of which produce 1 or 2 spinulose isidia which break off, leaving a decorticate tip to the papilla; eramulose (small fragile lateral branches sometimes simulate ramuli). Apothecia subterminal on the main axis and lateral branches, up to 9 mm. in diameter (often 2 develop less than 1 mm. apart so that the exciples are in contact but not concrescent), cupulate becoming plane, smooth to slightly impressed and verrucose or the larger ones few-ciliate; marginal cilia distant, up to 1.5 mm. long, annularly cracked and slightly constricted at the base, inflated in the middle and tapering to a subacute tip; disc concave to plane, smooth (deeply rng from folding of the thecium, teratological ?), ochraceous buff with a thin w pruina. [Vor. 44 46 ANNALS OF THE MISSOURI BOTANICAL GARDEN Cortex 30 & thick, outer half brownish, closely fastigiate, inner half hyaline, hyphae more loosely interwoven, lumen about 2 u; the whole in a gel; algal layer 15-20 un thick, continuous, cells protococcoid, 6-8 un in diameter; medulla 360 p thick, the outer 35 u densely woven, the rest almost wholly radial, of about 36 hyphae per circumference, dichotomous outward, very thick-walled, 5-6 p in diameter, lumen 1 y; axis 220 y in diameter, solid, hyaline, of conglutinate, thick- walled hyphae. This species is suggestive of U. cristata, which is more angled with a much larger hollow axis. U. complanata (FOVEATAE) has a thicker, complanate thallus without articulations or isidia and is partly ramulose. Turrall 55 p. p. min. is ap- parently a small moribund specimen, only about 2 cm. tall with apothecia 2 mm. in diameter. SUDAN: without locality, J. T. Bent, at Kew KENYA: Northern Frontier Province, summit of Ajao hill, 1030 m., near Buna, on Mere I. R. Dale L71; Mau Forest, Endabarrá, 2350 m., at Sieg of Acacia grove, on trees, e a Bally B4947, a small fragment ; all at Kew ANDA: Bunyoro, Busingoro, on bark of old Juenada tree, I. R. Dale L6g, type, at ew. i TANGANYIKA: EEN 2900—3045 m., on giant heather in clearings, R. G. Turrall 55 p. p. min., at Kew 63. USNEA TRACHYNA Motyka, Lich. Gen. Usnea Stud. Monog. 612. 1938. Usnea barbata v. aspera Müll. Arg., Flora 73:336. 1890, non alio loco. Type: Tanganyika, Pagani, von Hoebnel, at Turku. Thallus in type specimen 7 cm. long (incomplete, in other specimens cited by Motyka shorter, fruticose, in our plants about 6 cm. tall), somewhat prostrate, subflaccid, ashy olive green (fide Motyka, in our plants warm sepia to Verona brown, ramuli wood brown) ; holdfast a disc 2 mm. in diameter, scarcely darkened, base 3 mm. tall, 1 mm. in diameter below, expanding to 2 mm. just under the first branch; branching sympodial or more often unilateral, the axis and first branch curving downward and apparently nearly prostrate, the other branches erect, axils nearly right angles; articulate, joints variable in length, ends constricted to 1 mm. in diameter or less, inflated to 1.2-1.5 mm. below, to 1 mm. just below the apothecia; terete, papillae subdistant below, rather close above, small, short- cylindric, or larger, subcapitate or somewhat coralloid (perhaps from regeneration after small ramuli have broken off as the tips are more or less decorticate) ; ramuli quite close, mostly broken off on the older joints, 1-10 mm. long, the shorter smooth, but annularly cracked, the longer tuberculate to small-papillate, the longest with secondary ramuli (probably young lateral branches). Apothecia subterminal on lateral branches (often appearing terminal when the end of the branch has broken off), 5-8 mm. in diameter, nearly plane, but margin somewhat inrolled and folded in drying; exciple smooth, subfoveolate and partly minutely papillate on the larger ones; marginal cilia subdistant, 3-4 mm. long, 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 47 similar to ramuli but very fragile and mostly broken off, leaving only a few short young cilia; disc plane, pale olive buff pruinose. Ellis 4 is rather short, 3—4 cm. tall, with a thinner cortex (40—55 p) but agrees otherwise. Dale L40 is 4 cm. tall, more erect and densely ramulose, although it is abundantly fertile with apothecia up to 7 mm. in diameter. ERITREA: od Pass, P. E. E 4, at Kew. uge lu Hills, H. D. van Someren, in E. African Herb. GANDA: ee Avi oral 1630 m., I. R. Dale L40; Kigezi, Mafuga, 2410 m., I. R. Dale He A b.; both at ANG nja MES Marangu-Bismarck, 2100 m., R. G. Turrall 37, at Kew; analy Hildebrandt, ex and Sbarbaro, at Halow Hi erb. 64. USNEA scuTATA Motyka, Lich. Gen. Usnea Stud. Monog. 602. 1938. Type: Cameroun Francaise, Ntem, 950 m., Ledermann. Thallus fruticose, rigid, dirty glaucescent olive green (light brownish olive shading to Isabella color on slender branches and ramuli in our plants), attached by rooting fibers which penetrate the outer layer of the bark, without a visible holdfast, base about 3 mm. tall, 1 mm. in diameter below, expanding to 2 mm. at the first branch, rarely unbranched below, tapering gradually to 1 mm. in diameter below the apothecium; once or twice dichotomous above the base with very short internodes, then unequally dichotomous above, axils about 90°; articulate, the joints inflated, ends constricted, rather long, bearing 3—4 lateral branches 1 mm. or less in diameter, usually terminated by apothecia, a few sterile ones ending in a tuft of 3-4 slender branches about the size of the larger ramuli but minutely ramulose; terete, verruculose-papillate, very few papillae on the lower joints, denser above on low reticulate ridges and the surface appearing subfoveolate, some of the more cylindric papillae growing out as minute ramuli, leaving scars as they break off, but not sorediose; ramuli irregularly disposed, rather dense on some branches, distant on others, 2—3 mm. long, tips obtuse, annularly cracked and subinflated, smooth, some probably developing as small lateral branches, minutely papillate and minutely ramulose Apothecia subterminal, cupulate becoming plane, up to 15 mm. in diameter; exciple densely verruculose-papillate; marginal cilia in two ranks, distant, 2-3 mm. long, resembling ramuli; disc flesh color, covered by a dense white pruina. Our material agrees well with Motyka’s description except the apothecia are only 3-5 mm. in diameter and the disc is concave. Deighton 3503 has more dense ramuli and the papillae in less distinct rows. Allen 1779 from Kenya is larger in all dimensions but otherwise agrees. Thallus 10 cm. tall, rather flaccid; apothecia up to 23 mm. in diameter; cortex 130 y thick, highly gelified, outer 20 p deep brownish, of very loosely woven, predominantly radial hyphae, denser just below the colored zone; algal layer about 70 p thick, of colonies of Trebouxia, forming a nearly continuous layer; medulla 670 p thick, of very loosely woven thick-walled hyphae 6-7 p in diameter, predominantly radial and somewhat more densely woven . in a narrow zone just below the algal layer; axis 410 p in diameter, solid, of longi- [Vor. 44 48 ANNALS OF THE MISSOURI BOTANICAL GARDEN tudinal, relatively thick-walled hyphae, lumen about 3 p, becoming brownish in age. The large size is suggestive of U. ruvidescens, but the proportions of cortex, medulla and axis are nearer those of U. scutata, as are the conspicuous papillae. SIERRA LEONE: Lei, Sankon Biriwa, 1930 m., F. C. Deigbton 3503, at K ANGOLA: Northeast Lunda, ipee: Ani m., near the Luachima kenn on ian trees of: the gallery wood; E Cher 13928, a ya: Lake Ngunga, 1770 m., on pee of trees, G. M. Allen 1779, in Howe herb. at Bt Herb. 65. USNEA ALBOMACULATA Motyka, Lich. Gen. Usnea Stud. Monog. 620. 1938. Type: Ethiopia, Isaac Debr., Semenbachgebirge, Stender. Thallus pendent, 7-9 cm. long, probably olive buff or lighter when living, larger branches between wood brown and dark olive buff, slender branches and ramuli olive buff or pale olive buff after 3 years in the herbarium (pale fuscous after 75 years); holdfast conic, about 2.5 mm. in diameter, of radiating, con- crescent blackened root-like fibers, base 1 mm. in diameter below expanding to 1.5 mm. above, about 3 mm. long; branching unequally dichotomous below, closer and more sympodial above; articulate and 1.5 (—2.5) mm. in diameter below, tapering to about 1 mm. and scarcely articulate below the apothecia, with minute indistinct verrucae and large pseudocyphelliform elevated white ecorticate tubercles (appar- ently from regeneration where a small branch has broken off from unusually great proliferation of the medulla, not truly farinose-sorediose), round or somewhat oblong, more closely verrucose to minutely tuberculate above and subfoveate just under the apothecia; ramuli dense on the smaller branches, 2-8 mm. long, ascending (the longer perhaps young branches as they are verruculose at the base and some- times branched above), articulate and constricted at the base, inflated, tapering to an acute tip. Apothecia sessile on the smaller branches as well as terminal, 4—6 (-15) mm. in diameter, concave; exciple smooth or subpapillate near the base with a few small ramuli; marginal cilia close, thick, conic, slightly constricted at the base with acute tips, in two rows, the outer longer, sometimes branched, the inner shorter, simple; disc concave, flesh color, slightly white pruinose Chipp 57 p. p. is tentatively referred here sce the surface is almost papil- late, the ramuli are less dense, and the white pseudocyphelliform tubercles are smaller and rarer. FO: Simien, Minden. ramulicole, H. Scott a 3, at Kem, DAN: Onyir o, Issore, 1 770 m, 3. 50. N, 32^ 50 E. Chipp 57 p. p.; Mt. Eft on aed and cairn, 3360 m., T. F. Chipp 64, Sedi: sc with U. maculata Stirton, both at 66. USNEA BORNMUELLERI Steiner, Verhandl. Zool. Bot. Ges. Wien 53:227. 1903. Type: Cameroons, Cameroon Mt., Alfred Bornmueller, portion of cotype in Howe herb. com. Zahlbruckner, at "RPM Herb. 1957] DODGE——LICHENS OF TROPICAL AFRICA. II. USNEA 49 Thallus suberect to prostrate, 5-8 cm. tall, somewhat flaccid, "gelblich- graugrun” (Milbraed) when fresh, drying olive green (Motyka) but brownish olive below shading to Isabella color in our plants; holdfast subhemispheric, 3 mm. or more in diameter, black, the root-like fibers broken in collecting; main axis 1 mm. in diameter, rapidly expanding to 2 mm., closely dichotomous near the base, then less frequently so in the upper portion and again very closely so at the sum- mits, axils nearly 90° angles; branches somewhat curved, articulate, joints inflated up to 2.5 mm. in diameter in lower portion, tapering to 0.5 mm. just below the tips, 5-12 mm. long, terete, cracked between the joints but not separated (unless roughly handled in collecting); surface irregularly verrucose, sometimes in short ecorticate ridges on the lower joints, with somewhat larger, low tubercles on the upper joints, producing isidiose soredia which soon break off, leaving white puncti- form scars, the larger ones resembling small farinose soredia; summits densely branched, somewhat deformed by small tubercles, not soraliate, tips blackened (much less so than in species of the subgenus Neuropogon). Cortex 60 p thick, brownish, gelified, very hard, fastigiate, lumen of hyphae about 1 u; algae in scattered discrete colonies, about 35 p in diameter, cells proto- coccoid, 6—7 p in diameter; medulla about 980 yq thick, of very loosely woven hyphae 6—7 p in diameter, lumen 1 p, periclinal next the algal colonies and cortex, radial within; axis 200 p in diameter, outer surface slightly irregular, of thick- walled longitudinal, conglutinate hyphae. CAMEROONS: Cameroon Mt., grasslands, Alfred Bornmueller, portion of type, com. Zahlbruckner in Howe herb. at Farlow Herb.; mountain above Buea, 3900 m., saxicole, in grasslands above tree line, J. Milbraed 10913, at Kew; 1500 m., F. W. H. Migeod 144, at ew; 400 m., Pollier 1169bis, at Kew. 67. Usnea Frotown Zahlbr., Bot. Jahrb. [Engler] 60:541. 1926. Usnea barbata v. cornuta Flotow, Linnaea 17:16. 1843, p. p. Type: Ethiopia, Mont. Simens., 3000—4000 m., saxicole, Schimper. Thallus caespitose-fruticose, suberect, about 10 cm. tall, rigid; base hemi- spheric, 5 mm. in diameter, sympodially branched about 2 mm. above, forming prostrate principal axes from which arise erect lateral branches on the upper side, very close in the proximal half of some axes, less so in others; lateral branches di- chotomous, some once, others several times, especially in the upper portion, making a very intricate mass; main axes and secondary branches 2 mm. in diameter, taper- ing at each dichotomy to short, slender summits; articulate on main axes and secondary branches in lower portions, joints constricted at the ends, somewhat inflated, simply cracked above, surface closely and minutely white-tuberculate except the basal joints, the tubercles often in rows, forming reticulations or, where longitudinal, appearing slightly angular, mostly forming soralia of isidiose soredia; most axes and secondary branches eramulose, a few densely to sparsely ramulose; shorter ramuli 1-2 mm. long, nearly smooth, the larger 10-15 mm. long, coarser, [Vor. 44 50 ANNALS OF THE MISSOURI BOTANICAL GARDEN more fragile, a few branched, closely tuberculate to subpapillate, producing soredia; summits short, obtuse, rugose, farinose from soredia. Jarrett 402 is a single moribund fragment Closely related to this species but may be new. It is in too poor condition to describe; it is arboricole and comes from a much lower altitude. OPIA: Simien, Mindigabsa, H. ee 323 c, moribund fragments, portions in Howe a. w A: Chyulu hills, H. D. van Someren, i in E. African Herb. DA: Kigezi, Mofievura: 4230 m., saxicole, non. mosses, A. S. Thomas III4; Ruizi mpe 1390 m., arboricole, T. Jarrett e both at 68. USNEA SERPENTARIA Motyka, Webbia 8:392. 1952. Usnea Flotowii subsp. serpentaria Motyka, Lich. Gen. Usnea Stud. Monog. 625. 1958. Type: Ethiopia, Mont. Simens. near Dibell, saxicole, 3000—4000 m., Schimper. Thallus erect, forming dense tufts, 9 cm. tall from a common holdfast, up to 9 mm. in diameter with dense dichotomously branched root-like fibers about 2 mm long; branches simple or dichotomous just above the base, often again dichotomous about 1 cm. above the base and again about 5 cm., then more frequently dichoto- mous, the ultimate branches sparingly ramulose; base blackened to the first di- chotomy, shading upward from Saccardo's umber through buffy citrine to old gold at the summits and ramuli; branches 2 mm. in diameter below, slightly flexuous, thinner at each dichotomy, ultimate branches and ramuli 0.2-0.3 mm. in diameter at the base, tapering upward to obtuse, dark brown tips, not articulate nor inflated, subfoveate with low flat ridges, cortex subareolate but not flaking off on the larger branches, the smaller branches less rugose and sparsely papillate, the papillae short and broad, forming minute pseudocyphellae on all but the slender ramuli; ramuli rare or absent on some branches, unilateral on others, varying from 2 to 20 mm. vis curved upward. Only one moribund subterminal apothecium present in our plants. KENYA: Chyulu hills, H. D. van Someren, in E. African Herb. 69. USNEA caPENsis Motyka, Lich. Gen. Usnea Stud. Monog. 321. 1937. Type: S. Africa, Orange Free State, Taaboschkranz, Rebman. Thallus caespitose-fruticose, erect, about 5 cm. tall, pale ashy olive green (buffy olive in our plants); holdfast a black cone, 4 mm. in diameter, 2 mm. hi gh, ad- herent to the rock, giving rise to about two axes 1.5 mm. in diameter, expanding to 2 mm. at the lowest dichotomy, rather closely but unequally dichotomous — throughout; articulate, joints constricted at the ends, subinflated below, relatively ` long, tapering upward but relatively thick to the summits, subareolate below, smooth above (very faintly verrucose under 12X magnification) ; ramuli very dense but very fragile, up to 5 mm. long, tips acute, often blackened but more 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 51 often broken off and then digitiform, slightly constricted at the base, subinflated, densely beset with minute spinules with blackened tips near the summits; summits finally curved, 0.2—0.3 mm. in diameter, grossly tuberculate with isidiose soredia, appearing farinose after the soredia are shed. Apothecia not seen. Thomas 3631 is smaller, about 4 cm. tall, and very few of the sorediose re- curved summits have been formed, evidently the decadent spinules serving as isidia. The Forbes specimen from Nyasaland is 5 cm. tall and more sparingly branched; it may be a juvenile form of U. transvaalensis, although it agrees in other vegetative characters with U. capensis. UGANDA: Karamoja, Napak, 2250 m., on rocks in mountain grassland, A. S. Thomas 3631, in E. African Herb. e egen Mt. Mlangi, 1930-2250 m., A. P. S. Forbes (6 July 1939), in E. African erb. SOUTH AFRICA: Cape of Good Hope, without locality, Mund & Maire, in Tuckerman herb. sub U. barbata v. hirta, at Farlow Herb. / 70. USNEA PULVINATA Fries in Lehmann, Pl. Preiss. 2:145. 1846-47. Type: Western Australia, Preiss. Thallus up to 3.5 cm. tall, erect, fruticose, very rigid, pale greenish, drying olive ochre to honey yellow in the herbarium; holdfast of coarse radiating rootlets penetrating between the rock crystals, somewhat blackened and verrucose; branch- ing into three or more primary axes about 1 mm. in diameter, lowest internode variable in length, up to 8 mm., then twice or thrice closely dichotomous and rarely so above, angles about 45? and the branches curving upward, terete, surface smooth except for scars of broken ramuli; ramuli variable in length, 1-10 mm. long, slightly constricted at the base and inflated in the middle, about 0.2 mm. in di- ameter, curving upward and tapering to obtuse tips, some growing out as slender lateral branches and again short-ramulose, surface tuberculate, forming small soralia of isidiose soredia, usually black-tipped and appearing farinose after the isidia are shed, in old plants the soralia more or less confluent, giving a furfuraceous appearance under 12X magnification. Apothecia absent in our material, reported by Motyka as terminal, about 3 mm. in diameter; exciple smooth, sometimes with tubercles; marginal cilia short, dense, thick and obtuse; disc somewhat concave, yellowish flesh color, slightly pruinose. Cortex about 110 p thick, fastigiate and brownish in the outer 30 y, loosely and irregularly woven below, imbedded in a hyaline gel; algal layer of discrete colonies of Trebouxia, about 30 p. in diameter, forming a nearly continuous layer, cells 4-7 4 in diameter, somewhat polyhedral from mutual pressure; medulla 75- 80 p thick, white, K—, densely woven, of predominantly periclinal hyphae 6-7 u in diameter, lumen 1 p, inner 10-15 » brown next the axis; axis 700 p in diameter, outer 200 u of interlaced transverse hyphae, predominantly periclinal, the central cylinder predominantly of longitudinal hyphae with more slender lumina and much more transparent so that under lower magnifications the axis appears hollow. [Vor 48 32 ANNALS OF THE MISSOURI BOTANICAL GARDEN | The Robillard plant is somewhat doubtfully referred here, as the cortex and | axis are thinner and the soredia are less developed. The axis is uniformly composed of longitudinal hyphae. MADAGASCAR: Imerina, Andrangoloaka, saxicole, Hildebrandt 2172 (3 collections), at | Farlow Herb., Province < Betafo, native collector, spring 1901, ex herb. E. G. Paris, 1904 (2 Keen) ,at Farlow Herb. AURITIUS: without a Robillard ex Gray Herb. sub U. florida v. strigosa Ach., at Farlow Herb. 71. Usnea MYRIOCLADA (Müll. Arg.) Zahlbr., Cat. Lich. Univ. 6:588. 1930. Usnea barbata v. myrioclada Müll. Arg., Bot. Jahrb. [Engler] 20:244. 1894. Type: Tanganyika, Usambara, Derema, Fischer. Thallus erect, subfruticose, about 5 cm. tall, pale glaucescent green (Motyka, axes between buffy brown and citrine drab, ramuli between dark olive buff and deep olive buff in our plants) ; base of coarse rooting fibres penetrating dead wood; closely sympodially branched at the base and frequently so throughout, axils nearly | right angles, becoming nearly dichotomous below the summits; main axes sub- - articulate, joints very slightly inflated, rather long, very densely ramulose below, ` appearing smooth between the ramuli but slightly verruculose under 36X magni- - fication; ramuli mostly 1-2 (—5) mm. long, slightly constricted at the base, an- - nularly cracked, slightly inflated, fragile, many broken off at the base, giving a coarsely verrucose appearance, abruptly tapering to an acute black tip; slender branches and summits nearly devoid of ramuli, closely tuberculate, producing | soralia of isidiose soredia which may appear farinose after the soredia are shed; summits of the main axes all broken off, those of slender branches densely coralloid, | branched. | A fragment of another plant in the collection is fertile; the vegetative portion is identical with that of the sterile plant. Apothecia sessile on slender branches, 3 mm. in diameter, cupulate; exciple densely ramulose with short conical ramuli, appearing subfoveolate between them; marginal cilia erect, very dense, in two rows, | less than 1 mm. long, resembling small ramuli; disc concave, flesh color, densely white-pruinose. In Gossweiler 9991, the summits are nearly intact, tapering to acute tips, not coralloid. : OLA: Cuanza 3u, Amboin near Carloaongo-Cuvo River, 1000 m., on dead trees, T Gocwiil 9991, at K : RHODESIA: ene District, 1320 m., on dead wood, Frederick Eyles 419) 72. UsNEA Lynce! Motyka, Lich. Gen. Usnea Stud. Monog. 411. 1937. Type: St. Helena, near tomb of Napoleon, collector unknown, at Oslo. 3 Thallus pendent, rather flaccid, about 20 cm. long, pale ashy (main axes "m brown below, upper portions and ramuli deep olive buff in our plants); base about 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 33 10 mm. long, 0.9 mm. in diameter, lower part black, closely and regularly annularly cracked; closely dichotomous at nearly right angles just above the base, then di- chotomous at acute angles with relatively short internodes throughout, minutely tuberculate throughout, each white tubercle bearing a few isidia; ramuli somewhat irregularly disposed, variable in length up to 30 mm. long, flexuous, slender, annu- larly cracked, not tuberculate or very minutely so at the base of the longer ramuli, tips acuminate but mostly broken off. A single apothecial initial seen, lateral at the base of a large ramulus, but heavily infected by a parasite before the initial reached 1 mm. in diameter. War- necke IO7 is too ote gie for certain determination. TOGO: viue I07, in E. African Herb. FERNANDO PO: [Santa pue Peak], 2900 m., Gustavo Mann 682 p. p. min., at Kew. ST. HELENA: without locality, J. D. Hooker, in Taylor herb. at Farlow Her b. 73. UsNEA GosswziLERI Dodge, Ann. Mo. Bot. Gard. 43:395. 1956. Type: Angola, Cuanza Sul, Amboim, Capir near the Carloaongo River, 1000 m., on dead trees, J. Gossweiler 9991, at Kew. Thallus erect, fruticose, 5—6 cm. tall, between deep olive buff and tea green; holdfast about 4 mm. in diameter, 2 mm. thick; polychotomous at the top of the holdfast into 7-8 branches about 1.5 mm. in diameter, again densely branched 7-10 mm. above, then rarely branched sympodially in the mid-portion, branches about 1 mm. in diameter, dichotomously branched above, thinning at each dichot- omy, annularly cracked but not inflated; surface closely verrucose-tuberculate, forming isidiose soredia, appearing spinulose where the tubercles are crowded and finally farinose after the isidia are shed; ramuli 2-3 (—5) mm. long, irregularly dense, the shorter smooth, the longer minutely tuberculate with acute, often black- ened tips; sterile summits short, relatively thick, but tips acute. Apothecia rare, sessile on the main axes some distance below the summits, or subterminal, 3 mm. in diameter (perhaps immature); exciple smooth; marginal cilia subdistant, woe cracked at the base, 1-2 mm. long, conic, the longer sometimes dichotomous and minutely tuberculate; disc carneous, white-pruinose. Cortex 85-100 yu thick, fastigiate, highly gelified, hyphal lumina subdistant within, repeatedly dichotomous in the outer 30 p where the gel is more deeply stained but without crystals; algal layer 40-60 p thick, of discrete colonies of Trebouxia, about 25 p. in diameter, cells 5-6 p in diameter, sometimes scattered, sometimes aggregated into a nearly continuous layer; medulla about 160 p thick, of thick-walled hyphae about 3 y in diameter, very densely interwoven and heavily nubilated with hyaline or slightly brownish crystals, very lax in the inner 25 p; axis 230 & in diameter, of thick-walled longitudinal hyphae, conglutinate in the Outer portion, less compact toward the center, with longitudinal air spaces up to 10 1 in diameter between the strands of hyphae. : Cuanza Sul, pu Capir near the Carloaongo River, 1000 m., on dead trees, "x Ge reg at Kew [Vor. 44 54 ANNALS OF THE MISSOURI BOTANICAL GARDEN 74. UsNEA TORRIDA Motyka, Lich. Gen. Usnea Stud. Monog. 327. 1937. Type: Congo, without locality or collector, in British Museum. Thallus about 5 cm. tall, erect, rather rigid, citrine drab; holdfast penetrating cracks of bark on twigs; base up to 5 mm. long, about 1 mm. in diameter, scarcely blackened; branching once to thrice dichotomous at the base, occasionally so above; branches variously curved upward, slightly constricted at the base, soon about 1 mm. in diameter, tapering upward; surface smooth between the very dense ramuli; ramuli 0.5-2 mm. long, tips acute, straight or slightly curved upward, occasion- ally growing out to 5 mm. long as small lateral branches, again sparingly ramulose; on the lower portion the ramuli break off so that the lower internodes appear tuberculate to papillate. Apothecia terminal, up to 3 mm. in diameter, cupulate becoming nearly plane; exciple closely small-ramulose (not as densely as on the branches) ; marginal cilia dense, 3—4 mm. long, annularly cracked and bearing a few ramuli; disc concave, flesh color, densely white-pruinose. The Nyasaland plant is a fragment, less densely ramulose but it agrees otherwise. ENYA: Northern Frontier Nu. summit of Ajao hill, 1030 m., near Buna, on bushes, I. R. Dale L71 p. p., at Kew. ANGOLA: Northeastern "een? Chitato, Dundo near Rio Luachima, 750 m. (Expl. Companhia de Diamantes de Angola), on tall trees, J. Gogo 13928; Benguella, country of ds Ganguellas and Ambuellas, J. Gossweiler; both at Kew. NYASALAND: Mlanje District, Luchenya M Mt. Mlanje, 1890 m., L. J. Brass 16846 (Vernay Nyasaland Exp. 1946), in N. Y. Bot. Gard. Herb., a fragment tangled with U. Vainioana Zahlb 75. UsNEa PICTA (Steiner) Motyka, Lich. Gen. Usnea Stud. Monog. 325. 1937. Usnea ceratina v. picta Steiner, Sitzungsber. K. Akad. Wiss. Wien, math.-naturw. CL 106:1:210. 1897. Type: Kenya, Machakos, corticole, Liechtenstein. Thallus fruticose, rigid, ashy green, 8—10 cm. tall, base very rigid, several times subsympodially branched above the base, axils nearly right angles; branches irregu- larly divergent, up to 2 mm. in diameter throughout most of the length, tapering gradually to short, branched summits, nearly straight, slightly constricted at the bases, with verruciform papillae, obtuse, denser in the upper portion of the branches; ramuli rather thick, rigid, aculeiform, perpendicular to the branches, tips acute, often branched Apothecia abundant, terminal, up to 20 mm. in diameter, nearly plane and densely pruinose when young, becoming convex, often plicate-rugose in the center when old; exciple smooth, glabrous; marginal cilia rather rare, 5-10 mm. long, rigid, conical; disc the same color as the thallus, slightly pruinose; epithecium yellowish; ascospores 6 X 10 p. 1957] DODGE— LICHENS OF TROPICAL AFRICA. II. USNEA 55 Cortex about 100 p thick, very hard, chondroid, dilacerate within and perforate in places; medulla about 200 p thick, dense, K— or yellow next the axis; axis 420 p in diameter, somewhat effigurate. I have not seen material certainly referable here. The material Motyka referred to U. picta is evidently variable, as three of the six specimens cited are annotated as not typical. U. isabellina Dodge from the Northern Frontier Province of Kenya is related to U. picta; perhaps the same as the specimen cited by Motyka from southern Usambara, H. Meyer, "forma thallo substramineo, haud exacte certa." The Nyasaland plant is closer to Motyka's description but the axes are up to 3.2 mm. in diameter, apothecia only 5 mm. in diameter and the axis 2650 y in diameter, quite lacerate but not distinctly hollow. I hesitate to describe another new taxon in this group until I have seen more material. NYASALAND: Zomba Plateau, 1820 m., saxicole, L. J. Brass 16166 (Vernay Nyasaland Exp.), in N. Y. Bot. Gard. Herb. 76. USNEA ISABELLINA Dodge, Ann. Mo. Bot. Gard. 43:395. 1956. Type: Kenya, Northern Frontier Province, near Buna, summit of Ajao hill, 1030 m., on bushes, I. R. Dale L71, at Kew. Thallus erect, fruticose, rigid, 7 cm. tall, Isabella color below, shading to chamois above; holdfast 8 mm. in diameter, giving rise to about 10 main axes 1—1.8 mm. in diameter, some branching almost at the base, some up to 20 mm. to the first branch, infrequently dichotomous to subsympodially branched above, terete; cortex smooth, closely annularly cracked and subareolate below, less so above, appearing coarsely papillate from the ramuli breaking off about 0.2 mm. from their bases; ramuli dense, quite fragile, 2-4 mm. long, coarse, slightly verrucose but not be- coming sorediate, tips often slightly blackened, some longer ramuli growing out as small lateral branches; ultimate branches up to 40 mm. long, tapering slowly, tips broken off. Apothecia rather rare, about 5 mm. in diameter, sessile some distance below the tips of the branches, plane to subrevolute; exciple smooth to slightly rugose; marginal cilia fragile, of variable length up to more than 3 mm. long, resembling ramuli; disc avellaneous, slightly white-pruinose. Cortex 40-50 » thick, fastigiate, forming a relatively thin-walled pseudo- parenchyma, cells 5—6 p in diameter, the outer 25 p heavily nubilated with minute brownish crystals; algae Trebouxia, in small discrete colonies, some up to 40 L in diameter, cells 6—8 p in diameter; medulla 65—80—120 p thick, of slender (3 pin diameter) hyphae very densely interwoven and heavily nubilated with minute brownish crystals; axis 500 un thick, of predominantly longitudinal, thick-walled hyphae about 3 p in diameter, lumen 1—1.5 p. : This species seems closest to U. picta (Steiner) Motyka, and perhaps the speci- men cited by Motyka from Tanganyika, southern Usambara, H. Meyer “forma thallo substramineo, haud exacte certa” may belong here. U. isabellina differs from [Vor. 44 56 ANNALS OF THE MISSOURI BOTANICAL GARDEN U. picta in color, size, of parts, thinner cortex and medulla and much thicker axis. KENYA: Northern Frontier Province, near Buna, summit of Ajao hill, 1030 m., on bushes, I. R. Dale L71, at Kew. 77. UsNEA HorRIDULA (Müll. Arg.) Motyka, Lich. Gen. Usnea Stud. Monog. 335. 1937. Usnea longissima v. horridula Mull. Arg., Flora 74:372. 1891. Type: South Africa, Moorn River near Port Natal (Durban), Plant. Thallus pendent to creeping, up to 30 cm. long, shorter in our plants, olive green drying Mars brown to russet on the main branches, ramuli and tips of branches cinnamon buff, reddish-maculate; base conic, about 2 mm. in diameter, composed of black rooting fibers penetrating the bark, 7 mm. tall, 1.5 mm. in diameter, branching dichotomous at nearly right angles, then curving downwards and flex- uous but not parallel, tapering gradually to the short capillaceous summits; cortex cracked and areolate below, verrucose, becoming papillate above the second axil, many papillae growing out to form very dense ramuli about 2 mm. long, annularly cracked at the base, minutely tuberculate, forming soredia, others smoother, up to 7-8 mm. long (perhaps young branches), rather fragile, leaving slightly elevated farinose scars; summits about 10 mm. long, densely dichotomous, slender, tapering to acute tips. Cortex very variable in thickness in the same section, 80—135 p, the outer 55 p of predominantly periclinal branched hyphae, rather closely packed, reddish, the rest more loosely woven, all imbedded in a yellowish gel, paler next the algal layer; algal layer of discrete colonies about 35 p in diameter, cells Trebouxia, about 8 u in diameter, in some places the colonies quite close, in others distant and dying off; medulla about 210 » thick, the outer 70 p pale, of predominantly periclinal hyphae, the rest of very densely woven, very thick-walled hyphae 3 p in diameter, obscured by brownish crystals; axis 600 & in diameter, very dense and hyaline with scattered brownish areas about 35 p in diameter, due to minute brownish crystals which obscure the structure. KENYA: Chyulu hills, H. D. van Someren, in E. African Herb. 78. Usnea Date Dodge, Ann. Mo. Bot. Gard. 43:396. 1956. Type: Kenya, Northern Frontier Province, near Buna, summit of Ajao hill, 1030 m., on bushes, I. R. Dale L70, at Kew. Thallus pendent, subrigid, more than 16 cm. long, probably up to 40 cm.; main axes drying deep olive buff, rarely pink-maculate in younger plants and in the more slender distal portions of older plants, the lower portions of the axes of older plants becoming orange cinnamon, the ramuli remaining deep olive buff; holdfast of root- like fibers penetrating the bark, base about 1 mm. in diameter and 2 mm. tall, terete, somewhat blackened and annulate; closely dichotomous to sympodially 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 92 branched just above the base, then rarely dichotomous with long internodes throughout; main axes about 1 mm. in diameter, tapering gradually to the tips, acutely few-angled from concrescent rows of papilliform tubercles and the com- planate bases of the ramuli, sulcate between the ridges, with a few scattered papilli- form tubercles in the furrows; the tubercles growing out as coarse spinules serving as isidia; quite closely ramulose throughout along the acute corticate ridges, mostly about 4 mm. long, nearly cylindric, smooth with rather obtuse tips, a few up to 15 mm. long, curved or flexuous, closely tuberculate with relatively large tubercles in the lower portion, each bearing a few spinules; summits broken off, but such fragments as are present suggest that they are slender, unbranched and not capillaceous. Cortex 50—55 y thick, fastigiate, of thick-walled gelified hyphae, the outer 20 y very deeply staining; algal layer 40 p thick, a continuous layer of protococcoid cells 8 u in diameter; medulla 250 p. thick, of thick-walled hyphae 3 p in diameter, very densely woven next the algal layer and the axis, very loosely woven between; axis 550 p in diameter, nearly circular in cross-section, the outer 40—50 p of longi- tudinal, conglutinate thick-walled hyphae, the rest with single coarse brown hyphae 8 » in diameter scattered throughout the sclerenchyma, giving the appearance of a section of a monocotyledonous stem such as that of maize. U. Dalei suggests U. decipiens v. Rhodesiana Dodge, which is more angled with farinose soredia and which lacks any suggestion of a reddish color on the older axes. U. Dalei is the only sulcate species of the sETULOSAE, RUBIGINEAE, somewhat par- allel to U. chloreoides (GONIODES), but is quite distinct in color and other char- acters. None of the species of the sulcate subsections, GONIODES nor SULCATAE, bear tubercles in the furrows, nor develop a reddish color. KENYA: Northern Frontier Province, near Buna, summit of Ajao hill, 1030 m., on bushes, I. R. Dale L70, at Kew. 79. USNEA ERUBESCENS (Stein) Motyka, Lich. Gen. Usnea Stud. Monog. 333. 1937. Usnea articulata £. erubescens Stein in H. Meyer, Ostafrikanische Gletscherfahrten, P. 314. 1890; [tr. E. H. S. Calder] Across East African Glaciers, 353. 1891 (with very brief description) ; Müll. Arg., Flora 73:336. 1890, nom. nud. Usnea barbata v. articulata £. erubescens Stein, Jahresber. Schles. Ges. Vaterl. Kul- tur 66:134. 1888, nom. nud. Type: Tanganyika, Kilimanjaro, at tree-line, 3000 m., H. Meyer, in Mus. Bot. Univ. Breslau. llus pendent or possibly prostrate, about 20 cm. long, flaccid, drying reddish fuscidulous, dull; base not seen, sparingly dichotomously branched, one branch usually better-developed than the other; branches irregularly curved, up to 2 mm. in diameter; articulate, joints about 1 cm. long, inflated, constricted at the ends, [Vor. 44 58 ANNALS OF THE MISSOURI BOTANICAL GARDEN alternating with short cylindrical joints 0.3 mm. long, some of the larger joints deformed, rugulose and foveolate, almost epapillate, with minute tubercles along the ridges, other large joints smooth; ramuli absent; summits closely branched. Apothecia and soredia not seen. Cortex about 30 yp thick, soft, fuscidulous; medulla up to 400 p thick, very lax, white, K—, of irregularly woven hyphae; axis 120—150 y in diameter, irregular in cross section. Besides the specimen upon which Motyka based his description, Hans Meyer collected another among mosses of the grasslands at 4000 m., "mit gelbrothen Lager .. . " evidently smaller as he refers to the specimen from 3000 m. as “in sehr kräftigen langbärtigen PEERS Thomas 619 from Mt. Elgon, 4420 m., y be this entity. It is 5-8 cm. long, ochraceous tawny to ochraceous buff, ieee of long and Deeg joints present but somewhat irregularly so, summits (upper centimeter) rather abruptly more slender, not articulate, slightly tubercu- late but not sorediose and chamois to colonial buff. Only more material can show whether our plants are ecologic variants from higher altitudes or whether they represent a distinct taxon. Our plants somewhat resemble the more rigid U. terri- cola Dodge from Cameroons, which shows conspicuous soredia when much younger and comes from a lower altitude. If our reference of the Uganda material is correct and from Motyka's description, U. erubescens Motyka is much more closely related to the ARTICULATAE or the AMOENAE, in spite of its reddish color, than to the sETU- LOSAE, RUBIGINEAE, where Motyka places it. eem? Mt. Elgon, Masaba, 4420 m., on rocks in alpine meadow, A. S. Tbomas 619, at Kew. 80. UsNEA sUBFLORIDA (Zahlbr.) Motyka, Lich. Gen. Usnea Stud. Monog. 335. 1937. Usnea Steineri v. subflorida Zahlbr., Bot. Jahrb. [Engler] 60:541. 1926. Type: Tanganyika, Meru District, Fassmann. Thallus erect, rigid, fruticose, 4—5 cm. tall, roods brown to pecan brown to wood brown and avellaneous in South African specimens (under 12 magnifica- ` tion irregularly light yellowish olive with coral red spots, reddish fuscous spotted with greenish fuscous, fide Motyka); holdfast 3 mm. in diameter, hemispheric, of coarse, branched rootlets, nearly black, base about 1 mm. in diameter below expand- ing to 1.5 mm. at the first dichotomy, about 5 mm. tall; thrice dichotomous with short internodes just above the base, axes about once dichotomous above; branches straight or curved, about 1.5 mm. in diameter below, tapering to 1 mm. below the apothecia; articulate, joints long, inflated below and tapering above, constricted at the ends, terete, papillose, some papillae forming spinules which break off as isidia; ramuli rather dense, 3-5 mm. long, the longer slightly tuberculate, tips acuminate but mostly breaking off, leaving an obtuse end as the cortex regenerates. 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 59 Apothecia up to 6 mm. in diameter, subterminal but appearing terminal when they mature, the branch often passing within the exciple and the tip emerging as a stouter cilium, cupulate to nearly plane; exciple with branched radial ridges becoming subfoveate, slightly tuberculate along the ridges; marginal cilia moder- ately close to subdistant, up to 4 mm. long, radiating, straight or flexuous, annularly cracked at the base, tapering to acute tips, maculate, the larger tuberculate; disc densely white pruinose when young, becoming dark flesh color in age. ENYA: Mau forest, ae ae 2355 m., on trees at edge of Acacia grove forest, P: R. O. Bally B4947 p. p. m t Kew congo: Mt. Kahusi, US m., feral See F. L. Hendrickx 4300, small and moribund, OUTH AFRICA: Transvaal, Zoutpansberg district, near “Punch Bowl” Inn, on decay- ing ec trunks, O. Almborn, Lich Afric. 19, in Dodge Herb. 81. USNEA BicoLoRATA Motyka, Lich. Gen. Usnea Stud. Monog. 336. 1937. Type: Congo, Eastern Province, Kivu Mts., de Witte. Thallus subpendent, 12 cm. long, rigid, pale ashy olive green, irregularly red- spotted; base quite slender, rigid, blackened, slightly attenuate, infrequently sym- podially branched almost to the summits, axils 90? or nearly so; branches up to 2 mm. in diameter, rather uneven in thickness, slightly attenuate at the base, abruptly so at the summits, irregularly curved, rarely deeply cracked; more or less terete, indistinctly deformed, and slightly foveolate, papillae obtuse, concolorous, tubercles orming soralia in the upper portion of the thallus, soralia relatively broad, white- farinose, becoming eroded; ramuli quite close, of variable length, similar to branches from dichotomies; summits indistinct, dense. Apothecia not seen. Our plants from Cameroons are tangled with Usnea speciosa and U. hispida. They are nearly all Vandyke brown to Natal brown on the main axes although under 12 mignification small patches of citrine drab and dark olive buff are visible and the broken cortex is red-maculate; the medulla is as thick as the axis; otherwise they agree with the above description translated from Motyka. In Bally B4947, the soralia are apotheciiform, up to 0.5 mm. in diameter, wider than the branches which bear them. CAMEROONS: Victoria, in Meyer Crater, 2580 m., on Aquaria salicifolia, T. D. M. 017 b. re min., ex hb. Bot. gmi Victoria, at Kew. Ya: Mau forest, Endabarrá, 2355 m., on trees at edge of Acacia grove forest, P. fe O. Bally B4947, "pied to sheet of U. aequatoriana Motyka, at Kew 82. UsNra sprLotomes Dodge, Ann. Mo. Bot. Gard. 45:396. 1956. Type: Kenya, Chyulu hills, H. D. van Someren, in East African Herb. Thallus probably procumbent, perhaps pendent, subrigid, about 10 cm. long, russet below, shading through tawny to zinc orange and ochraceous salmon, with smaller ramuli pale olive buff; base not seen; branching sympodial or very closely [Vor. 44 60 ANNALS OF THE MISSOURI BOTANICAL GARDEN dichotomous in basal portion where axes are up to 0.9 mm. in diameter, tapering to short subcapillaceous summits, irregularly dichotomous above, some internodes long, others quite short, axils nearly 90°; branches very flexuous, summits mostly recurved, articulate, joints slightly inflated, ends constricted, very variable in length, rarely the cortex regenerating to form very short, darker joints up to 0.3 mm. long, but usually joints separated by a narrow crack; minutely white verru- cose, producing single or very few isidia; ramuli very irregularly disposed, some- times very close, 4 (-10) mm. long, very fragile, slightly verrucose, producing isidiose soredia, becoming minutely farinose as the soredia are shed; ramuli on more slender branches subdistant, about 1 mm. long, smooth, tips acute with long nude areas; summits mostly broken off, up to 20 mm. long, slender, tips acuminate, some nude, others very sparsely short ramulose. Cortex 55-80 y thick, of yellow conglutinate thick-walled radial hyphae 7-8 p in diameter, lumen 1 p; algal layer 40 y thick, of protococcoid cells up to 12 p in diameter, dying in places; medulla 75 & thick, the outer 35 p of predominantly periclinal, thick-walled hyphae 5—6 u in diameter, lumen 3 p, forming a pseudo- parenchyma, the inner 40 u of slender, densely woven thick-walled hyphae, heavily nubilated with minute brown crystals; axis 650 p in diameter, solid, of longitudinal, conglutinate thick-walled hyphae 4—5 y in diameter, lumen 3 p. This species belongs in the group with U. erubescens and U. spilota. Perhaps some of the material referred by Motyka to the latter species may belong here, as Motyka states that the collections from South Africa are "minus certa" and one "juvenilia, possibile ad U. borridulam non evolutam spectantia." U. horridula differs from U. spilotoides in its non-articulate, more regularly ramulose thallus with the isidiose soredia confined to upper parts of the ramuli. KENYA: Chyulu hills, H. D. van Someren, in E. African Herb. 83. UsNEA MEYERI (Stein) Motyka, Lich. Gen. Usnea Stud. Monog. 338. 1937. Usnea cornuta v. Meyeri Stein, Jahresber. Schles. Ges. Vaterl. Cultur 66:134. 1888. Type: Tanganyika, Kilimanjaro, 4000 m., saxicole, H. Meyer. Thallus 4—6 cm. tall, erect, very rigid, greenish-stramineous, reddish-maculate, main axes becoming Van Dyke brown, slender branches and ramuli clay color to cinnamon buff, summits nearly white from soredia in the herbarium; holdfasts of radiating, closely branched rootlets, spreading over the surface of the rock and penetrating between the crystals; base about 10 mm. tall, 2 mm. in diameter, cortex annularly cracked and areolate; about thrice dichotomous with very short inter- nodes (appearing polychotomous), each axis with a long internode (10 mm. or more) then again densely branched, the branches terete, curving upward to form a depressed hemispheric tuft; the lowest internode about 2 mm. in diameter, thinning at each dichotomy, the ultimate branches still about 1 mm. in diameter; the lower two internodes blackened, the upper branches clay color to cinnamon buff, testa- 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 61 ceous-maculate; surface smooth or low-tuberculate, forming soralia of black-tipped isidiose soredia becoming confluent on some of the ultimate branches; ramuli very irregularly distributed, absent on some branches, subdistant on others, about 2 mm. long, annularly cracked at the base, slightly inflated, tapering to an obtuse tip, some up to 5 mm. long, slightly tuberculate and occasionally forked (perhaps young secondary branches). Apothecia not seen. Cortex 100 y thick, so blackened that structure is not clear; algal layer about 60 p thick in discrete colonies of Trebouxia (mostly disintegrated, leaving regu- larly spaced lacunae just under the cortex, with an occasional cell or small colony remaining; medulla about 160 p thick, of mostly periclinal thick-walled hyphae about 3 p in diameter, axis 1000 p in diameter, solid, of longitudinal conglutinate hyphae interlaced with strands of transverse hyphae. KENYA: west slope of Mt. Kenya, on trail from West Kenya Forest Station to summit, 3630 m., in giant heath zone, Edgar A. Mearns 1521, 1536 (T. Roosevelt Exp.), at Farlow Herb. 83a. Var. cHONDROCLADA (Steiner) Dodge, Ann. Mo. Bot. Gard. 43:396. 1956. Usnea cbondroclada Steiner, Verhandl. Zool. Bot. Ges. Wien 53:228. 1903. Type: Cameroons, Fako, 3600—3800 m., on lava blocks, Alfred Bornmueller. Thallus 4—6 cm. tall, erect, rigid, axes Sayal brown shading to light brownish olive, ramuli Isabella color, indistinctly reddish-maculate; holdfast broken off, base black, 5 mm. tall, 1 mm. in diameter below, expanding to 2 mm. just below the first dichotomy; repeatedly dichotomous just above the base, then sparingly so above; branches 1.3—1.5 mm. in diameter, annularly cracked, not inflated below, slightly so above, thinning somewhat at each dichotomy, slightly verruculose, some verrucae growing out as small ramuli; secondary branches slender, up to 0.6 mm. in diameter, paler and some more clearly reddish-maculate; ramuli from less than 1 mm. to 5 mm. long, slightly inflated, tips acuminate, blackened; the smaller ramuli easily break off leaving minute white scars and probably serve as isidia, especially in the upper portions of the smaller secondary branches and the obtuse summits. Apothecia not seen. Cortex 65 p thick, fastigiate, gelified, outer 15—30 y dark brown, algal layer about 30 u thick, of Trebouxia, cells 7-8 p in diameter; medulla 100 p thick, densely woven under the algal layer, looser next the axis, hyphae 3 y in diameter, lumen about 1 p; axis solid, about 300 y thick, of longitudinal conglutinate hyphae interlaced with strands of transverse hyphae. CAM Ns: Cameroon Mt., Alfred Bornmueller sub U. inc f. ramillosa Steiner, e nom. 3 com. Zahlbruckner, in Howe herb. at Farlow Herb 84. USNEA MACULATA Stirton, Scottish Nat. 6:293. 1882. Usnea Flotowii v. subhispida Zahlbr., Bot. Jahrb. [Engler] 6:541. 1926. Type: S. Africa, Cape of Good Hope, Somerset East, P. MacOwan & J. M'Lea, in Art Galleries of the Glasgow Corporation. [Vor. 44 62 ANNALS OF THE MISSOURI BOTANICAL GARDEN Thallus erect, fruticose, up to 7 cm. tall, pale greenish drying pecan brown to vinaceous tawny (under low magnification spotted with deep olive buff) ; holdfast a black disc up to 10 mm. in diameter, composed of closely woven rooting fibers giving rise to erect axes, mostly unbranched below, some dichotomous just above the holdfast, blackened for 2—5 mm. at the base, sparingly unequally dichotomous in the upper portion; axes forming a dense erect tuft 10—40 mm. in diameter; sub- articulate, joints relatively long and slightly constricted at the ends, up to 1.6 mm. in diameter a short distance above the base, more slender above; surface rather closely tuberculate to subpapillate, papillae either single or concrescent in short irregular ridges, some growing out as caducous, conical spinules about 0.5 mm. long, a few growing out as fragile ramuli, 5-10 mm. long, the larger annularly cracked with obtuse tips, curved, ascending; scars from fallen spinules and smaller ramuli resembling small farinose soralia. Apothecia sessile on smaller branches, 3-4 mm. in diameter, cupulate; exciple citrine drab to deep olive buff, smooth to slightly tuberculate near the base; mar- ginal cilia dense, up to 5 mm. long, coarse, annularly cracked, slightly constricted at the base, tips obtuse, many broken off; disc concave, somewhat deformed, coarsely cream buff-pruinose. The apothecia of our plants are apparently young and agree with those de- scribed by Motyka for U. subflorida (Zahlbr.) Motyka except those of the latter are terminal instead of lateral. The vegetative portion of the plants agrees with - that of U. maculata. Perhaps when a fuller series of specimens is available, it will be necessary to unite U. subflorida with U. maculata. Some of our plants are from farther north than either species has been reported. Greenway 6532 is very young, about 1 cm. tall and sterile, but probably should be referred here. SUDAN: Mt. Kinelli, 3360 m., on pole and cairn, T. F. Chipp 64, at Kew UGANDA: Ruizi River, 1390 m., on trees in riverine forest, T. Jarrett 402 3 bo ew. LI S. Pare Mts., Mt. Mtonto, 2100 m., ramulicole, P. J. calcis bee at UTH AFRICA: without eng Drége, ex herb. W. Sonder sub U. fragilis Laurer, herb. nom.?, in Tu S erman herb. at Farlow b. AGASEARE ina, Andrangoloaka, Hildebrandt 2173, Nov. 1880, com. C. Rensch ex herb, Sbarbaro es vu barbata v. scabrosa (Ach) Müll. Arg., at Farlow Herb. 85. USNEA ACANTHERA Vainio in Motyka, Lich. Gen. Usnea Stud. Monog. 358. 1937; Type: S. Africa, Bechuanaland, Crocodile River, Klingsberg, in Riksmuseet, Stockholm. e Thallus prostrate to suberect, rigid, about 6 cm. long, ashy green, darker on the main axes; divaricately dichotomous just above the base and sparingly so below the _ summits; axes 1 mm. in diameter, sigmoid-curved, continuous, terete, surface un- | even, not papillate although very young ramuli may simulate papillae; ramuli dense, _ strigose, perpendicular, about 2 mm. long, pale green, smooth. 1957] DODGE—LICHENS OF TROPICAL AFRICA. Il. USNEA 63 Apothecia terminal, up to 7 mm. in diameter, almost plane; exciple slightly uneven or rugose, densely ciliate; marginal cilia dense, 4-5 mm. long; disc subplane, margin often broken, flesh color, almost epruinose; epithecium fusco-fuligineous; ascospores 6 9 u in narrow asci. Cortex 30 y thick, chondroid, pale green; medulla 120—140 y thick, rose color, dense, K—; axis 360 » in diameter, white with narrow central cavity, slightly spotted with a few black hyphae. Our specimens seem to be juvenile forms of this species. The branching is dense at the woody base, 4 mm. in diameter, 5 mm. tall, only once dichotomous above in about the middle of the axis, branches nearly straight; ramuli 3-5 mm. long; summits slender, the uppermost 10 mm. eramulose. Sterile but not sorediose in the Kenya specimen, a few very young apothecia present in the Uganda specimen. KENYA: Chyulu hills, H. D. van Someren, in E. African Herb. UGANDA: Kigezi, Mafuga, 2410 m., I. R. Dale L41 p. p. min., at Kew. 86. USNEA MAROCANA Motyka, Lich. Gen. Usnea Stud. Monog. 359. 1937. Type: Maroc, Camp Monod, arboricole, Mouret. Thallus 5-8 cm. tall, fruticose, pale ashy green, drying deep olive buff to olive buff; dichotomous just above the holdfast and twice or thrice dichotomous below with rather short internodes and once or twice dichotomous at the summits; nearly terete with occasional low ridges where the ramuli are close in rows; branches mostly 1.2-1.3 mm. in diameter in the lower three fourths of their length, then tapering upward to slender acute summits; surface smooth except for young ramuli or isidia; ramuli rather close, tending to form vertical rows, bases of the larger ramuli somewhat complanate, the smaller less so; mostly 5-10 (—15) mm. long, often annularly cracked just above the base, slightly inflated in the middle and tapering to acute tips, mostly smooth below, verrucose and isidiose above, the verrucae appearing minutely farinose after the isidia are shed. Cortex 65-90 p thick, probably fastigiate but so gelified as to appear structure- less; algal layer continuous, 30—40 p thick, of Trebouxia, cells 5-6 p in diameter, ` with some colonies invading to the middle of the cortex; medulla pinkish, compact, about 130 y thick, of thick-walled interwoven hyphae; axis about 1100 y in di- ameter, appearing solid, but lacerate with scattered small cavities about 10 p in meter. SIERRA LEONE: Gebense, Sefadu, on trunk of Elaeis guineensis, F. Adames, com. F. C. Deigbton M4756, at Kew. 87. USNEA AMPLISSIMA Stirton, Scottish Nat. 6:106. 1881. Usnea artbroclada subsp. arthrocladodes Vainio, Cat. Welwitsch's African Pl. 2:397. 1 Type: S. Africa, Cape of Good Hope, Somerset East, P. MacOwan. Type of U. artbroclada subsp. arthrocladodes Vainio is: Angola, Golungo Alto, Serra do Alto Queto, Welwitsch 69 p. p.; both in British Mus. TVor. 44 64 ANNALS OF THE MISSOURI BOTANICAL GARDEN Thallus about 60 cm. long, pendent, very rigid, ashy or pure green, partly paler and very slightly reddish; base very large and firm, infrequently but several times dichotomous, axils 90°, branches arcuate, pendent, of nearly the same diameter throughout, up to 2.5 mm. in diameter, tapering slightly at the ends, neither exactly terete nor clearly angled; surface uneven, irregularly subareolate, closely annularly cracked, the cortex regenerating at the cracks, paler or pinkish, wherefore the surface appears irregularly rugose and scabrous on account of the areolation and the coarse concolorous tubercles; ramuli close, slightly curved, very thick at the base, tapering gradually to the tips, the thicker ones tuberculate and sometimes branched, the more slender ones smooth, tips obtuse, closely annularly cracked and subulate. | Apothecia abundant, usually at the tip of a lateral branch, 10—20 (—30) mm. in diameter, nearly plane; exciple smooth (subfoveolate on very large apothecia); ; marginal cilia not dense, radiating, about as long as the diameter of the disc, usually ` not branched, annularly cracked, tips acuminate, usually smooth but some of the longer ones slightly tuberculate; disc plane or slightly rugose in the larger apothecia, dirty flesh color, slightly pruinose; thecium about 70 p tall, epithecium greenish fuliginous, ascospores 11 X 7 p in slightly inflated asci. Cortex about 100 y thick, chondroid, but soft, fragile, green with reddish spots; medulla about 180 y thick, white, dense, K—; axis about 500 p thick, white. : he above description is translated from Motyka as our specimens are rather young, only about 15 cm. long and sterile. The cortex is areolate but not slightly | rugose. Since no suggestions of soredia on the tubercles have been seen, it is possible i that our material may be young U. distensa; however, in our younger plants of that ` species, about the size of the present plants, soredia have already developed. The specimen from Kenya is only a fragment from the middle of the thallus but it is | abundantly fertile. : ENYA: western slope of Mt. Kenya along trail from West Kenya Forest Station Do Geer 3630 m., in giant heath zone, Edgar A. Mearns 1533 (T. Roosevelt Exp.) ex DG K. pN foi. at Farlow Herb. A: Cuanza Sul, pee, Capir near the Carloaongo-Cuvo River, 1000 m., |: : TOES 9991, at Kew. 2 88. USNEA DISTENSA Stirton, Scottish Nat. 6:104. 1882. Type: S. Africa, Cape of Good Hope, Griqualand West, Kimberley Diamond Fields, T. Shaw, in Art Galleries of the Glasgow Corporation. E Thallus pendent, rigid, up to 70 cm. long (only 15 cm. in the type); greeny drying deep olive buff; holdfast up to 8 mm. in diameter, blackened, base 2 mm. m diameter, about 10 mm. long, closely annulate, blackened for a variable distance in different plants, then closely dichotomous above the base (axils 90°) and les. frequently and more irregularly so above; main axes about 1.5 mm. in diametef a thinning at each dichotomy to the capillaceous tips (about 6-10 cm. long), closely 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 65 annulate and subareolate on the main axes, smooth with a few minute pale tubercles, becoming more and larger tuberculate on the more slender branches, sometimes somewhat foveolate and angled; sparingly and ee ramulose on some main axes, quite closely so on others; ramuli about 25 mm. long, closely tuberculate, forming soralia of isidiose soredia, the more gett branches more regularly and closely ramulose, the ramuli shorter, 5-15 mm. long, more slender and less tuberculate, all ramuli rather fragile. Apothecia not seen. Our Central African material is rather short, mostly about 30 cm. long, and may represent young plants although abundantly sorediiferous. The main axes are not subscrobiculate from regenerating cortex of the fissures; otherwise it agrees well with Motyka's description. NDA: Mt. Elgon, 1290 m., on trees, W. Small 218 (fragments from terminal por- tions e Ca thallus TANGANYIKA: Up bara, Amani, m., in rain forest, P. J. Gree enway 998, in African 1 Herb. and at Kew; Amani- Marii. 10950 m. P. T: Green enway 4151, in E. "Bion SOUTH AFRICA: Cape of Good Hope, Griqualand West, Kimberley Diamond EN T. Sbaw, in Art Galleries of the ea rae Cape of Good Hope, Knysna, on Quercu. O. Almborn, Lich. Afric. 18, in Dodge H CONGO: km. 30, route Kam, FL. Hendrickx 4330, in E. African Herb. 18, in Dodge He rb. without locality or collector, ex herb. Macreight in Tuckerman herb. sub erb. U. Nin. at Farlow H 89. UsNEA Promontori Motyka, Lich. Gen. Usnea Stud. Monog. 386. 1937. Type: S. Africa, Cape of Good Hope, in primeval forest, Rebman. Thallus pendent, quite rigid, more than 100 cm. long, pale ashy green (Mo- tyka) drying chamois to cream buff in our material; base not seen; branching dichotomous, axils usually 90°, complanate, the branches curving downward until the axes are nearly parallel, internodes very long, subflexuous to, straight, the ultimate 20 cm. unbranched, about 1 mm. in diameter below, tapering to the long slender tips; cortex of main axis coarsely areolate, margins of areoles slightly elevated, giving a rough, subrugose appearance, the more slender portions with scattered low, pale tubercles, often in somewhat oblique rows, perhaps giving rise to small soralia but not present in our South African material (a few in our Kenya Specimens); ramuli irregularly disposed, sometimes quite close, in other places subdistant, fragile, 5—15 (-20) mm. long, curved, axils not complanate, slender, tapering to acuminate tips. Except for greater length and color, our plants agree closely with Motyka's description. CONGO: Mt. Kahusi, 2700 m., F. L. Hendrickx 4316, basal fragment only, in E. es SS rb. Mau forest, Endabarra, 2030 m., in Acacia shrub forest, P. R. O. Bally B4947, in E. / African Herb. SN AFRICA: Cape of Good Hope, Knysna District, Hermansdorp Division, on [Vor. 44 66 ANNALS OF THE MISSOURI BOTANICAL GARDEN branches of Podocarpus Thunbergii, A. W. Woodcock, ex hb. Mus. Austro-Africanum, at ew. MAURITIUS: without locality, Sieber 43, in Tuckerman herb. sub U. plicata, at Farlow Herb. 90. USNEA GONIODES Stirton, Scottish Nat. 6:108. 1881. Usnea angulata £. goniodes Hue, Nouv. Arch. Mus. Paris, IV. 1:45. 1899, quoad nomen sed non descriptio. Type: South Africa, Cape of Good Hope, H. Bolus 517, in Art Galleries of the Glasgow Corporation. Hue based his new combination on a plant from Paraguay, Peribebuy, Balansa 4140, probably U. sulcata v. neutra Motyka, from his description. Thallus pendent, subrigid, about 30 cm. long, green, drying deep olive buff (finally sayal brown to snuff brown in the type) ; holdfast a dark disc about 2 mm. in diameter, on the upper side of a twig from which two or more axes arise, 0.5— 0.7 mm. in diameter, about 2 cm. long, cylindric, annulate with very few ramuli, curving over either side of the twig, then dichotomous and becoming triangular in cross-section; cortex irregularly annular and subareolate, ridges usually corticate, sometimes cracked, up to 1.2 mm. in diameter (only 0.8 mm. in our rather young material), very rarely dichotomous above; axes flexuous and subparallel through- out, thinning gradually to the very slender, almost cylindric filiform summits; ramuli from the ridges, bases complanate, quite close, mostly about 10 mm. long, very slender, an occasional one growing out as a very slender lateral branch up to 50 mm. long, with small ramuli, the larger ramuli tuberculate and sometimes isidiose sorediose. hecia up to 15 mm. in diameter, at the tips of lateral branches, plane; exciple slightly to deeply rugose-reticulate and sublacunose, epapillate; marginal cilia few and often missing, up to 10 mm. long, tuberculate; disc plane, distinctly rugose, flesh color, chalky, pruinose (olive buff in type and in Brass 16992). Cortex 55—65 p thick, fastigiate, of very thick-walled, conglutinate hyphae, the ` outer 20 y slightly brownish; algal layer 25 p thick, of closely packed cells of Trebouxia 6—8 p in diameter; medulla about 130 y thick, of densely woven thick- walled hyphae, heavily nubilated with dark brown minute crystals, more loosely woven about 40 p outside the axis and tearing on sectioning; axis 730 p thick, appearing solid but minutely lacerate with many small cavities about 10 p in diameter in the inner three fourths of the axis, of thick-walled hyphae 6-8 p in diameter, lumen 3 y. Some of our material is young and sterile. In several places, some injury to the cortex has been followed by a dense branched tuft of ramuli. Gossweiler 13654 consists of a few fragments as if broken from the middle of the primary axes, or perhaps younger plants; Brass 16992 is a fragment from the terminal portion of the thallus, fertile, quite typical except that the soredia are less developed than in the type. 1957] DODGE——LICHENS OF TROPICAL AFRICA. II. USNEA 67 ANGOLA: pm de Lunda, Dundo near Rio Luachimo, 700 m., J. Gossweiler 13654. NYASALAND: Mt. Nchisi, corticole, L. J. Brass 16992 (Vernay Nyasaland Exp., 1946), in N. Y. Bot. Ga Si He rb. FOTOKAS EAST AFRICA: without locality, J. Stocks 39; Mozambique, Inhambane, on Copaifera, without erii a fragment separated from material of U. trichodoides Vainio; both at Kew. SOUTH AFRICA: Cape of Good Hope, H. Bolus 517, type in Art Galleries of the Glas- w Corporation; Adelaide, Miss Brugton, com. Rev. W. Lillie, in Howe herb., det. U. disdehs by M. Bouly de Lindo cited by Motyka (Monog. p. 397), at Farlow Herb. 91. USNEA CHLOREOWEs (Vainio) Motyka, Lich. Gen. Usnea Stud. Monog. 397. 1937. Usnea angulata f. chloreoides Vainio, Hedwigia 37:40. 1898. Type: Uganda, near Mt. Ruwenzori, Nzowi Nez, 0° 5’ S., Scott-Elliott. Thallus pendent, subrigid, about 50 cm. long, between citrine drab and deep grape green; base not seen; frequently dichotomous, branches somewhat flexuous, quite uniformly 1.2 mm. in diameter, frequently and irregularly transversely cracked, angulate, the ridges formed of coalescent rows of coarse papillae, many of Which develop into ramuli, with occasional tubercles between, axis not conspicu- ously sulcate, some ridges quite high and almost alate; ramuli dense, very fragile, annularly cracked, tuberculate, 3-6 mm. long, tapering upward but tip obtuse and slightly brownish. Apothecia absent in the type, abundant in our plants, up to 10 mm. in diameter, lateral on short lateral branches about 0.3 mm. in diameter, stalks coarsely papillate, not angled, quite (ECH SEN deeply incised, margin lobed, smooth, nearly plane; marginal cilia —3 mm. long, d quite fragile; disc plane, vina- ceous cinnamon, dh SE when youn Under two of the apothecia, the branch Se grown out beyond the apothecium for 7 cm., about 0.5 mm. in diameter. With some hesitation I have referred our fertile coastal lowland material to a sterile species from the highlands of Mt. Ru- wenzori. The thallus agrees quite well with Motyka's description, except for more abundant and shorter ramuli and more alate ridges, the upper portion of which often break away, leaving long white scars. 27, at e SE gege? SC nas em? ëm T roe gua ex herb. G. K. Merrill, at Farlow Ag Fort Dauphine, native collector, com. F. G. Paris ex herb. H. E. Hasse, at em Herb. 92. UsNEA FERNANDIAE Duvigneaud, Bull. Soc. Roy. Bot. Belg. 85:110. 1952. Type: Congo: Bas-Congo, Ngombe-Matadi, Paul Duvigneaud 305 U3, in Herb. sigle portion of type in Bot. Mus. Lund., loaned by Dr. Ove Almborn [Vor. 44 68 ANNALS OF THE MISSOUR! BOTANICAL GARDEN Thallus pendent, subrigid, 70 cm. or more long, flexuous, grayish green; pri- mary holdfast not surely seen, but secondary holdfasts formed where lateral branches come in contact with a twig, consisting of discrete, radiating rooting fibers forming a disc 1 mm. in diameter; secondary bases terete, of variable length, cortex smooth, subannularly cracked; irregularly dichotomous at nearly right angles but the branches soon curving downward and subparallel, axils slightly complanate; branches mostly 4- to 5-angled with low obtuse ridges, occasionally acute but not alate, from concrescent rows of tubercles, scars of minute fallen spinules and the complanate bases of ramuli, but in some places almost terete, plane or slightly convex between the ridges, irregularly annularly cracked; slender lateral branches perpendicular to the main axes, terete, 20—40 mm. long, 0.4—0.6 mm. in diameter, sparingly ramulose, minutely tuberculate from scars of fallen spinules, abruptly tapering to acute tips which are sometimes forked; ramuli mostly 5-7 mm. long, about 0.2 mm. in diameter, cortex smooth, rarely annularly cracked above the somewhat complanate base, tips acute but very fragile and mostly broken off; summits more slender and fragile but not very distinct. Apothecia (absent in specimen available) at tips of slender lateral branches, up to 15 mm. in diameter; exciple smooth; marginal cilia few or none; disc almost plane, white-pruinose. Cortex 35 p thick, gelified, fastigiate, outer 6 p filled with dark brown gran- ules, thinning over the ridges and regenerating in cracks until it reaches the bottom of the algal layer; algal layer about 80 p thick, of discrete colonies of Trebouxia 15-20 y in diameter, forming a nearly continuous layer on the outside, less close within, cells 6—7 p in diameter; medulla 80-100 p thick, densely woven just under the algal layer and next the axis, very lax between, hyphae 3 y in diameter, lumen less than 1 A: axis 420 p in diameter, with two cavities near the center, each about 50 p in diameter, of longitudinal, thick-walled hyphae 6-7 p in diameter, lumen nearly 2 p, with scattered thin-walled hyphae 6—7 p in diameter, filled with a deep yellow, highly refractive material (oil?), scattered singly or lying side by side in short rows, not radially arranged. The portion of the axis available for study is younger than that of U. bakong- ensis; perhaps in the older part of the axis, the central cavities coalesce to form a relatively large central cavity as in that species, which it resesmbles in structure of the axis. The medulla is better differentiated and laxer in U. Fernandiae. — Bas-Congo, Ngombe-Matadi, Paul Duvigneaud 305 U3, cotype in Bot. Mus. und. 93. USNEA AFRICANA Motyka, Lich. Gen. Usnea Stud. Monog. 398. 1937. Type: Tanganyika, Usambara Muandara forest, C. Holst. us pendent, subrigid, up to 110 cm. long, dark olive buff, ramuli olive buff (from soredia) ; base absent in our material, frequently dichotomous just above the base, then occasionally so throughout, branches often more or less twisted about 1957] DODGE—LICHENS OF TROPICAL AFRICA. Il. USNEA 69 each other (wind action?), branching at right angles, the two branches curving downward below the axil to become almost parallel, somewhat flexuous, the prox- imal portion up to 1.8 mm. in diameter, tapering evenly to the summits; cortex irregularly transversely and longitudinally cracked, with a few low longitudinal ridges (more in Eyles 1023) formed from the complanate axils of the ramuli; ramuli 4—10 (—20) mm. long, densely verrucose with soralia of SC soredia, quite fragile, tips acute. Similar soralia from the verrucae are abundant on the more slender portions of the main T. W. Maitland 396 is doubtfully ud here. It consists of a coarser, more rigid plant, axis 2 mm. or more in diameter, more conspicuously angled and less frequently dichotomous; the ramuli are 20-30 mm. long and thicker. The Kenya plant is also doubtful. It consists of fragments about 10 cm. long, the main axes tawny olive to clay color, the ramuli chamois to cream buff. It seems closer to U. africana than to the STRAMINEAE, SULCATAE. KENYA: uplands, 2250 m., e S S ex mede Mus. 5161, at Kew. UGANDA: Busoga, Jinja, 15 E. Brown or, ex Bot Gard. Entebbe; Mile 8 on Ma- saba Road, 1130-1240 1 m., F. fr ases 1891, moribund; Bangala, Sese, 1190 m., T. W. Maitland 396; all at K TANGANYIKA: Me pwa, Kiboriani Mt., 1770—1930 2 B. D. Burtt 3880; Kondoa a tek on Oxyanthus speciosus DC., B. D. Burtt 1183, at - RHODESIA: Matopos, 1610 m., "Frederick Eyles 10090) Me District (Gazaland), ni 2250 m,, C. F. M. Swynnerton 617; both at 94. USNEA BAKONGENSIs Duvigneaud, Bull. Soc. Roy. Bot. Belg. 85:109. 1952. Type: Congo: Bas-Congo, between Thysville and Tumba, Pau] Duvigneaud 582 U2, in herb. Duvigneaud, portion in Bot. Mus. Lund, loaned by Dr. Ove Almborn Thallus pendent, 70 cm. and more long, subrigid, grayish green; primary es fast broken off in specimen studied, secondary holdfasts fibrous discs 3—4 mm. diameter, dark brown; secondary base about 1 mm. in diameter, of variable Si cortex annulate and subareolate; main branches 1-1.5 mm. in diameter below, tapering gradually to more closely ramulose, capillaceous summits about 7 cm. long; rather sparingly dichotomous at about 45^ angles, the branches slowly curv- ing downward; below nearly terete except for complanate bases of small lateral branches, rather closely annularly cracked, the cracks regenerating as transverse ridges with minute low tubercles producing 1—3 spinules, becoming irregularly angular above from short longitudinal ridges formed by the complanate bases of small lateral branches, or ramuli, and the upturned edges of longitudinal cortical Cracks with rows of scars of fallen spinules, closely but incompletely annularly cracked forming transverse ridges as in the lower portion, with scattered scars of fallen spinules between; ridges less conspicuous toward the summits which are terete with rare scars; ramuli mostly broken off below, subdistant, becoming closer above, especially on the capillaceous summits, 5-10 mm. long, the shorter smooth, the longer covered with minute, fragile spinules. [VoL. 44 70 ANNALS OF THE MISSOURI BOTANICAL GARDEN Cortex 65—80 p thick, gelified, fastigiate, hyphal lumina about 1 y in diameter, the outer 10 p heavily nubilated with brown granules; algal layer about 225 p thick, of discrete colonies of Trebouxia, about 25 p in diameter, forming a nearly continuous layer on the outside with similar colonies less closely packed almost to the axis, the inner 25 » with only an occasional colony; medulla not differen- tiated as a distinct layer unless the inner 25 p of the algal layer be so considered; axis 800 p in diameter, wall 170—200 p thick, of longitudinal thick-walled, con- glutinate hyphae 3.5 p in diameter, lumen 1 p, with thin-walled hyphae 5-6 p in diameter arranged in nearly continuous, mostly radial rows (suggesting the appear- ance of rays in transverse sections of coniferous wood), cell contents highly refractive as if filled with an oil or resin, colored deep brownish yellow; central cavity stellate in outline of transverse section, 400—460 p in diameter, probably formed by the coalescence of 7—8 central cavities, loosely filled with contorted, thick-walled hyphae 6—7 p in diameter, lumen about 2 p, nubilated with minute hyaline granules. In the original description, Duvigneaud states "medulla et axis K lutescentes et dein rubescentes." In the specimen studied, on shaving off the cortex of an upper portion of a main branch and treating with K, a yellow color promptly develops which goes into solution in K, leaving the exposed medulla white when the solution is blotted off. Mounting transverse sections near the base of a main branch in K, all color disappears from the sections except for algal cells. coNGo: Bas-Congo, between Thysville and Tumba, Paul Duvigneaud 582 U2, in herb. Duvigneaud, portion of cotype in Bot. Mus. Lund. 95. UsNEa Livipa Motyka, Lich. Gen. Usnea Stud. Monog. 419. 1937. Type: Cameroons, Mt. Cameroon, Winkler. Thallus 100 cm. long, pendent, flaccid, uniformly ashy green, partly slightly fuscescent (dark olive buff to deep olive buff in our material); base not seen, in- frequently dichotomous throughout, branches irregularly flexuous, somewhat con- torted, about 0.8 mm. in diameter below, slightly compressed and up to 2 mm wide where the complanate axils of ramuli are nearly opposite, the more slender portions terete, slightly longitudinally rugose from areolation of the cortex which cracks away in places leaving farinose decorticate areas; ramuli somewhat irregu- larly disposed with complanate axils on the larger axes, less so on the slender ones, often curved, tips acuminate, sometimes blackened, variable in length, fragile. Apothecia sessile on short lateral branches or near the base of a large ramulus, 3-4 mm. in diameter, nearly plane; exciple slightly impressed; marginal cilia about 12, very variable in length; disc slightly concave, carneous, densely white-pruinose. Cortex 40—45 y thick, highly gelified, the outer 15 » brownish; algal layer 20- 25 p thick, of discrete colonies of Trebouxia, cells 6—8 p in diameter; medulla 40-55 p thick, of densely woven slender hyphae, heavily nubilated; axis 670 X 490 p, outer surface eroded longitudinally, appearing solid, but slightly lacerate 1957] DODGE——-LICHENS OF TROPICAL AFRICA. IIl. USNEA 71 with a few narrow bundles of slightly brownish, coarser hyphae, resembling the rays of vascular plants. Most of our material is rather young and is densely tangled, somewhat smaller in dimensions than the type, but agrees well in essential characters. SIERRA LEONE: Sugar Loaf, 480 m., H. Donovan Tindall 16, at Kew. OAST: C. ; 5 T. F. Chipp 445, at Kew. CÓTE Rene Gommilé, Agnibile-Krou, Guy Roberty 13553, in Herb. Mo. Bot. Gard. CAMERO Victoria, in Meyer Crater, 2580 m., on NM salicifolia, T. P. M. 917 b. b. min. yec tangled with Usnea speciosa Mot., at 26. UsNEA AMANIENSIS Dodge, Ann. Mo. Bot. Gard. 43:396. 1956. Type: Tanganyika, East Usambara, Amani-Maramba, 1030 m., growing in festoons on the topmost branches of tall evergreen rain-forest trees, P. J. Greenway 4151, ex Herb. E. African Agr. Res. Sta., Amani. Thallus prostrate and creeping along the larger branches of trees, the more slender branches pendent, rigid, more than 60 cm. long, chamois to deep colonial buff; base black, conic, 5 mm. in diameter, buried in humus (decaying mosses, leaves, etc.) and attached to the bark, tapering upward to 1.4 mm. in diameter and about 10 mm. long, then several times dichotomous with short internodes, then rarely dichotomous throughout; principal axes about 1 mm. in diameter, thinning at each dichotomy to the capillaceous summits; short lateral branches or ramuli penetrate the humus and form secondary holdfasts on the underlying bark, giving rise to secondary axes; terete, smooth, without verrucae or tubercles; main axes closely annularly cracked, the cortex regenerating over the crack, producing a white annular appearance of the larger axes, longitudinal cracks rare and the are- oles do not crack off; the more slender branches with very narrow annular cracks, sometimes close, sometimes distant on the same branch, terete or with single minute paler verrucae; ramuli 10-20 mm. long, very slender, straight or curved, rather fragile, tips acute, sparsely and minutely tuberculate, no soredia, isidia nor propagula seen, Apothecia abundant on some axes, absent on others, 3-5 mm. in diameter, cupulate, becoming subplane, sessile on ramuli, eciliate; exciple smooth; disc fuscous, white-pruinose. Cortex about 55 y thick, apparently fastigiate but so highly gelified it is im- Possible to see the structure clearly; algal layer 45 p thick, of discrete colonies of Trebouxia about 30 p. in diameter, forming a nearly continuous layer; medulla 105— 120 y thick, of very densely woven hyphae, so heavily nubilated with minute brown crystals that structure is not very clear; axis 450 p in diameter, of strands of longi- tudinal, conglutinate hyphae, lumen 1—1.5 p in diameter, separated by thin-walled hyphae 6-8 p in diameter very heavily nubilated with brown crystals (giving a brown reticulate appearance to the transverse section). The Nyasaland fragment has a single apothecium 9 mm. in diameter, a slightly thicker cortex; the axis appears solid, i.e. the thinner walled cells have not yet VOL. 44 IR ANNALS OF THE MISSOURI BOTANICAL GARDEN turned brown and are not nubilated. The color of U. amaniensis is suggestive of the GLABRATAE, OSSEOLEUCAE, but all the other characters clearly relate it to the ELONGATAE, CERATINAE Or LONGISSIMAE. TANGANYIKA: East Usambara, Amani-Maramba, 1030 m., growing in festoons on x topmost branches of tall Freien rain-forest trees, P. J. Gr reenway 4151, ex Herb. E Reesen: MS ing Sta. oe Mlan ron Es Luchenya Plateau, Mt. Mlanje, 1890 m., L. J. Brass DEL 16828, DNE (Vernay Nyasaland Exp., 1946), in N. Y. Bot. Gard. Herb. 97. USNEA EBURNEA Motyka, Frag. Flor. Geobot. 1:29. 1954. : Madagascar, Manakambahiny Est, large wood between Alaatra and Tamatave, near Saharakara River, 800 m., com. H. des Abbayes Thallus pendent, quite rigid, about 80 cm. long, yellow-stramineous (cream buff in our plant); holdfast a black disc 5 mm. in diameter, surface with radial ridges (coalesced rooting fibers), base 2-3 mm. long, closely annulate, dichotomous at very obtuse angles, soon dichotomous at a 90? angle and occasionally so above with long internodes, axes quite flexuous, very slightly angled, almost terete, rarely with cristate ridges on larger branches, closely annular, terete on more slender branches and less closely annulate; surface with scattered subpapilliform tubercles, sometimes in irregular rows on the larger branches, apparently elongating and breaking off as a rather stout isidium, paler than the thallus; ramuli close except near the base, mostly 10 (-20) mm. long, curved, tapering to an acuminate point, minutely and sparsely tuberculate. Apothecia not abundant (absent in our material), subterminal on larger ramuli, 2—4 mm. in diameter, concave becoming plane or even revolute, deformed; exciple smooth, eciliate; disc deep flesh color, white-pruinose. MAURITIUS: without locality or collector, at Kew. 98. USNEA SAVANARUM Duvigneaud, Bull. Soc. Roy. Bot. Belg. 85:112. 1952. Type: Congo: Bas-Congo, between Thysville and Tumba, Paul Duvigneaud, 582 U4, in Herb. Duvigneaud, two fragments in Bot. Mus. Lund, loaned by Dr. Ove Almborn. Thallus pendent, 50 cm. or more long, flaccid, grayish green; holdfast and base not seen; primary axes somewhat flexuous, 0.5 mm. in diameter, subterete, occa- sionally subangular from slightly elevated isidial scars in rows; surface subareolate- corticate (not scaling in the material studied, perhaps so in the lower portion of the thallus) appearing decorticate from coalescent scars of fallen isidia and minute spinules up to 0.3 mm. long, partly tartareous from protruding medullary hyphae; ramuli 3-5 mm. long, perpendicular, abruptly tapering to acute tips, close, smooth or nearly so, about twice annularly cracked; slender short lateral branches up to 25 mm. long, unequally dichotomous, minutely spinulose, appearing minutely 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 73 farinose-tuberculate after the spinules are shed. Apothecia (absent in specimen studied) 3-4 mm. in diameter; exciple sub- smooth to very slightly pitted; marginal cilia few, resembling the ramuli; disc white. Cortex about 40 y thick, fastigiate, hyphae 6—7 y in diameter, lumen 1 p, geli- fied, the outer 10 p heavily nubilated with dark brown granules; algal layer of discrete colonies of T'rebouxia, about 15 p. in diameter, forming a nearly continuous layer, with some colonies much deeper in the medulla; medulla 30-35 p. thick, of rather densely woven, predominantly radial hyphae, 6-7 p in diameter, laxer next ` the axis, heavily nubilated with dark brown crystals, soluble in KOH but not in lacto-phenol; axis solid, brownish, 250 p in diameter, of conglutinate, longitudinal hyphae, about 4 p in diameter, lumen about 1 p; axis K reddish brown; medulla K yellow becoming orange as the solution dries. Transverse sections are promptly bleached by K except for the algal cells. Occasionally a member of a dichotomy of a short lateral branch is very densely branched, forming a compact cluster 3—4 mm. in diameter, probably teratological. Only two fragments about 10 and 12 cm. long, were available for study. These are clearly from the upper half of a main branch and are clearly corticate through- out, but it is probable that the basal portion of the thallus would show some scaling, due to the thin, laxly woven region of the medulla next the axis. CONGO: Bas-Congo, between Thysville and Tumba, Paul Duvigneaud 582 U4, cotype fragments in Bot. Mus. Lund. 99. USNEA INDIGENA Motyka, Frag. Flor., Geobot. 1:27. 1954. ype: Madagascar, Manakambahiny Est, large wood between Alaatra and Tamatave, Sahamalaza, 800 m., com. H. des Abbayes. allus pendent, flaccid, about 30 cm. long, chamois to cream buff; dichoto- mously branched, the branches curving downward until they are soon nearly parallel as main axes, 0.7—0.8 mm. in diameter, subsequent dichotomies rare, axils complanate, 90?; very indistinctly few-angled, lower portion of axes annulate, smaller branches continuous, terete, smooth, tapering to capillaceous summits; ramuli moderately close throughout, slender, tapering to acuminate tips, mostly 4-10 mm. long, a few longer and dichotomous at their tips, bearing a few minute spinules which break away and serve as isidia. The ramuli are capable of forming a holdfast where they come into contact with a twig, as a black disc about 1 mm. in diameter. ew. MADAGASCAR: Cercle M. des Merie [? illegible], native collector, vere 1901, ex herb. H. E. Haase, at Farlow Herb. TVor. 44 74 ANNALS OF THE MISSOURI BOTANICAL GARDEN 100. USNEA CONTORTA Jatta, Malpighia 19:163. 1905. Type: Madagascar, Piunarantsoa, Beforna, collected by a native. Thallus pendent, subflaccid, up to 60 cm. long, pale stramineo-green (Motyka), dark olive buff to deep olive buff in our plants; holdfast a thick black disc 5 mm. in diameter, base blackened, 5 mm. long, 1 mm. in diameter, closely annulate, branching closely dichotomous just above the base, axils sometimes complanate, the branches curving and becoming subparallel, very rarely branched above, often one branch larger than the other; axes 1 mm. in diameter, areolate below, sometimes with simple cracks, in other plants the margins of the areoles obtusely elevated, subscrobiculate with predominantly oblique ridges, giving a somewhat contorted appearance, rarely low-tuberculate between as well as on the ridges, more slender branches closely but indistinctly annulate with scattered single tubercles; ramuli irregularly disposed, rare on the lower portions of the main axes, closer above but not crowded, 10-20 (—35) mm. long, faintly tuberculate, some straight, others curved or flexuous, tapering uniformly to acuminate, slightly darkened tips. Apothecia not abundant, sessile on the more slender portions of a main axis, or more frequently lateral in the middle of a very long ramulus, cupulate, becoming plane; marginal cilia irregularly disposed, quite close, erect, of variable length, similar to ramuli; exciple smooth, margin splitting and becoming crenate; disc pale flesh color, white-pruinose (in some, the thecium has been eaten away by insects and small erect cilia develop on the old disc). TANGANYIKA: Southern Highlands, Iringa District, Morgana, 1930 m., W. Carmichael as in iem African Herb. Pages gia District in forest, Ecklon, ex herb. W. Sonder, in Tuckerman herb. at a Farlow TIUS: Mecum prae Robillard ex Gray Herb. at Farlow Herb. Abe UNION: Salazie, Frére riguez, com. Boule de Lesdain ex herb: Hasse sub U. mascarena Motyka olim herb. nom., at Farlow Herb. 101. UsNEA arcuta Motyka, Lich. Gen. Usnea Stud. Monog. 406. 1937. Type: Congo, Congo de Lemba, Goliat. Thallus up to 40 cm. long, rather more rigid than U. trichodeoides, pendent, axes pale olive buff when decorticate with isidiose soredia, corticate portions and ramuli olive buff; base not seen; infrequently dichotomous with obtuse axils, the branches curving downward and soon parallel; axes about 1 mm. in diameter below; somewhat flexuous, terete to slightly longitudinally rugose, cortex cracking off and then farinose with abundant isidiose soredia, patches of cortex remaining with minute sorediose verrucae; ramuli quite close, 10-20 (—30) mm. long, the longer subdichotomous, tips acuminate, straight or curved, corticate with minute, irregu- larly disposed pale tubercles, forming soralia on the longer ramuli, but never as . dense as on the axes Apothecia (on Chipp 57 only) sessile on the smaller branches, up to 6 mm. in diameter, nearly plane; marginal cilia few, radiating, 5-10 mm. long, similar to 1957] DODGE—LICHENS OF TROPICAL AFRICA. II. USNEA 75 ramuli; disc flesh color, white-pruinose, becoming nude when old. ur plants are mostly short fragments about 20 cm. long, but agree with Motyka's description. Perhaps the variable amount of decortication may be cor- related with habitat, depending on dryness and wind. Plants referred here from semi-swamp, dense forests are much more corticate than from open scrub woodland and savannas. Eyles 1023, a short fragment from the mid-portion of a main axis, is somewhat doubtfully referred here. The axis is mostly corticate and more distinctly rugose than the other plants, but much less so than in U. aequatoriana Mot. Curious spinules with dark tips are irregularly borne on some ramuli. SUDAN: Issore, Onyiro, 3° 50’ N., 32° 50’ E., 1720 m., T. F. Chipp 57, more corticate and e at Kew ENYA: Chyulu hills, 1290 m., on trees in mixed savanna forest on lava flows, P. R. O. Bally 1236, at Kew UGANDA: Rude River, 1390 m., in riverine forest, T. Jarrett 402 p. p., at Kew ANGANYIKA: Bukoba, KEE Mingiro Forest Reserve, 1550 m., on trees tof dense semi- genge siue forest, G. Watkins 492 (native name: inacbankimá 4), at Kew ODESIA: E. Isoka venere PM 1480 m., on trees in Bracbystegia mimosaefolia nb Vdcdladd: C: G. Trapnell, a - RHODESIA: Matapos District, T m., Frederick Eyles 1023 p. p. min., at Kew. 102. USNEA TRICHODEOIDES Vainio, Ann. Acad. Sci. Fenn. A 67:8. 1915; emend. Motyka, Lich. Gen. Usnea Stud. Monog. 421. 1937. Usnea longissima f. Ebersteinii Stein, Jahresber. Schles. Ges. Vaterl. Cult. 66:133. 1888. Type: Cape of Good Hope, com. Persoon, in herb. Acharius at Helsinki. Mo- tyka based his description on Tanganyika, Usambara, C. Holst 2642. Type of Usnea longissima f. Ebersteinii Stein, from Tanganyika, Kilimanjaro, Rabai Ndaru in scattered forests in steppes, H. Meyer. Thallus up to 40 cm. long, pendent, very flaccid, ecorticate, axis colonial buff to white, corticate, ramuli deep olive buff; dichotomous just above the holdfast, then rarely dichotomous throughout, the branches soon parallel, straight or slightly flexuous; axis 0.4-0.6 mm. in diameter, slightly thinner at each dichotomy to the capillaceous summits, terete to slightly flattened and occasionally very faintly longitudinally rugose; cortex soon flaking off, leaving a farinose, subsorediose surface on most of the branches except the smallest, sometimes with a few minute spinules (isidia?); ramuli close, somewhat variable in length, 3-10 (-15) mm. long, corticate, terete, smooth, tips acuminate, the longer dichotomous near the tips. Apothecia (absent from most of our material) subterminal on stouter ramuli, 3-8 mm. in diameter; exciple slightly lacunose to smooth; marginal cilia few, about 10 mm. long, often broken off; disc nearly plane, flesh color, lightly but distinctly pruinose. The material BET here is rather variable. It differs from U. arguta Motyka in having axes more slender (0.4—0.6 mm. instead of 1 mm.), straight rather than tortuous, shorter, smooth ramuli 3-15 mm. long (instead of 10-30 mm.) with [Vor. 44 76 ANNALS OF THE MISSOURI BOTANICAL GARDEN irregularly disposed small sorediose tubercles; axis with pale yellow strands instead subfuscous axes from strands of fuscous hyphae. SUDAN: Onyiro, Issore, 1770 m., 3° 50’ N., 32° 50’ E., T. F. Chipp 57, a fragment, t Kew. KENYA: Mt. Elgon, 2790 m., E. J. Lugard & Mrs. Cyril Lugard 470, at Kew; Loita plains 40 e 60 pros southeast af Narok, 1610-2250 m., Anita G. Curtis 669b, 743c, in Dodge herb. det. Motyka, and at Farlow Herb.; Lake Baseng, 1775 m., on trees, G. M Allen 1 os in Howe herb. sub U. ^y at Farlow H erb. UGANDA: Ruizi River, 1550 m., T. Jarrett 402; Bugishu, EMEN 2410 m., A. S. ae? 484; K igezi, Kasatoro, Kabale River, 2250 m SA R. Dale L43; Mt. Elgon, 1290 m., W. Small 216; without locality H. Fyffe 167; all at ONGO: Mt. Kahusi, ca. 2700 m., on Cer s po. F. L. Hendrickx 4302, in E. TANGANYIKÀ: Bukoba, Kabobwa, Mingito Forest Reserve, 1550 m., G. Watkins 492 (Forest SCH Herb. 3231, native name: machan ma), sheets at Kew and in E. African Herb.; Maruessa between oe? and Taita, Hildebrandt com. C. Rensch ex herb. Sbar- baro, at Y: mee? erb. ZANZIBAR ISLAND: without locality, J. H. Vaughan 1359, at Kew. IAM UE: epee on Copaifera, without collector, at Kew RICA: Gno e Di Drége, com. Laurer ex herb. Sbarbaro: Uitenhage District, in Mehdi Ecklon, ex herb. W W. Sonder in Tuckerman herb.; both at Farlow Herb. STUDIES IN THE CAPPARIDACEAE. III EVOLUTION AND PHYLOGENY OF THE WESTERN NORTH AMERICAN CLEOMOIDEAE* HUGH H. ILTIS** One of the best ways to visualize evolution is to arrange related forms in the order of their complexity. With fairly large and complete series, it is possible to show graphically the homology of unlike extremes by observing the intermediates. Use of the so-called reduction series (Woodson, 1935) such as the modifications described by Bower (1923) and by Zohary (1948) is open to severe criticism, since they are composed of contemporary types not in direct line of descent (Lam, 1951). Nevertheless, they help to elucidate such problems as relationships between species and species groups, direction of evolution, and morphology of derived structures, particularly in cases where paleobotanical and cytogenetical evidence is lacking. The Capparidaceae, as well as the whole of Rhoeadales, provides a wealth of series illustrating many evolutionary problems (Camp, 1950). Four western North American genera of the subfamily Cleomoideae, namely Cleome sect. PERITOMA, Cleomella, Wislizenia, and Oxystylis, constitute one of the most clear-cut "reduc- tion series," and their taxonomy, morphology, geographical distribution, intra- and inter-generic relationships, and intra-specific variations are discussed in this paper. Based on the above, an attempted evolutionary interpretation is correlated with the geologic history of North America and is discussed in the light of the probable great importance of desert evolution in this group and the angiosperms generally.! I. TAXONOMY The Capparidaceae is a member of the Order Rhoeadales, with the Cruciferae its closest, most specialized relative. The 40 genera and 700 species of Cappari- daceae, occurring mainly in the tropics of both hemispheres, are customarily Ges arated into two major subfamilies: Capparidoideae and Cleomoideae. The Cap- paridoideae includes forms, mainly shrubby or arborescent, with fleshy, indehiscent fruits. Its largest genus is Capparis, the pickled flower buds of the Old Worl C. spinosa furnishing the capers of commerce, a seasoning for food. The Cleo- TT should like to express my thanks to > keng of Science and dere ur d a = and Sciences, Universit y of Arkansas, and t onte ee of Botany, Univer. both of which gave me aw during the course win t s study; also to the S Society Gë SC X. fee à grant-in-aid. Many (friends have given me much encouragement and helpful criticism C. Johnson, - C. Rollins, L. H. Shinners, and parti deier Drs. Royal Shanks a arp, of the University of Tennes essee; and Drs. George B. Van aack, R. M. Iryon, Ae ae ` Missouri Botanical Garden, St. Louis, where a mino tion of the present paper was used as p a master’s thesis (Iltis, 1950). My wife, Grace Theis, da provided pss ee th reire this study, and her interest and counsel are here gratefully acknowledged. —— * This paper was presented at a symposium on "The Contribution y Systematics to Evolutionary Studies," held inde id auspices of Se Missouri Botanical Garden, St. Louis, on November 4—5, un and supported by the National Science Cal rst ** Department of Bot tany, University of Wisconsin, Madison. (77) [Vor. 44 78 ANNALS OF THE MISSOURI BOTANICAL GARDEN moideae are mostly herbaceous and annual, rarely shrubby. The fruits are usually dehiscent, unilocular capsules (siliques) with two generally deciduous valves at- tached to a persistent replum (placenta). This subfamily includes the small North American genus Polanisia and the world-wide genus Cleome, with about 180 taxa, 83 of which occur in the Western Hemisphere (Iltis, 1952, 1955b). Of these, some are known for their showy flowers, such as C. houtteana Schlecht. (CC spinosa” hort.), the Spider Flower, and C. serrulata Pursh, the Rocky Mountain Bee Plant. SYNOPSIS OF THE TAXA CONSIDERED IN THIS STUDY (Probable relationships diagrammed in fig. 16) A. Cleome? sect. PERITOMA (DC.) Baill. 1. C. lutea Hook. (including var. jonesii 4. C. serrulata Pursh—Rocky Mountain Bee Macbr.)—Yellow Bee Plant, Stinking Mus- Plant tard 5. C. multicaulis Sessé & Moc. ex DC. (C. 2. C. isomeris Greene (Isomeris arborea Nutt.) sonorae Gray —Burro Fat, Bladder Bush 6. C. sparsifolia Wats. 3. C. platycarpa Torr. B. Cleomella DC.? 1. C. angustifolia Torr. 6. C. billmanii A. Nels. (including C. mac- . C. longipes Torr. brideana Payson o * 3 oN 2 3 "nis Iltis . C. obtusifolia Torr. & Frem., sensu lato 4. C. mexicana Sessé & Moc. ex DC. . C. plocasperma Torr. & Frem. (sensu 5. C. palmerana Jones (a complex including C. Abrams, 1944 Rydb., C. montroseae Payson, C. 9. C. parviflora Gray (including C. alata Eastw., C. gracilis Brandg.) 10. C. brevipes Wats. C. Wislizenia Engelm.* 1. W. refracta Engeim, (including all nine species described by Greene (1906) and others.)— Jackass Clover, Stinkweed, Spectacle Pod D. Oxystylis Torr. & Frem. - O. lutea Torr. & Frem. IL. GENERAL MORPHOLOGY The eighteen species enumerated above share such a large number of vegetative and floral characters that it is sometimes difficult to distinguish between flower- ing plants of two genera. All species are small or medium-sized plants (figs. 4-6), annual (except Cleome isomeris, Cleomella perennis, C. mexicana, and, rarely, forms of Wislizenia), though not always ephemeral, and are essentially glabrous (except Cleome platycarpa and Cleomella obtusifolia). ‘The leaves are compound, with 3 (in Cleome lutea 5) entire leaflets less than 6 cm. long, succulent and thick, more or less conduplicate and often glaucous, these usually tipped by a thick, usually varnished mucro or by a slender scarious hair. : : 2 A comprehensive monograph of Cleome in the New World, written as a doctor's thesis, is await- ing publication (Iltis, 1952). In the only published revision of Cleomella (Payson, 1922), the species concept is too narrow. 's (1929) unpublished thesis was useful in my studies, which cover all species except C. an- gustifolia and the C. palmerana complex. 4 The many variants here are best divided into the eastern W. r. var. refracta, the northern W. r. var. melilotoides (Greene) Johnston, the southwestern W. r. var. palmeri (Gray) Johnston (cf. ` Johnston, 1924), Kuntze (1903) considered Wislizenia a section of Oxystylis. E 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 79 One significant peculiarity of all but two taxa (Cleome isomeris, C. platycarpa) is the possession of microscopic whitish, filiform or laciniate, scarious stipules? on either side of the petiole base or, in ebracteate flowers, at the base of the pedicel (figs. 7-8). These stipules are inconspicuous, and commonly break off in old stems. Their frequency emphasizes close relationship, since in other Cleomoideae such structures are found only in a few species of Cleome and rarely in Polanisia. Morphologically, they seem to be homologous to the minute, gland-like stipules of the Resedaceae, concerning whose nature there have been so many divergent views (Bolle, 1936). In the opinion of Hennig (1929) they are probably reduced stipules. The presence of these anomalous structures is additional evidence of relationship between the two families. The tetramerous, usually weakly zygomorphic flowers, borne in racemes (figs. ^, 5, 8) are characterized by frequent occurrence of persistent or tardily deciduous, synsepalous calyces, by sessile or subsessile petals, and tightly coiling mature, dry anthers. The flowers, in general, are larger in Cleome than in the other genera. The petals are yellow in all species except the purple or pink Cleome serrulata and C. multicaulis, and have a convolute (closed) aestivation. A structure that varies considerably is the more or less eccentric nectariferous disk between corolla and androecium which is highly complex in some species. All species seem to be insect- pollinated (Pellett, 1930; Jepson, 1936). The truncate or pointed stigma (in Cleome multicaulis, somewhat capitate) and the slender, indurate style (fig. 3) point to the general specialization of this group, since I consider sessile, capitate stigmas as primitive. This view agrees with that of Zohary (1948) regarding the Cruciferae. The seeds vary in size and ornamentation, but in structure are basically uniform, being reniform, globular to ovoid, somewhat flattened and characterized by a deep invagination, with the embryo, as in all Cleomoideae, curved and notorhizous (conduplicate). The two parallel ends, or “claws,” may remain open (as in Polanisia and most Old World species of Cleome), or they may be fused exteriorly by a thin membrane or the thick testa, as in nearly all New World Cleomoideae. In the four genera treated here the two “claws” are fused internally as well, sye at the very tip. The lack of a cavity inside the seed is also a specialized condition, for in primitive tropical species, e.g., Cleome parviflora or C. spinosa, the internal sinus between the claws is large. All species inhabit semi-arid or arid regions of western North America (figs. 9, 12), such as deserts, semi-deserts, sagebrush ( Artemisia) associations, alkali flats, and grasslands, occurring often in “badlands and in gravelly or sandy, dried-up river beds and shores. Most species occur in ine and saline soils, with some definitely halophytic. ae In summary, the uniformity of most vegetative and floral structures justifies the assumption of this as a coherent group of closely related taxa. — The “stipular” thorns or spines of Cleome and Capparis are apparently epidermal outgrowths, not foliar in nature. [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN i VALVE VIEW eg x sr diis beach 41 l Jo CLEOME CLEOMELLA WISLIZENIA OXYSTYLIS REPLUM VIEW ig. 1. Fruit reduction series of the North American Cleomoidea A CE € Ze Kee pee Eaäeecier C, Cleome multicaulis; D. T Ee siliculifera dieu E, F, lla palmerana; G, Cleomella ag erma; H, Cleomella obtusi- iik Chocolat billment, J; Cleom ella obtusifolie var. jenrocien nos; K, Wislizenia Lë var. palmeri; L, Wislizenia refracta var. refracta; M, Oxystylis lutea. M CLEOME CLEOMELLA WISLIZENIA OXYSTYLIS 10:010]0/0]0» 8 DADA E OTO oO Fig. 2. Silique Er ot series of the North American Cleomoideae. Letters cor- ones to those of fig. 19571 ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 81 IIl. MorruHorocicat Basis FOR AN ASSUMED GENERIC PHYLOGENY The fruits, racemes, and stipules of the four genera discussed may be arranged in phylogenetic series, in which we can follow each from Cleome, through Cleo- mella and Wislizenia, to its greatest modification in Oxystylis.8 1. Modification of the Fruit.— The changes are most striking in the fruit. Indeed, this “reduction series” suggested the current study.” The basic fruit type of the Cleomoideae is the silique, but of the four genera only Cleome has siliques, a fruit type that is con- sidered primitive in this family as well as in the whole order (Zohary, 1948). All species of Cleome sect. PERITOMA have dehiscent siliques formed by two valves attached to an elongated replum on which the 8 to 30 or more seeds grow. When the fruits are ripe the valves fall off, permitting the seeds to fall free while the replum remains attached to the plant (fig. 10A). Interjected between silique and receptacle is the gynophore (fig. 2), a slender stalk 5-40 mm. long. Both &ynophore and pedicel elongate somewhat after flowering. Some Cruciferae have similar siliques except that a false septum is drawn across the replum making the pod two-chambered. The siliques of two Cleome species may help us to understand the evolution of the Cleomella silicle. (1) Cleome lutea, primitive as the only species with five leaflets, has fruits that vary from only 8 mm. long in depauperate plants of var. lutea to 80 mm. in var. jonesii. In the short fruits and general habit, var. lutea at times strongly resembles Cleomella angustifolia and C. longipes. (2) Cleome blatycarba, primitive in its floral structure but otherwise specialized, has siliques anomalous in their strong compression (fig. 1B), with flat valves attached to a broadly oval replum tipped by a long style. A roundish replum, rare in Cleome, is usual in Cleomella. Hence, in replum shape Cleome platycarpa is intermediate between Cleome and Cleomella. Young plants of Cleome platycarpa closely resemble those of Cleome billmanii, a similarity probably due to convergent or parallel evolution. Cleomella fruits (fig. 4) are silicles, short siliques not much longer than wide. The repla are oval to round (fig. 1 E-J, top row), similar to those of Cleome, but wider and smaller. Viewing the fruits from the side (fig. 1 E-J, bottom row), however, we find a variety of shapes caused by lateral expansion of the valves, ranging from hemispheres to obtuse or acute cones or horns (fig. 15). ome Cleo- mella species may have fruits very similar to the unrelated South American Cleome siliculifera Eich), (fig. 1p) or abnormally small-fruited C. lutea. As in Cleome, the seeds fall free (fig. 108). à The tendency of Cleomella valves to expand laterally correlates with: (1) the ——— ome and Cleomella was derived main] Nevertheless, the phylogeny hin a * Evidence for establishing the generic Soe of Cle dtr i i enus. x ` ` primitive to spect from the morphology of the primitive taxa in eac i , e 1 : ge tylis Species are homologous to and parallel those of the series Cleome —> Oxystyits. : 3 I am grateful to my esteemed teacher, R. E. son, Jr., who interpreted for me this sequence, thus introducing me to the concept of “reduction series” (Woodson, 1935). [Vor. 44 82 ANNALS OF THE MISSOURI BOTANICAL GARDEN obvious decrease in length of gynophore, pedicel, and fruit (Oé to Yo the length of Cleome siliques), particularly in the Great Basin species; and (2) the radical decrease in seed number (2-12 (—20) per silicle, cf. fig. 2). Here, while decrease in fruit size is accompanied by only a minor decrease in seed size, the seed number decreases markedly. By contrast, in other groups of Cleome with reduced siliques (West Indian *UNIFOLIOLATAE"; North African sect. THYLLACOPHORA Franch.), seed size is usually directly PME to silique size, with the seed number essentially constant. MEDIAN OBLIQUE SECTION. SECTION CLEOME CLEOMELLA WISLIZENIA OXYSTYLIS Fig. 3. Longitudinally sectioned fruits (e, gynophore-replum-style) after removal of valves and seeds. Most variable in fruit shape is the Californian Cleomella obtusifolia, with ob- tusely conical valves in the typical form (fig. 1H) to narrowly conical ones in var. taurocranos (fig. 1, and 27). In this most specialized fruit of Cleomella the valves are constricted at intervals. Due to this a single seed is often loosely im- prisoned in the tip of each valve (fig. 10c), the valve and seed falling together at maturity—a rather insignificant and accidental matter were we to stop at this point! However, what we see is a feeble expression of an evolutionary tendency, hereafter called the "schizocarp tendency" which reaches great prominence in Wislizenia and Oxystylis. 1957] ILTIS—-EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 83 The polymorphic, monotypic Wislizenia (fig. 5) differs sharply from Cleomella in fruit characters, the evolutionary trends already apparent in Cleome and Cleo- mella (figs. 1 and 2, K-L) being greatly increased. Thus, the repla, reduced to minute rings 1 mm. in diameter, tipped by long, slender styles (fig. 3), are borne on gynophores and pedicels, each only 5-12 mm. long. Because of the small replum, the shape of Wislizenia valves differs from those of Cleomella. Valves of Cleomella are hemispheric to narrowly conical, being widest at the point of attach- ment to the large replum and narrowest at the apex, while those of Wislizenia vary from obovoid or subspherical to obconical, being narrowest near the point of attachment to the tiny replum and widest near the middle in obovoid valves (fig. IL) or at the apex in obconical ones (fig. 1K). Thus, when the fruits mature, each valve permanently encloses its one or two seeds (fig. 10D). In Wislizenia, then, the fruit is really a schizocarp, with each of the two valve-seed units a mericarp. These could be thought of as nutlets, in which all but the seed tip (which may project partially through the basal pore of the valve) is enclosed by the ovary wall. We have here a continuation of the schizocarp tendency indicated in Cleomella obtusifolia var. taurocranos. O four ovules in each ovary, usually only two mature into seeds, although all may develop in the larger-fruited, primitive W. r. var. palmeri. This low seed number is a continuation of the trend already manifest in Cleomella. The valves of var. palmeri are obconical and elaborately ornamented, while those of var. melilotoides are obovoidal and nearly smooth. However, the latter not only reach ultimate size and valve reduction, with but one seed in each valve, but the valves also resemble those of Oxystylis, the final genus in this sequence, in having small rounded mericarps with tightly held seeds. Furthermore, while both Cleomella and Wislizenia usually have divaricate valves, some races of the special- ized northern (melilotoides) and eastern (refracta) varieties have valves which Point unilaterally away from the plant, resembling those of Oxystylis (fig. 11, top row). Oxystylis lutea (fig. 6), endemic to the Death Valley region, shows to what igh degree a generalized structure such as the Rhoeadalean silique may become modified. While its seeds and valves resemble those of Wislizenia r. melilotoides, the structure of style, replum, gynophore, and pedicel is radically different. The replum of Wislizenia (fig. 3) is a tiny ring. In Oxystylis (figs. 2 and 3) this ring undergoes great lateral and downward expansion and thickening, resulting in an "inverted V-shaped" tube with down-facing open ends (fig. 2M). From this modified “replum” a greatly thickened, elongated, spinescent style projects upward. e gynophore and pedicel, each about 1 mm. long, are also much stouter and shorter than in Wislizenia (fig. 8c, 8D). Attached to the down-facing orifices of the replum are the two parallel and unilaterally protruding, obovoidal valves (figs. 1M, 6, 8D). Like those of Wis- lizenia r. melilotoides, they enclose their single seed tightly and permanently (fig. 10E), with but a minute pore at their attachment point. However, the thinner [Vor. 44 84 ANNALS OF THE MISSOURI BOTANICAL GARDEN and very smooth Oxystylis valves fit the seeds like seed-coats, which, functionally, they are. We have traced the fruit modifications from Cleome through Oxystylis: (1) from large, elongate siliques, through round or laterally pointed silicles with di- vergent horizontal valves, to small schizocarps with nearly vertical, unilateral valves; (2) from fruits with many free-falling seeds to fruits where each valve and its single seed are inseparably joined and fall as one unit; (3) from fruits with short, slender styles and large elongated, hoop-shaped repla to those with long, spinescent, stout styles and small, tubular, "inverted V-shaped" repla; and finally (4) from elongate pedicels and gynophores to fruits in which these structures are short, stout, and nearly obsolete. 2. Modification of Racemes and Bracts.— e may trace the modifications of racemes and bracts from primitiveness in Cleome to specialization in Oxystylis. The primitive Cleomoid inflorescence is a bracteate raceme. Through loss of bracts, ebracteate types have evolved independ- Ne ` $52 Zl S ^i JS D SS L KE Ry e [S8 x Jh n bracteate inflorescence of 5. Dense ebracteate inflorescences Fig. 4. Ope Cleomella ‘angustifolia. Note long region of of Vieni. vegetative growth before flower production 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 85 ently at different times, with both bracteate and ebracteate racemes occurring in several genera (various groups of Cleome and in Podandrogyne). ‘The ebracteate type generally characterizes specialized taxa. Racemes are bracteate in Cleome sect. PERITOMA. One species (C. multicaulis) has 3-foliolate leaves not specially differentiated into bracts and of the same size and shape whether subtending flowers or not. In the other species of PERITOMA, however, all but the lowest bracts are 1-foliolate and easily distinguished from the multifoliolate cauline leaves. cause racemes tend to be densely flowered at the apices and young bracts are very small, inflorescences may at first appear to be ebracteate. However, in time the bracts become foliaceous and after flowering the raceme axis elongates greatly, spacing bracts as well as the pedicels, so that the lower, older portions of a raceme are open and clearly bracteate (fig. 8a). Of the many flowers in each raceme, most arc deciduous, only a few ovaries remaining to mature into siliques at well- spaced, cyclic intervals. Most older bracts therefore subtend only pedicellar scars. Since Cleome racemes are indeterminate inflorescences, they continue apical growth for months, becoming quite long, over 6 dm. in some species. The indi- vidual racemes are clearly separate. Racemes and bracts in Cleomella, particularly in the primitive C. Jongipes and C. angustifolia (figs. 4 and 8B) are much like those in Cleome. In some specialized species 3-foliolate bracts, which look like cauline leaves, occur throughout (C. brevipes, C. obtusifolia, C. mexicana; cf. Iltis, 1956). In other species a tendency toward bract reduction is quite evident, and racemes are often apically ebracteate. Plants of Cleomella tend to be smaller and more branched than those of Cleome. Their racemes, though separate, are more clustered and usually appear after less vegetative growth. Fruiting racemes are denser, for many more flowers mature into silicles. In Wislizenia (figs. 5 and 8c), racemes are always ebracteate, sometimes resem- bling nearly bractless extremes of certain Cleomella species (C. longipes). Some forms of W. r. refracta or W. r. melilotoides have profusely branched stems, each branch tipped by a raceme 1—2 cm. long (fig. 5). From this specialized type we can go all the way to the primitive var. palmeri where little-branched stems bear a few racemes 10-20 cm. long. Mature racemes of Wislizemia are very crowded (fig. 8c), because up to 95 Per cent of the flowers develop into fruits, while at the same time the raceme axes elongate only slightly after flowering. Though these inflorescenses are structurally racemes and therefore indeterminate, they are determinate in that they grow often less than 1 or 2 cm. in W. r. refracta and W. r. melilotoides, before growth is curtailed, SC If a lateral branch system of Wislizenia (lower right, fig. 5), with its many short inflorescences, were telescoped into a head-like “bur” about 149 its original size, one of those peculiarly congested, axillary, compound inflorescences character- istic of Oxystylis lutea (fig. 6) would be obtained. Each of these raceme clusters [Vor. 44 86 ANNALS OF THE MISSOURI BOTANICAL GARDEN is borne singly in the axil of a well-developed leaf, and each leaf, even the very first, subtends such a cluster. The center of each cluster is composed of one to several stout, short, straight or branched central axes from which diverge 5-15 stubby side branches, densely covered with flowers or fruits. Each side branch, rarely over 1 cm. long, represents a raceme and may bear as many as 40 flowers though not all develop into fruit (fig. 8D,E). That this congested multitude of racemes is homologous to a branch system of Wislizenia is further borne out by the occasional occurrence of a large, normal leaf on the central axis at the base of one of the side branches (fig. 6, bottom right hand "bur"). Such leaves correspond to those usually found in Wislizenia at the base of each raceme. "le WZ V sg JC » a ) — Ta 4 Gj í er Vaz HABIT DIAGRAM J} N .. Fig. 6. Highly congested ebracteate inflorescence clusters of Oxystylis. Plant on right, with mature fruits; plant on left, very young and in flower (after Jaeger, 1941). Note exceedingly short region of vegetative growth before flower production. The high development of four tendencies is responsible for the uniqueness of the Oxystylis inflorescence: (1) the essentially determinate, short growth of ra- cemes; (2) the nearly total lack of post-floral elongation of raceme axes; (3) the almost complete elimination of non-floral sections of branch systems; and (4) early initiation of flowering. Thus again we have traced modification from primitiveness in Cleome to specialization in Oxystylis. 1957] ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 87 3. Modifications of the Stipules. — We can trace modifications of stipules from Cleome, in which they are gener- ally lacking, through Cleomella and Wislizenia, with progressively larger stipular bristles, to their eventual fusion into the flat complex structure in Oxystylis. Their occurrence in sixteen of the eighteen species under discussion emphasizes the phyletic coherence of the genera. All species of Cleome sect. PERITOMA, except C. isomeris and C. platycarpa, have scarious, yellowish, bristle-like stipules, from 1-2 mm. long, occurring either Omm STEM A STEM ki 1 2 STP e B CLEOMELLA A GLEOME LUTEA ANGUSTIFOLIA È STEM haa \ STEM ue d = : Se d LE = Sy "Nae , 4, } pune A397 Ge SE are ‘WISLIZENIA j REFRACTA D oXYSTYLIS LUTEA Fig. 7. Stipules of cauline leaves. singly or in tufts of 2 or 3, on either side of the petiole base, of both cauline leaves and bracts. Stipules of Cleomella generally resemble those of Cleome, but in some species tend to be longer (1—4 mm.) , each tuft having 3-8 bristles. The largest are those of C. obtusifolia, noted by Watson (Gray, 1895), who erroneously cited the ab- sence of stipules in all other species as a key character! Jepson (1936) made a imilar error. . As in Cleomella, stipules in Wislizenia® occur in tufts each having up to 10 basally joined bristles of unequal length. ———M. S These ihe only stipules of the Cleomoideae regularly mentioned in the literature (Munz, 1935; Jepson, 1936; Pax and Hoffman, 1936). [Vor. 44 88 ANNALS OF THE MISSOURI BOTANICAL GARDEN In Cleome and Cleomella stipules are associated with the petiole bases of both the cauline leaves and flower-subtending bracts. Wislizenia and Oxystylis have cauline leaves with stipules but ao bracts! Yet on the axis of the inflorescence, just beneath each pedicel base, well-developed stipules are found in the place where the petiole of the bract-blade would have originated were the inflorescence bracteate. This loss of blade and petiole but retention of stipules is an interesting case of un- equal reduction of an organ. Furthermore, although the stipules on the inflores- cence occur occasionally as two separate tufts, usually they are fused into a single prominent strip which makes them more conspicuous than in the two bracteate genera. Oxystylis has the most specialized and the largest (2-4 mm. long) stipules of the four genera.) The other three genera have bristles which increase in num r, length, and basal cohesion, from Cleome to Wislizenia. In Oxystylis these trends reach a climax in rather peculiar structures (fig. 7p): flat, very thin, and trans- parent central blades with 3-8 greatly contorted, flagellate, slender teeth. The stipules are most abundant and most easily observed on the stubby bractless racemes (fig. 8, D and x), but they also occur paired on the stems at the base of each cauline leaf (fig. 7D). They may be interpreted as groups of bristles such as occur on Wislizenia, joined for half their length to form the flat lamina, with the free ends forming the teeth. 4. Other Trends.— In addition to the fruit, raceme, and stipule, there are less-prominent series which may be traced from Cleome to Oxystylis. Among these are reduction in size of flower and of seed, and minor changes in leaf shape. A summary of primitive versus advanced characteristics in the Cleomoideae has been published (Iltis, 1956). Contrary to Bessey’s dicta, the perennial habit of Cleomella perennis, C. mexi- cana, of some plants of Wislizenia r. palmeri, and of Cleome isomeris is definitely not primitive (Iltis, 1956). It can be interpreted as a specialized secondary char- acter which in the Clecmoideae appears sporadically in several groups of grassland and desert species through convergent evolution. IV. INTERPRETATIONs OF THE REDUCTIONS AND THEIR RELATIONS TO ARIDITY A study of the plant's environment considered in conjunction with its mor- phology may help to interpret the structural trends of the North American Cleo- moideae from the primitive (Cleome sect, PERITOMA) to the most specialized genus (Oxystylis) . l. Distributions in Relation to tbe Prevailing Climate.— The distribution of the four genera corresponds roughly to prevailing climatic patterns in western North America (figs. 9, 12). The comparatively unspecialized sect. PERITOMA of Cleome is widespread, occurring generally in more mesic regions, ? Despite the large size and peculiar structure of Oxystylis stipules, botanists have failed to notice them; some have even stated that they are lacking (Jepson, 1936). 1957] ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 89 "a 5 D M, v 1 Aa vie LN ¢; ( ANGUSTIFOLIA -D F oxvsrvuis LUTEA E s S d E ET i LIZENIA == S@ REFRACTA Fig. 8. Portions of inflorescences with stipules: D, drawn on same scale; enlarged portion of D i stipule size and length. A- " a E, ote progressive condensation of raceme axis and increase in in less alkaline habitats, and at higher elevations. The more specialized genera, Cleomella and Wislizenia, occur in progressively more arid and alkaline habitats of lower elevations, and the most specialized genus, Oxystylis, is limited to the most arid and alkaline area of the continent, the Death Valley region of California and Nevada. This correlation between arid climate and morphological specialization’ — 1 Thi i : iali i ithin Cleome sect. This applies not only to the generic level! The most specialized species within Cle ^ PERITOMA (sparsifolia) and in erg (obtusifolia, parviflora, brevipes) are likewise associated with Some of the most alkaline, arid habitats of California and Nevada. [Vor. 44 90 ANNALS OF THE MISSOURI BOTANICAL GARDEN Sech `~ | IZA CLEOME, -S _SECT. PERITOMA E WISLIZENIA ` EEE OXYSTYLIS le — —— = ——& o — M Mrá— ee "wé C ino Fig. 9. Total distribution of the western North American genera of Cleomoideae. implies that these modifications are related in some way to desert environment, and it is tempting to make the conventional assumption that they are "adaptations" and have “survival value". 2. Climatic Notes, Especially on Deatb Valley and tbe Mojave Desert.— In nearly all environmental aspects, Death Valley is extreme (Thompson, 1929; Anon., 1954; Went, 1955). According to Munz (1935): [mean?] temperature in the Mojave Desert is about 10 ° F., this engen, between ` five and eight months out of the year extreme Sg of 134° E at Greenland the wettest year (Went, 1955)], and a minimum of zero climate is one of unusually high summer tem- perature with resulting low humidity and high evaporation. [italics mine] 1957] ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 91 Coville (1893) states: The observations on the — gei of Ge Miri in 1891 for the hottest month of the year, July, show an B ese h e ximum , an average daily minimum of 87°, and a mean of 102? F. A m um — be? 122 WH five times during the summer; these temperatures were ben under cover of a er constructed especially to eliminate the effect of dës iation from the ground... In Pre sunlight bodies ien sa ly absorb heat must reach a temperature many deeg | in excess of the recorded max According to Parish (1930) the record of 134? F., “when the mercury rose to the top of the tube", would have been exceeded in an instrument of greater range, and "such temperatures in the shade would indicate at least 150? F. in the sun". This temperature “is the highest natural-air temperature that has been recorded on the earth's surface by means of a tested, standard thermometer exposed in a standard ventilated instrument shelter" (Thompson, 1929).!! Soil temperatures must be even higher. MacDougal (1908) reports 148° F. in sandy soil surrounding roots of annuals at the Grand Canyon in Arizona. On a sunny summer day, soil temperatures in Death Valley must certainly greatly exceed 48° F.! Precipitation varies from year to year, with the average annual rainfall (Went, 1955) 1.35 inches. Over a ten-year period (Thompson, 1929) the average monthly rainfall during November (0.20"), January (0.48"), and February (0.40"), accounted for 64 per cent of the yearly precipitation. During the rest of the year only rarely was there rain in excess of 0.10 inches per month, and in 64 out of 114 consecutive months there was no rain whatever. The rains that do come are usually in the form of hard, short showers. 3. Some Prerequisites for Desert Survival.— With the environment so extreme and water supply so limited, a plant has to be specially adapted to survive; in short, it has "to make hay while it rains". The ability of seeds to take up water quickly after one of the rare showers, rapid ger- mination, a short life cycle,!2 economy in growth and transpiration, and as rapid and large an output of seeds as the circumstances permit are all adaptations which seem essential to survival of desert annuals. The fruit and raceme modifications of the xerophytic Cleomoideae seem to have "developed in response to” two of the major demands of a desert environment: (1) the immediate and rapid germination of the seed during the rare times when water is available, and (2) the subsequent most rapid, most direct (or most economical) completion of the life cycle, i.e. Production of seeds. —— gy world record was recently exceeded in Tripoli (136.4°, fide Anon., 1954) Dr. W. E Blair (personal puren gg ts tells me that in some Aen (e.g. Sell crate : engt the immature aquatic phase of the life cycle is generally cor with aridity of native habitat. Toads in mesophytic areas reach the adult stage — gud mon e Of larval existence, while those in some parts of Texas complete metamorphosis in as s as twenty days. [Vor. 44 92 ANNALS OF THE MISSOURI BOTANICAL GARDEN 4. Xeropbytic vs. Mesophytic Cleomoideae.— The differences between mesophytic and xerophytic Cleomoideae, of both Old and New World (cf. Iltis, 1956), may be generalized as follows: Mesophytes, as a rule, are 1—6 m. tall, have fewer, thinner, larger leaves with 3—13 narrowly elliptic, flat leaflets with long-attenuate apices, and produce flowers after considerable vege- tative growth. Their fruits, commonly borne on long gynophores, are usually 5-10 cm. long (some 25 cm.!) and usually produce many fairly large seeds. The number of mature fruits per plant is frequently low. Xerophytes are rarely over 2 m., usu- ally less than 1 m. tall, and have many small leaves. The 1—3 (—5, or rarely more) leaflets are often conduplicate, with obtuse to rounded or emarginate apices, those of the American species generally glaucous and fleshy. Some species flower and fruit when only 10 cm. tall; the Saharan Cleome scaposa may fruit when only 3 cm. tall (Kachkarov and Korovine, 1942) and Cleomella brevipes when only 1 cm. tall (Iltis, 1955b). The small fruits, often 0.5—3.0 cm. long, have many small seeds or few larger ones. The gynophore, often lacking or short, frequently reflexes over the pedicel. The number of mature fruits per plant tends to be very high. The above generalizations apply to most desert Cleomoideae, including such widely separated, unrelated groups as the western North American species, the West Indian “uni- foliolate” species of Cleome, the Eurasian Cleome sect. THYLLACOPHORA Franch., etc. The occurrence of these characteristics in widely separated groups of desert plants is evidently due to convergent, adaptive evolution (cf. Briquet, 1914). 5. Interpretations of the Reductions.— Most species of Cleome sect. PERITOMA and the primitive taxa of Cleomella (fig. 4) grow tall, have many leaves, greatly accrescent, bracteate inflorescences (fig. 8A) and large, many-seeded fruits, all structures requiring relatively large quantities of water and organic materials, and long growing seasons. Most taxa of Cleomella (fig. 4) are smaller and often more branched than those of Cleome sect. PERITOMA. Inflorescences are shorter and frequently nearly ebrac- teate, and may appear earlier in the life cycle. The fruits are smaller, have fewer seeds, and mature more rapidly. The earlier appearance of flowers and the smaller fruit would result in shorter, or more economical life cycles. Young plants of Oxystylis flower almost immediately after germination, with the earliest inflorescences in the axil of the first or second foliage leaf, an inch or less above ground (fig. 6). Compared with primitive species of Cleome or Cleo- mella (fig. 4), these plants do not waste time or water on much vegetative growth. The stems of Oxystylis are usually “unbranched”; its leaves are few. The inflores- cences do not grow indefinitely as in Cleome, but after flowering they elongate at most a few millimeters, in sharp contrast to the striking postfloral elongation of the peduncle in most other Cleomoideae (fig. 8). Since the two-seeded fruits are very small, maturation is rapid. The lowest inflorescences of a plant 3 dm. tall may carry nearly mature fruits while the upper are still in flower. o summarize: The series from Cleome sect. PERITOMA to Oxystylis shows 1957] ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 93 progressively less vegetative growth, increasingly early flower production, and more rapid maturation of fruits, the latter reflected in the reduction of fruit, pedicel and gynophore length, and in inflorescence elongation (cf. fig. 8). That these are adaptive modifications seems obvious. For example, if a 3 dm. tall Cleome serrulata should perish through environmental vicissitudes, it would probably leave no seeds to perpetuate the species, while even smaller plants of Oxystylis would have mature seeds. To quote Parish (1903) It is in these LÀ areas that there is the greatest dere a of the protective adapta- tions which enable a plant most fully to utilize a scanty supply of water. The methods by which this is Käsch are three: by habits of growth; by pro picis for storing supplies of in times si T as reserves A times of need; and by contrivances for ugh evapor The first of these Deen is xd exemplified by most of the Kremer annuals. They spring up at once after light rains and p e rth no more than a leaf or two before proceeding to the production of a flower and a Ge If moisture now E reproduction is assu red; should it continue to be s eent. branches a are sent out and flowers and seed multiplie TI us : DA when receiving x a little moisture may fulfill the eyele = existence and provide the continuance ai ts specie $, "ees at d, an inch o ature; but under more ea conditions it may attain teg of two or three feet. Any change, therefore, that resulted in more economical vegetative growth and consequently more rapid production of flowers and seeds in some of the ancestral populations of Cleome sect. PERITOMA would have been advantageous either to their survival in a climate becoming progressively more arid or in the invasion of more arid regions. Those populations of Cleome that adapted to aridity by reducing the length of vegetative and fruit-producing periods appear to have given rise to Cleo- mella, whose ancestors, by the same process, produced Wislizenia and Oxystylis. That reduced silique size probably evolved as a result of selection favoring shorter maturation periods and not for other reasons of economy, such as seed number reduction per se, is borne out in another way. In some species of Cleomella, and Particularly in Wislizenia and Oxystylis, reduction in number of seeds from each flower is compensated for largely by the much greater number of flowers that mature into fruits. In fact, in some plants of Wislizenia nearly every flower pro- duces fruits (fig. 8) which clothe the inflorescence so densely that the total number of seeds is greater than that of most plants of Cleome. 6. Interpretation of the Schizocarp Tendency.— The evolution of the fruit reductions appears to have been indirectly associated with that of the schizocarp tendency in Wislizenia and Oxystylis where one or two seeds are permanently enclosed by the ovary wall (valve) (figs. 2, 10). This addi- tional "seed coat” might help the plant to survive in deserts since it facilitates water uptake and retention, resulting in rapid seed germination, a well-known character- istic of desert plants. Thus, Bell (1953) reports a minimum germination time of nineteen days for a desert species and twenty-nine days for a mesophytic species of ot Seed germination in desert areas is necessarily closely tied up with rare, rt and often violent showers.!* At other times, according to Coville (1893), A discussion of seed germination in Death Valley and its relation to hard rains is given by Wen: (1955). TVOoL. 44 94 ANNALS OF THE MISSOURI BOTANICAL GARDEN 0. Dehiscence of fruits and evolution of schizocarp tendency: A, Cleome lutea Fig. 1 ir var. lutea; B, Cleomella longipes; C, Cleomella obtusifolia var. taurocranos; D, Wislizenia refracta var. palmeri; E, Oxystylis lutea. water is practically non-existent since rain water quickly re-evaporates into a very dry atmosphere or is soaked up by a parched earth. The ovary wall appears to function in two ways: (1) the soft tissues of the valve imbibe water and swell (fig. 10 c—, black areas); (2) by capillarity, water enters through the mouth of the urceolate valve and fills the narrow spaces between valve and seed (fig. 10 c—e, black dots). Imbibition is probably of secondary im- portance in the development of the schizocarps since many Cleomoideae with free- falling seeds already have an outer testa with large blister-like cells which swell up when wet, thus retaining water for a time (e.g. Cleome serrulata, C. lutea, Cleo- mella hillmanii in North America; Cleome arabica, with very long, elongate hair cells, in the Sahara, etc.) Capillarity, however, appears to me to have been im- portant in the schizocarp evolution. We may postulate that the tendency started in an ancestral Cleomella with fruits like those of C. obtusifolia var. taurocranos (fig. 10c), where one seed is loosely held by a constriction near the tip of the funnel- shaped valve. 1f such a valve falls off and becomes filled with water, the captive seed will have more water available for a longer period and hence a greater advan- — tage over the free-falling seed in the availability of sufficient water to commence — germination. Such a hypothetical race gave rise to Wislizenia. Here the valve is urceolate with a small basal opening, through which, on immersion, water enters - filling all spaces between the seeds and valve wall, the latter, in addition, becoming ` soft and spongy (fig. 100). The one or two seeds are thus held in a water-filled ` "pitcher" which dries out slowly. 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 95 If this interpretation is correct, selective forces that furthered seed inclusion must have acted during the earliest germination phases of the life cycle. In con- trast, selective forces that produced smaller fruit and foreshortened branches must have acted on later stages: vegetative growth, flowering, and fruit maturation. The two sets of adaptations, though produced by independent selection factors during different stages of the life cycle, are nevertheless historically interconnected since such schizocarpous fruits could not bave been produced from Cleome siliques unless preceded by drastic size reduction. Not until fruits were as small and as laterally expanded as those of Cleomella was there any chance for the valves to curve around and imprison seeds. The fruit-size reduction from Cleome to Cleo- mella was then a prerequisite, a preadaptation, on which the schizocarp tendency of Wislizenia and Oxystylis was based. 7. Differential Selection and the Critical Phase of the Life Cycle.— That fruit-maturation and pre-germination are the phases of the life cycle in these xerophytes during which selection pressures are greatest is suggested by the pronounced modification of fruits and racemes, and by the relative lack of m cation in most other plant parts. In the Cleomoideae, desert survival is evidently not as closely tied up with the floral phases. The flowers, though specialized, are similar in all the genera, without any pronounced morphological trends as in the fruits or racemes, save decrease in size. There are sporadic instances of highly specialized floral structures, such as petal-nectaries 4 Cleome sparsifolia) and various specialized nectariferous disks (Cleome and Cleomella). The thickish leaves, gen- erally adapted to xerophytism and similar throughout, were changed only slightly, mainly from narrow and pointed to broad and emarginate leaflets These specific points agree with the generalizations of Stebbins (1949) that in sedentary plants "problems connected with FER and dispersal are far more important than those involved with maintenance". They also agree with many examples given by Stebbins of families (Goes Cruciferae, etc.) in which flowers and leaves vary little, while fruits and inflorescences show a wealth of modifications. Although the interpretations of the reductions are mostly conjectural, it would not be difficult in the Southwest to obtain data on the exact length of life cycle phases of each taxon, or to compare length of germination of artificially freed seeds of Wislizenia with that of seeds enclosed by valves. 8. Orthoselection.— It is evident from the morphological series (fig. 1) that we are dealing with a kind of orthogenesis, in the descriptive sense only, which can be interpreted as the result of orthoselection (Huxley, 1943). This assumes that the same environ- mental forces that produced Cleomella continued on (though in increasing in- tensities) to produce Wislizenia and Oxystylis, and that if the same selective forces are operating on the same phyletic material, the same adaptive trends will continue. TVor. 44 96 ANNALS OF THE MISSOURI BOTANICAL GARDEN Furthermore, a decreased gene pool resulting from high selection would seem to permit a group to evolve only within the limits of its available genetic potential. 9. Classification of Modifications.— imofeeff-Ressovsky (1940) classified morphological evolutionary trends into three groups: (1) Neutral, which cannot be interpreted as obvious adaptations, there being no plausible relation between structure and environment. The larger stipules, or the tendency of the gynophore to deflex over the pedicel (Wislizenia, Cleomella spp.) are placed here though they might be consequential to certain other adaptive trends. (2) Adaptive, where a plausible relationship between morphology and environ- ment is evident. Most reductions studied here appear to belong to this group. (3) Harmonious modification. In examining fruiting racemes of Wislizenia or Oxystylis, it is obvious that raceme foreshortening could not have taken place with large siliques but became possible after the fruits were reduced in size. Con- versely, the many fruits could not have matured if the siliques were large because of spatial interference and problems of nutritive supply. Likewise, bract reduction is also prerequisite to foreshortening for there is not enough space for each Oxystylis flower to be supplied with a bract. Another “harmonious reduction” is the decrease in flower size with greater raceme density. V. GEOGRAPHIC DISTRIBUTION 1. Geographical Distribution of the Genera and the Degree of Allopatry or Sympatry of the Species. — Cleome, with 180 taxa, is of almost world-wide tropical and sub-tropical distri- bution (fig. 11).1* Of the 55 species and 25 subspecies indigenous in the Western Hemisphere, most are tropical. The six species of sect. PERITOMA are restricted to temperate, rarely subtropical, North America (figs. 12a, 13a) with their center of variation in eastern California and adjoining Nevada, where four species occur. All six species are to some extent sympatric, some with an overlap of ranges that is complete or nearly so. The total range of Cleomella, (figs. 128, 138) though smaller than that of Cleome sect. PERITOMA, is similar in outline. The distribution of individual species is rather different. The five eastern and Mexican species are allopatric with sharply separated ranges. On the other hand, the Great Basin and Pacific slope area, with five species, contains many regions where two or three species occur together, but only one small region in south-central California, the center of variation of Cleo- mella, where four species are sympatric. 14 Distribution maps have been plotted from material of major U. S. herbaria (Iltis, 1955b); particularly the Missouri Botanical Garden Herbarium, and from sources that list individual local- ities, such as floristic works. Especially helpful was the unpublished thesis of Ethel Crum (1929); dealing with Cleome isomeris, Cle I izenia, which cites many specimens. A microfilm CO i i : ` rnt Annetta Carter, to whom I should like to express my thanks. Generalized ranges, as given in floras, were found to be unreliable and were not used. 19571] ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 97 Fig. 11. Distribution of the genus Cleome: black with white dots, indigenous; black dots, adventive. Old World, about 90 taxa; New World, 79 taxa. Wislizenia refracta occurs in the Southwest and adjoining Mexico. Its range is smaller than that of Cleomella and appears to be more or less continuous (figs. 12c, 13c), forming a horseshoe around the northern end of the Sierra Madre Occi- dental Wéislizenia spread as a weed into the San Joaquin Valley and other parts of California during historic times (Jepson, 1936), so that its natural range is smaller than the present-day, artificial range. Its morphologically diverse populations appear to be allopatric. Oxystylis lutea is limited to deserts and washes of the Death Valley region in California and Nevada (fig. 12D, 13D). Most collections come from the Valley itself along the Amargosa "River" (marked on fig. 12D by a black line) where the species is reported to be abundant (Jaeger, 1941). A few specimens have been collected further north in Nevada. Oxystylis is a monomorphic taxon, and its populations are, therefore, allopatric. : The total generic ranges progressively decrease in size from the generalized Cleome sect. PERITOMA to the specialized Oxystylis. The regions in which the most widely distributed group (PERITOMA) occurs are generally rather mesic, while those of Cleomella, Wislizenia and Oxystylis are progressively more xeric, with Oxystylis thriving in the driest and hottest climate of the continent. As to the geographic distribution within a genus, taxa in Cleome sect. PERI- TOMa are very highly sympatric; in Cleomella both sympatric and allopatric; and in the monotypic Wislizenia and Oxystylis the taxa (being the more or less differ- entiated populations) are allopatric. 2. Geographic Distribution and Origin of Genera.— Taxa frequently evolve through geographic separation of one part of a popula- tion from another, and subsequent genetic and morphological differentiation of either or both of these isolated populations through different or unequal selection Pressures (Mayr, 1942). That part of the original population which is isolated in [Vor. 44 98 ANNALS OF THE MISSOURI BOTANICAL GARDEN » DISTRIBUT aya: z| DISTRIBUTION OF WISLIZENIA IBUTION OF THE SIX SPECIES OF CLEOME SECT. PERITO, Or HIGHLY Sant Set VI DISTRIBUTION OF OXYSTYLIS ` ONE UNIFORM SPECIES EISTOCENE pl NY NA d : ^ D TH Sec See HEIL [* udi d e E: LD Ü SEN S GT IS LIST TS IA ‘i EE d — eH H | | PP Ce VEH ei CAN TEEN ea A ama LEJJITT ONES S VAN CLEOME MULTICAULIS Ae = * CLEOME PLATYCARPA Vamos ?] \ OME LUTEA S AS. |. CLEOME LUTEA ` —---—--- LENS ( ta edet; nm e eene! D ——— OME SPARSIFOL|A cs Fig. 12. Distribution of all taxa in the western North American Cleomoideae and the degree of sympatry and allopatry in each genus, an environment fast changing to a more xeric one will evolve rapidly. The other part, if its environment does not markedly change, may remain more generalized and will evolve slowly. Generally, the more primitive taxa (subspecies, species, genera) will be found in the more mesic climates, the more specialized in more xeric ones. If two taxa have thus evolved from one original population through geographic isolation they would a£ least for a time occupy allopatric ranges. By deciding which of the two related taxa is more specialized and which more primitive we can speculate on the geographic direction of evolution and on the geographic origin of taxa. That such a scheme is applicable only to relatively recent evolution need not be emphasized (cf. Deevey, 1949). Wagner’s ideas as amplified in “Jordan’s Law” or “Rule of Proximity”, that “given any species in any region, the nearest related species is not likely to be ound in the same region nor in a remote region, but in a neighboring district separated from the first by a barrier of some sort” (Cain, 1944), has been found to apply chiefly to species or subspecies within the same genus. The fact that it — apparently also applies to related genera, or to closely related species, one in one — genus and one in another, greatly extends Jordan's Law. However, such com- — plementary ranges appear here too frequently to be accidental. In the western Ce North American Cleomoideae, the geographic ranges of the two most closely d 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 99 related taxa of two related genera or, sometimes, the total ranges of two related genera, are highly allopatric, a trend that appears congruent with the phylogeny as outlined on morphological grounds. Cleome.—The geographic origin of the genus Cleome is unknown, though the apparently primitive species occur in the New World tropics. Sect. PERITOMA is similar to somé New World Cleome species and apparently has been derived from them. The occurrence of C. multicaulis in central Mexico (fig. 124) suggests that section PERITOMA as a whole originated in the Mexican highlands, one of the two important distribution centers of Cleome in the New World (the other being South American), the area from which Madro-tertiary elements of the southwestern flora appear to have originated (Munz, 1935; Chaney, 1947; Axelrod, 1948). The total range of PERrroma is allopatric to that of the other sections!5 of Cleome (fig. 13a). In spite of the fact that the range of PERITOMA overlaps that of Cleomella nearly completely (fig. 9), offering little help in tracing the origin of Cleomella, we can profitably compare the ranges of the most Cleomella-like species of Cleome sect. PERITOMA (Cleoma lutea) with those of the most Cleome-like species of Cleomella (the midwestern Cleomella angustifolia and the Mexican C. longipes). All are very much alike in habit, inflorescence, and seed structure. Cleome lutea is clearly allopatric to both species of Cleomella, although its range nearly touches theirs in two places (fig. 135). The range of Wislizenia complements that of Cleomella (fig. 13c) except for a major overlap in northern Mexico and adjacent New Mexico and Texas, two most closely related taxa of the two genera are Cleomella longipes and Wis- lizenia r. palmeri (a southwestern, primitive variety). Both are allopatric, being separated by the northern Sierra Madre Occidental. We can postulate, as shown y arrows in fig. 13c, that Wislizenia originated in western Sonora, from a branch of Cleomella (similar to C. longipes ?) isolated by the rising Sierras, and subse- quently migrated north and east. Eventually, as W. r. refracta, the genus spread through the wide gap of the continental mountain axis in southern Arizona and New Mexico, thus reentering the territory of its ancestors (C. longipes), where the two genera are now sympatric. EE Wislizenia and Oxystylis are exactly allopatric (fig. 13D), their dividing line the "Death Valley drainage-Colorado River drainage" watersheds (Miller, 1950; Hubbs and Miller, 1948). Wislizenia, to my knowledge, has never been collected or reported from the Death Valley region (Coville, 1893, and others). The above data suggest the origin of these genera as follows (fig. 13A-D): Cleome sect. perrroma evolved from a mesophytic Mexican Cleome. In some drier regions, perhaps in northern Mexico, southern Arizona, New Mexico or Western Texas, one of the taxa of Cleome (C. lutea ?) gave rise to Cleomella which ——— tenuis Wats., which belong to a section (to be 15 zom. E $ è H T xcept for the unifoliolate species and C f d World taxa than to any of the New World. named in the near future) more closely related to Ol [Vor. 44 100 ANNALS OF THE MISSOURI BOTANICAL GARDEN TIED, Y X829 MEUM Piae za EY Ü ZV ee ei OME, SECT. PERITOMA air EN E ee CLEOME, ALL OTHER SECTIÓNS C. tenuis and unifoliolate species) ] > A Fig. 13. Geographic distribution and origin of genera. Arrows point from the less specialized to the two most closely related taxa, one in one genus and one in the other. spread and evolved into many species. The more primitive species apparently remained in the mesic regions with little climatic change, while the more special- ized evolved mainly in the California and Nevada deserts and in the Mexican plateau (Iltis, 1956). A southwestern primitive population of Cleomella (C. longipes ?) spread to Sonora and California and evolved into Wislizenia which, spreading north and east, evolved a number of variants. One of these (W. r. melilotoides ?) became isolated in the Death Valley region, where it evolved into Oxystylis. Although the center of variation (region with the most genera or species) for the western North American Cleomoideae lies within southern California and Nevada, their putative centers of origin appear to be more to the southeast. This supports Cain's (1944) contention that the two types of generic centers are not necessarily synonymous. VI. DEGREES oF INTER-GENERIC AND INTER-SPECIFIC RELATIONSHIP AND INTRA-SPECIFIC VARIABILITY Any systematist knows that there are different kinds of genera, just as there are different kinds of species. These differences between genera are in part due t0 variations in: (1) the degrees of inter-generic relationships within a group of closely related genera; (2) the degrees of inter-specific relationships; and (3) the intra-specific variability. 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE X CLEO SU c MELLA ` AZ Be Ka LA ANGUSTIFOLIA 7 Pre T Lo a NGI ee WESTERN DESERT ei S SE PALMERANA < " we ; PS : SHADED LJ SPECIES! 27 DOTS- WoL ZENS AREA: CLEOMELLA, d ‘ANGUSTIFOLIA IC Ch H 3 : ipn VCI PERENNIS C MEXICANA MF Se |! e d mmi - Si. Se ZER SEAN OXYSTYLIS | Se? f Y wet oe M IT CH : a3 Së SS -— y : f 1 D WISLIZENIA D i ^ è Fig. 13 (Continued). Valley drainage system Heavy line in D encloses the Death 101 [Vor. 44 102 ANNALS OF THE MISSOURI BOTANICAL GARDEN We can ask: (1) How closely are the genera in the group related to one another?; (2) How closely are the species in one genus related to each other?; and (3) How much does each species vary? For the taxa dealt with in this study, at least, the answers to these questions fall into well-defined patterns and correlate with the phyletic series previously discussed as well as with the geographic distribution and therefore prove signif- icant to phylogeny. (1) Degrees of Inter-generic Relationsbips.— Cleome can be considered the "mother genus,” the genus primordioides or prim- itivus of the Cleomoideae (Buxbaum, 1951), from which all other genera in the subfamily, the genera progressiva, have been derived. Whatever primitive relatives Cleome still may have, they must be sought in the Capparidoideae!® or, preferably in other families. As the primitive genus, Cleome contains the least specialized taxa within the subfamily as well as a considerable number of specialized ones that cannot be segregated conveniently from the primitive generic plexus. Thus, the natural section PERITOMA can be thought of as a specialized branch within the genus, a contention which is supported by the allopatry of perrroma with respect to the other sections of Cleome (fig. 13A). The exact position of PERITOMA within Cleome is not certain. The fused claws of the seed, the closed, convolute aestivation, and other characters indicate a fairly close relationship to the other New World species of the genus. Although the species of PERITOMA do not differ from these by any outstanding qualitative char- acters, they do differ in the way they combine various characters which occur in diverse groups of Cleome. Hence, a generic segregation of PERITOMA from Cleome, although natural, would be impractical, though McVaugh (1945) has made just such a recommendation for this type of situation. Cleomella is exceedingly close to Cleome sect. PERITOMA. Biologically, it is nothing but a specialized segment of the PERITOMA plexus which, because of its short, transversely expanded fruits, can be easily segregated. While there are two rather tenuous, quantitative differences between the two genera, the only qualita- tive difference, and then not an absolutely constant one, is the transverse expansion of the valves. If the fruit of certain, quite unrelated, South American species were placed in this sequence (e.g., C. siliculifera Eichl., fig. 1p) there would be no way of telling where one genus begins and the other ends. However, in North Amer- ica, the break between Cleomella and pERITOMA is fairly sharp, and Cleomella, a useful genus, is undoubtedly also a good biological unit. 15 The members of this subfamily are shrubs and trees of the tropics; in nearly all characteristics of fruit and seed, as well as in many flowers and leaves, they appear to be, if anything, more special- ized than the Cleomoideae, from which, contrary to current opinion, they may have been derived. 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 103 Wislizenia has close affinity with Cleomella, although it is not as close as be- tween Cleomella and Cleome, with a£ least three quantitative differences (reduction of seeds, bracts, replum) and one complex, qualitative difference (shape and orna- mentation of the fruit) to separate the genera. Oxystylis differs sharply from Wislizenia by at least three quantitative characters and two qualitative ones (extreme modification of replum; telescoped inflorescences) . Assigning one point for each "simple" quantitative inter-generic difference and two points for each qualitative difference, the morphological discontinuity between , any two genera follows a trend from Cleome to Oxystylis, with three points or less between Cleome and Cleomella, about five points between Cleomella and Wislizenia and seven or more points between Wislizenia and Oxystylis. Although this arbi- trary point system takes into account only the more obvious characteristics, the relative values are essentially correct as shown by fig. 1. This progressive increase of inter-generic difference, approximately represented by the numbers 3—5—7, sug- gests some accelerative type of evolution which will be discussed toward the end of this paper (cf. fig. 155). (2) and (3). Degrees of Inter-specific Relationships and Intra-specific Variability.— Cleome sect. PERITOMA is composed of six species so very distinct that some botanists have deemed it necessary to establish segregate genera for them: Isomeris Nutt. (for C. isomeris) ; Carsonia Greene (for C. sparsifolia) ; Peritoma Nutt. (for C. lutea, C. serrulata, and C. multicaulis) ; and Celome Greene (for C. platycarpa). Variability within the species of Cleome is not uniform. Some, such as C blatycarpa, C. multicaulis and C. sparsifolia, all taxa with comparatively small ranges, are monomorphic. Others, such as the widespread C. serrulata and C. lutea, vary considerably, mainly in silique shape and length. Cleome lutea is the only species easily divisible into geographic varieties, with C. l. lutea widespread and C. l jonesii Macbr. occurring from central Arizona to Baja California. That this divisible species is also the most primitive taxon of PERITOMA is significant. Cleome isomeris, highly variable in fruit shape, has been divided into three geographic Varieties which overlap and intergrade and are difficult to maintain. The ten species of Cleomella are more closely related to ane another and compose a more tightly-knit group than do the six species of Cleome sect. PERITOMA. owering plants of different species of Cleomella sometimes are almost indistin- guishable, a situation not encountered in PERITOMA. Great variability in fruit shape or in pedicel and gynophore length is characteristic of some species, as in the relatively widespread C. plocasperma or the more local C. palmerana and C. obtusi- folia. Most other species are monomorphic (Crum, 1929; Iltis, 1955b, 1956). No Species is divisible into geographic varieties. ds EU fferences of opinion regarding subgeneric (ie. subspecific) divisions within Wislizenia refracta sensu lato are due largely to the great variability in shape, size, [ Vor. 44 104 ANNALS OF THE MISSOURI BOTANICAL GARDEN ornamentation and divergence of fruiting valves and in the length of the raceme. ile the specialized smaller-fruited, obovoidal- and smooth-valved plants gener- ally occur in the eastern (W. r. refracta) and northern (W. r. melilotoides) parts, and the primitive larger-fruited, obconical- and spiny-valved plants (var. palmeri) in Baja California and Sonora, the variability does not lend itself easily to taxonomic resolution. The many local races apparently replace each other complementarily (Greene, 1906), without much of a break in morphological pattern. Ten species of Wislizenia have been described to date (Greene, 1906). Jepson (1936) considers all, except W. palmeri, as varieties of W. refracta. I. M. John- ston (1924) interprets the genus as monotypic and very variable, recognizing three geographic varieties, while Crum (1929) considers the primitive and specialized extremes and the intermediates as separate species, all subdivisible into a number o weak varieties. ` Wislizenia refracta sensu lato is indeed highly polymorphic. It deserves to be studied by the methods of modern systematists. Specimens of Oxystylis lutea are not common in herbaria and information about its variability rests on meager evidence. I have examined sixteen different collec- tions, usually several plants in each; it is doubtful whether very many more have ever been made. If there were any appreciable variability among them it was not apparent. VII. Discussion Any phylogenetic interpretations of the facts of morphology, geography and variability are highly subjective. In the words of Lam (1951) they “may have evolutionary significance, but . . . represent mo phylogeny, for the simple reason that these schemes are always one-time schemes, with no fossils whatever involved”. A different viewpoint, stressing the value of phylogenetic series to the understand- ing of taxonomy, is that expressed by Buxbaum (1951), with whom I concur. 1. Correlation of Geographic Distribution and Intra-generic Relationships with Morphology to Determine the Relative Age of a Genus.— e fruit, raceme and stipule morphology indicates that Cleome is the primitive member of this sequence and that Oxystylis is the most advanced; in other words as genera Cleome is the oldest and Oxystylis is the most recent. The direction of evolution is thus inferred and the relative age of each genus is implied on morpho- logical grounds. Is it possible to correlate patterns of geographic distribution and degrees of intra-generic relationships with the morphological evidence? The distinctiveness of species in Cleome sect. PERITOMA indicates that we are dealing with very old taxa. The species, as we know them, may be young, but the point of divergence of the evolutionary lines that have given rise to these species dates far back in time. Not only are these species very distinct morphologically, but the high degree of sympatry in the group indicates they must also be very distinct genetically (fig. 12a, 13a), for in many places two or three species grow — ` side by side without the slightest indication of hybridization or introgression. For — 1957] ILTIS——EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 105 instance, in the alkaline semi-deserts of the San Luis Valley of southern Colorado, Cleome serrulata and C. multicaulis, the only species in the section with purple flowers and, in some ways, closely related, grow together with branches intermixed and flowers often touching. Yet, despite careful searching, there was no evidence of hybridization. At Vernal, Utah, Cleome serrulata and the very similar C. lutea 8row in abundance on the same slope. Mass collections by Edgar Anderson show no evidence of introgression! It is almost axiomatic to declare that sympatric species, as a rule, are old. They have to be, for only after prolonged isolation are two sibling species likely to be sufficiently differentiated that they can exist side by side without re-merging. Therefore, geographically evolving populations have to be allopatric before they can become sympatric (Mayr, 1942). One would expect then that the genus Cleomella, with a Cleome-like group as its morphological prototype, would be younger than Cleome in other ways as well, and that this age difference would be reflected in the geographic distribution of the species. A glance at fig. 125 shows that this is true. Of the ten species of Cleo- mella, five are allopatric and five are more or less sympatric. As one would also expect in a younger genus, the species of Cleomella are usually much more closely related to each other than are those of Cleome. Also, the ranges of some of the allopatric species (e. g., the Mexican taxa) still bear the earmarks of having evolved from a once-widespread species through disjunction and subsequent differentiation (Iltis, 1956). The hypothetical wide-ranging species differentiating into smaller units has passed out of existence in Cleomella. However, it is well-exemplified by the “super-species” Wislizenia which in its distribution is in a younger stage of evolu- tion (fig. 14). The genus has a single widespread species with a scattered but not markedly disjunct range (fig. 12c). It is highly polymorphic, each of its many Populations differing from the next only in minutiae. One series of populations, the southwestern var palmeri, is morphologically distinct enough to be called a Variety, a subspecies, or, by some, even a species. Most of the other local races are so similar to each other that they have not always been dignified with even sup- specific recognition. In studying the diverse fruits, one can almost "see" this Species in an active stage of evolution, differentiating into smaller, more local units. Since these populations appear to replace each other geographically, the intra- specific (i.e. intra-generic) allopatry is complete. Oxystylis, the most specialized and advanced genus morphologically, is Leni Sg in its distribution-variation patterns. With one possible exception, its range is nearly continuous, and the variability within its single species negligible. : The following table gives the number of species in each genus with overlapping ranges (per cent of sympatry). 1 The sequence is quite clear: from the primitive genus with large, overlapping ranges and very distinct species, through the next genus with smaller, partly over- [Vor. 44 106 ANNALS OF THE MISSOURI BOTANICAL GARDEN lapping, partly separated ranges and more or less closely related species, through the third genus with small allopatric ranges of closely related populations, to the last, most specialized genus Oxystylis, where the single species is so uniform mor- phologically that ranges for individual populations cannot be plotted. If we start the sequence backwards from the youngest to the oldest, we can, in a sense, "see" the evolution of a genus exemplified by the accompanying conceptual scheme (fig. 14). Thus, a genus would have to start in an “Oxystylis” stage to get to a “Cleome” stage. Per cent Genus Number of species of Kinds of taxa sympatry 1 sympatric with 1 species 17 Cleome ; S ie 2 e 3. Species exceedingly distinct; ranges 6 spp. 1 = i : ` = often large or medium 1 an » 5 an 83 5 sympatric with no species 0 : Cleomella 1 o ge. ed 20 Species distinct to very closely 10 spp. 2 G WR SS 30 related; ranges medium to small 2 » » 4 » 4 0 oro cro tn ME SO ien ; x Ge p. s or Taxa (populations and geo raphical more: subsp 1 sympatric with no species 0 varieties) closely related; ranges divisions small and local ES eer Taxa (populations) uniform; Osyeiylis f individual ulations cannot 1 ic wi : of individual populati 1 sp. sympatric with no species 0 e e eg du morphol: gical grounds because of uniformity 2. Willis “Age and Area".— When the ranges of all four genera are plotted on the same map (fig. 9), they are seen to roughly follow a pattern of what has been called "progressive equifo areas”. Centering essentially around the range of Oxystylis, the ranges of Wis- lizenia, Cleomella, and Cleome sect. PERITOMA become progressively larger. Such a pattern seems, on the surface, to fit perfectly into Willis’s (1922) ideas of “Age and Area”. According to this theory, the youngest genus, Oxystylis, would occupy the smallest range because it has had little time in which to spread its seeds; E oldest genus, Cleome, would occupy the largest range because it has had eons of ` time in which to spread from its original home. 2 This reasoning, however, does not apply to the situation in the Cleomoideae. A — comparison of this distribution map with some generalized climatic maps will reveal 4 1957] ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 107 =<. 7 STAGE III "CLEOMELLA" STAGE I! "WISLIZENIA" STAGE ! "OXYSTYLIS" Fig. 14. Hypothetical scheme showing stages in the geographic evolution of genera as exemplified by the four genera of western North American Cleomoideae. Explanation in text. that the primitive Cleome usually occurs in more mesic regions and that the ranges of Cleomella, Wislizenia, and Oxystylis are progressively more restricted to regions of greater aridity. Since only those plants adapted to aridity were selected, the plants with "mesic" genes tended to be eliminated by progressively more arid environments. By the time such a highly modified plant as Oxystylis had been produced by the narrow Channels of selection, its genotype had been shorn of all but "super-xeric" genes, Producing a homogenic species (Stebbins, 1942). Its range is small, not because of the limited time in which its seeds could have been disseminated, but because of its narrow physiological tolerances. Selection produced a highly specialized plant, not only morphologically but physiologically and ecologically. Plants of Oxystylis, being biotypically poor, are literally the "prisoners" of their own genes. is is also true of the other genera, but less so. Most species of Cleome sect. PERITOMA are widespread but they have their limits, for none are native in the moister, wooded regions east of the Mississippi. In general, morphological conservatism is matched by relatively great physiological (ecological) tolerance; morphological specializa- tion by physiological intolerance (specialization).17 This statement applies to the over-all generic ranges as well as to specific ranges; thus, the most highly special- oO 17 Ss : $ i : OC Tie seges E Een Sold fuo, seems reasonable and has been discussed by Stebbins (1952). [VoL. 44 108 ANNALS OF THE MISSOURI BOTANICAL GARDEN ized taxa, Cleome sparsifolia, Cleomella obtusifolia var. taurocranos, C. brevipes, C. parviflora, etc., Wislizenia refracta var. melilotoides, all occur in the vicinity of Oxystylis (although not in Death Valley). Likewise, Cleomella mexicana, the most specialized of a phylad of three species, is restricted to the vicinity of two salt lakes in central Mexico (Iltis, 1956). 3. Tbe Probable Age of Oxystylis.— Since Oxyst3lis is apparently an indigenous endemic (Sinnott, 1916) of the Death Valley region and the range of its closest relative is allopatric to the south, it can be deduced that it is a highly specialized northern segment of Wislizenia which became isolated and evolved im Death Valley. The physiography of the region (fig. 13D) suggests this strongly, for the contemporary watershed between . the Colorado and Death Valley drainage systems (Miller, 1950) corresponds exactly to the line between the ranges of Oxystylis and Wislizenia (fig. 12). Furthermore, Miller (1946) postulates, on the basis of fish distribution and geologic evidence (Blackwelder and Ellsworth, 1934), that there was once a connection between the two basins in that area and that it may have served as a migratory route for Wislizenia. Death Valley functions as a kind of island whose isolation, together with the extreme climatic conditions, evidently furthered speciation. This is supported by the considerable number of endemic taxa known from there, both plants (Munz, 1935; Dressler, 1954; Tryon, 1955 [Selaginella leucobryoides]) and animals (Miller, 1950), which appear to be indigenous endemics usually with southern affinities that evolved from relics of once more widely distributed taxa (apparently not plain relics as Miller states). Living in that region, they must be adapted to its present- day climate of extreme heat and drought. If these adaptations originated in situ, they are probably recent, since, as Parish (1930) phrases it, Death Valley and the Mojave Desert had once a “greener youth,” for "there is abundant evidence that in the Pleistocene Age, perhaps 20,000 years ago, this whole region enjoyed an abun- dant rainfall, and—as a necessary sequence—must have had vegetation very dif- ferent in its character from that which prevails today." Many authors (Campbell and Campbell, 1937; Antevs, 1937; Miller, 1946; Hubbs and Miller, 1948) present strong evidence that the whole region of eastern California was one of large, deep lakes in the late Pleistocene or even more recently. Blackwelder and Ellsworth (1934) think that an immense Pleistocene lake filled Death Valley, whose waters overflowed to the south into troughs that eventually emptied into the Colorado River. Miller (1946) names the integrated river system of that time the “Death Valley system” because the Death Valley formed the sump for its waters. Accord- ing to Antevs, the lakes probably date from the last pluvial period, between 20,000 and 25,000 years ago, and Lake Mojave may have been at its fullest as late as 15,000 years ago. Indians then lived in this region (Campbell and Campbell, 1937, and others in same symposium) which, in its aridity and alkalinity, is now desolate an forbidding to human habitation. 19571 ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 109 It is unlikely that Oxystylis, or its ancestral forms similar to Wislizenia, were able to grow in Death Valley during this moister period! It is also unlikely that Oxystylis or its ancestors migrated to Death Valley from other places than those to the south, since the northern regions were then even cooler and moister than they are now. There is furthermore no evidence that Oxystylis has ever grown outside its present range. “The distribution of the existing Flora of California indicates that the majority of the endemic species originated in their present area" (Campbell and Wiggins, 1947). Instead, we might postulate that the Wislizenia- like ancestor of Oxys£ylis migrated north after the Death Valley region had become about as arid as the habitats in which Wislizenia grows now and that the evolution of Oxystylis took place within Death Valley in the extremely short time of 15,000 years or less. Fifteen millenia for the evolution of a genus seems short indeed, even for the evolution of a species under most circumstances! According to Mangelsdorf 2): volution is usually dee as a process which moves with m E unhurried dern whos illion Bech — gical periods e progress is measured in mi of yea . and h . is ord paris thought ES as so slow that even tbe second hand on the deeg clock does oio perceptibly in dw span of a human lifetime or, in many cases, during the period. ad kafe human hist In some areas and in some taxa the "evolutionary clock" has practically stood still for a tremendous time span as, for example, in the temperate mesophytic wooded regions of the southeastern United States and eastern China. In many genera, the differentiations that have occurred there since the late Eocene or Mio- cene, 30 million or more years ago, have been on specific, varietal or even lower levels (Cain, 1943; Sharp, 1951; Li, 1952). Yet, as Mangelsdorf and others find exception to this condition in the very rapid evolution of animals and plants under domestication, we may postulate exceptions in organisms, particularly annual plants, which live in arid environments or which find themselves in environments changing rapidly toward aridity.!? Xerophytic and cultivated plants have certain similarities as to environment and evolution. Both occur ipii: in full Cen and in acia soils. Both are often saniert in whic h evolu- tion has been mos Both favor open habitats with h little eet d mpetition. For cultivated annuals, ‘the ti is kept open artificially by weeding and plow Ch annual xero- Phytes, the habitat is also open in the sense that: (1) a desert wash o z playa i is ot disturbed and disrupted YI Y vith bare = and sand a EE of sporadic periodic hard floods; ry Gi annuals being th e characters plants of a Lem their habitat is not closed or preempted at the beginning of de? growing season by sii already there, as in a sedge mead i : est high number of vegetatively reproduced perennials. Thus the os of — ul ge tions in deserts is ofte ery gre a good on, the nearly pure patches o: See species resembling those of DE plants in a field of send where E of interspecific com- h partl cultigens. Éiere vironment, which periodically allows only a few favored to grow [Vor. 44 110 ANNALS OF THE MISSOURI BOTANICAL GARDEN 4. Arid Environment and Rapid Evolution.— That and hot environments favor more rapid evolution than do mesophytic organisms, for example, has been substantiated by studies on fossil and recent or- ganisms as well as on theoretical grounds (Muller, 1942, and Stebbins, 1947, 1949, for plants; Miller, 1950, for fishes of desert hot springs). This seems to be due to the following factors (Stebbins, 1952): (1) Ecological fragmentation of habitats, and therefore populations, in arid regions is much more pronounced than in mesic ones, thus producing many situations favoring isolation and consequently specia- tion. With climatic fluctuations one can expect rápidly changing selection, and with the remerging of habitats and populations, introgression. (2) The fragmen- tation of the habitats produces large or medium-sized populations with isolated sub- units interchanging genes only occasionally, a situation most favorable to genetic drift and therefore to rapid evolution. (3) The number of special adaptive struc- tures that can be produced in these environments may be, and often is, very large; (4) the selective pressures are very high because of the extreme climate; and (5); in annuals only, the brevity of the life cycle and the rapid multiplication of genera- tions would increase the evolutionary rate (Sinnott, 1916). A factor not mentioned by Stebbins, but discussed for fishes by Miller (1950), is the increase in mutation rate due to direct influence of heat. Very little ex- perimental evidence is available on this subject. Navashin (ex Babcock, 1947) reported that the mutation rate in resting seeds of Crepis tectorum was accelerated by increased temperatures. Shkvarnikov and Navashin (1934) found that when resting seeds of C. tectorum were exposed to temperatures of 54° to 55° C. for 20 to 44 days, the surviving seedlings showed chromosomal abnormalities, par- ticularly translocations, and an increase in the mutation rate. They call attention to the fact that since heat is widely distributed in the world, it may play a con- siderable role in the production of mutations; and that resting seeds are often exposed to temperatures equal to or higher than those used in their experiments. This work is not mentioned in “Temperature and Evolution” (Plough, 1942, and others in same symposium) which summarizes evidence, mainly from Drosophila, that high temperatures (Muller 1928, 1942), as well as temperature shocks (a distinct phenomenon!), increase mutation rates. Fankhauser (1942), working with Triturus, showed that extreme temperatures in the laboratory induce poly- ploidy, and he suggested that heat might produce polyploid individuals in nature. That resting seeds of Cleomoideae, particularly Oxystylis, are exposed to ex- treme heat is evident, considering that Death Valley shade temperatures may be as high as 134° F. and soil temperatures certainly much over 160° F. Temperature shocks may be provided in the tremendous diurnal fluctuations which, because of low humidity, are rather typical of deserts. | Miller (1946, 1948, 1950) studied the effects of desert habitats on the evolution ` of Death Valley fishes. He concluded that if Lake Manly (the Pleistocene Death — Valley lake) was correlated with the Tahoe stage of glaciation as postulated by — * 1957] ILTIS—-EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 111 Blackwelder, the two Death Valley sibling species of Cyprinodon, now separated by 75 miles but originally of common ancestry, have probably been isolated for about 20,000 years. These two species live in comparatively stable habitats and are not exposed to as great fluctuations and extremes of temperature as land plants, yet they differentiated specifically in only 20,000 years. He believes (1950) that in warm spring populations, high temperatures accelerate the time rate of evolution by in- creasing the mutation rate as well as the number of generations per year. Although he deals with fish, his evidence seems compatible with this study. 5. Relative Age of Genera and Relative Rates of Evolution.— Evidence from morphology, distribution, etc. all point to the relatively old age of Cleome sect. PERITOMA and the progressively greater youth of Cleomella, Wis- lizenia, and Oxystylis. The phyletic divergence of Cleomella from PERITOMA must therefore date further back than that of Wislizenia does from Cleomella, which in turn must have occurred before the divergence of Oxystylis from Wislizenia. If we assume in the evolution of each genus an equal evolutionary rate (an uncalled- for assumption!), we would expect (fig. 15A) the greatest morphological gap be- tween the genera separated for the longest time (Cleome and Cleomella) and minor morphological differences between the genera which had diverged most recently (Wislizenia and Oxystylis). he situation is, however, exactly the reverse! Inter-generic differences are greatest between Wislizenia and Oxystylis, and only very slight between Cleome and Cleomella. The relative morphological discontinuities between the genera were found to be about 2 to 3 points between Cleome and Cleomella, about 5 points between Cleomella and Wislizenia, and 7 or more points between Wislizenia and Oxystylis (fig 158). Evidently, over-all rates of evolution were not uniform, but increased with the formation of each new genus. Cleomella evolved relatively slowly, M faster, and Oxystylis most rapidly.!? This "speeding-up effect" of arid habitats on evolution is consistent with the erra of Stebbins (1947); namely, that “rapid evolution requires: (1) the presence of a rapidly changing environment" and (2) a group of plants "pre- adapted in the direction of the secular change". Both of these conditions seem to apply here. Since the four genera of Cleomoideae evolved in the arid and semi-arid regions of southwestern North America, the ancestral PERITOMA branch of Cleome must have been preadapted to xerophytism. We also know that during the past W One of the obvi ous atid. perhaps ee en end to this mg ta E is t that the gap MAE Wislizenia and Oxystylis i great because intermediate taxa have die Za nore , the older the taxon the greater the opportunity for extinction. Cleome and Cle omella, there are only the weakest of differences Bo: age youngest. Pronounced in the subfamily. [ Vor. 44 112 ANNALS OF THE MISSOURI BOTANICAL GARDEN ux 4 X ES 3 2 el o0 $ mm u > o E a ul E ae e = N Lg H = = N Hd u c o A A u- £L Oo sc S E ul EX zx hh wu o n o a d = o Ba uw z o 3 gd ; : Fig. 15. A. If evolutionary rates are the same with the origin of ea ter- : r ch new genus, in generic gaps would be expected to decrease from most primitive to specialized (from oldest to youngest) genus. e . Inter-generic gaps as they occur in the western North American Cleomoideae. ote that the putative evolutionary rate is proportional to the angle between the phyletic lines. 20 million years the climate in the Southwest has oscillated tremendously from moister, subtropical in the Lower Oligocene (Chaney and Sanborn, 1933) to the initiation and development of drier and cooler grassland and desert environments of subcontinental extent. This trend progressed, with fluctuations, from warm and mild climates in the middle Pliocene (Axelrod, 1948) to the present day, except for marked interruptions during the Pleistocene, when climatic zones shifted alter- nately south during glacial periods and moved north during the interglacials. Since the last glacial period the change toward greater aridity has been rapid in such places as Death Valley and the Great Basin (Hubbs and Miller, 1948). This ` aridity is in part a result of rain shadows due to mountain building, some of which | occurred as late as the Pleistocene and may be continuing today (cf. Deevey, 1949): Cleome sect. PERITOMA may have originated during the second major burst © angiosperm evolution (Stebbins, 1947) in the middle Tertiary in the wake of cli- matic shifts toward aridity, either from Madro-tertiary forms that evolved in the 1957] ILTIS— EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 113 Mexican Plateau and moved northward, or from mesophytic, forest-inhabiting Cleomoideae which were left behind in the southward movement of the vegeta- tional zones. Cleomella, judging from its scattered distribution in the Great Basin and east and south of the Rocky Mountains, might have had its origins also during that period. It apparently originated far east of the Sierras before their elevation in the late Pliocene, for only one of its taxa was able to reach and cross that mountain range (a putative "pre-Wislizenia", into western Sonora). “At the close of the Pliocene and the beginning of the Quaternary the present desert had its origin in the rain shadow resulting from the structural differentiation of the Sierra Madre- San Bernardino chain.” (Cain, 1943). The ancestral "pre-Wislizenia", isolated from its relatives in Cleomella by a rising Sierra Madre, evolved into Wislizenia, eventually to spread north and east around the tip of the Sierra Madre, probably during and after the Pleistocene. During the glacial periods the less specialized species of Cleome and Cleomella, judging from their present great altitudinal and habitat tolerances, very likely were able to survive in approximately the same regions they occupy today; Wislizenia was probably restricted, as it is today, farther to the south. With the post-Pleistocene changes toward extreme aridity, the opening-up of alkaline valley floors to plant colonization and the correlated great disruption of previously existing habitats, we can postulate a northward expansion of ranges of xerophytes and an evolution that filled some of the newly available niches by producing not only Oxysfylis in Death Valley, but also such exceedingly specialized species as Cleomella brevipes in the Mohave Desert, Cleome sparsifolia in Nevada, and the local diversification of Wislizenia and other species of Cleomella. "Accelerative evolution" is interpretable through organism-environment rela- tionships as postulated by Simpson (1944) and Stebbins (1949, 1950) ; in our case, through the obvious direct relationship of morphological specialization to habitat specialization and to evolutionary rate. We might say that mild environment would result in mild selection, the consequence of which is slow morphological or Physiological evolution; a change to a progressively more arid environment wo Produce greater selection, resulting in faster evolution. Accelerative evolution, then, will be exhibited by semi-desert plants, or plants preadapted to arid environ- ments, evolving into progressively better-adapted types which are able to invade (or survive in) progressively more arid regions. The result would be increased selection, mutation, etc., making invasions of still more specialized habitats possible, Which in turn would intensify the processes that have been outlined.?? Deserts and desert evolution must have occurred on the earth for a long time. If there is one definite fact we know concerning the world's geologic past, it is that mountain ranges have formed only to be worn down and to be replaced by new chains. That mountain ranges once had rain shadows with deserts as they do today is a warranted assumption, as is the former presence of subtropical ncrease in body size of the American 20 ; : Whether this applies generally I do not kn geg bith for the Megebippus ow. The steady i Equids as illustrated in Romer (1949) is essentially an accelerativ as well as the Equus line. ANNALS OF THE MISSOURI BOTANICAL GARDEN 114 paz104auob Kjubiu ei sauoz aArdopo o $101582U0 joo1doay onKudoso T ail | 9 m ~ i 19571 ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 115 high pressure belts and their resultant arid zones. The high speed of evolution in deserts may therefore explain the problems of "macro-evolution" of the paleon- tologists and the seemingly sudden origin of phyletic groups from morphological types so different that they cannot be recognized by us as ancestral. This is par- ticularly true if we assume that with reverse changes toward milder climates, rapid evolution in deserts may be followed by a reverse and slower change toward meso- phytism. In some instances the “new” types of organs developed as desert adapta- tions may have had, by change in function and by some readjustment of structures, superior survival value to other, more ancient mesophytes. That organisms may take the circular route from mesophytism to xerophytism back to mesophytism has been documented by Stebbins (1952), and certainly has occurred again and again, in the larger sense, in the evolution of major divisions of the plant and animal kingdoms (‘‘mesophytic” fish to "xerophytic" terrestrial quadruped to “mesophytic” whale, etc.). The general significance of such possibilities in the more recent evolutionary history of angiosperms! has not been fully appreciated by most evolutionists. All desert evolution was not necessarily dead end! VII. SUMMARY The four genera of western North American Cleomoideae (Capparidaceae) can be arranged in a remarkably complete phyletic reduction series, from the primitive Cleome sect. PERITOMA (6 spp.) through the progressively more specialized Cleomella (10 spp.) and Wislizenia (1 sp. with 3 vars.) to the highly specialized Oxystylis (1 sp.). The reductions run from many-seeded siliques through few- seeded silicules to two-seeded schizocarps; from open, bracteate racemes through enser, ebracteate ones to congested, ebracteate raceme-clusters; from small to very complex stipules; and from plants with much vegetative growth and delayed Owering to ones in which even the earliest leaves subtend flowers or racemes. Nearly all these reductions are interpretable as adaptive modifications, a result of selection by arid environments which favored plants with more rapid or economical life cycles, The geographic distributions also imply that these reductions and specializa- tions are due to adaptive evolution. The primitive forms are widely distributed in more mesic climates; the specialized have progressively smaller, essentially equi- formal areas centering about the Death Valley region of California and Nevada, where Oxystylis is endemic. The ranges and degrees of relationship of individual taxa within each genus follow patterns corresponding to their morphology. The taxa of the primitive (therefore oldest) genus Cleome are all more or less sympatric ———— ei any of the present-d £ pee xe our maj ogen e, may y aig ib rm > Perhaps be explainable as an adaptation to xerophytic life. If t true, the lack of Zä gengt might be ue to extremely fieri eologically “instantaneous” ) desert evolution, with consequen short span j ch ead fossilization might have taken place, as well as the re = which widesp tunities for bester Macheren of desert habitats (in contrast to mesic or hydric egen [ Vor. 44 116 ANNALS OF THE MISSOURI BOTANICAL GARDEN and sharply distinct, those of the more specialized genus Cleomella both allopatric and sympatric and less sharply distinct, and those of the monotypic, polymorphic Wislizenia allopatric with closely related populations. The most specialized, and therefore youngest genus Oxystylis is monotypic and monomorphic. The fact that there are different kinds of genera can in part be explained by differences in their age. In evolution through geographic isolation, taxa of a genus have to be allo- patric before they can become sympatric, and by studying their geographic ranges and the amount of sympatry or allopatry, we can correlate geographical distribution with morphology and variability, and determine the relative age of a genus. By extending "Jordan law” not only to apply to the most closely related taxa within a genus, but also to a pair of monotypic genera or to the two most closely related species of two related genera, we can determine (if evolution has been geographic and sufficiently recent) the geographic direction of evolution by decid- ing which of the two taxa is the most primitive. Thus, Cleome sect. PERITOMA originated from primitive tropical species allopatric to the south. The primitive species of Cleomella seem to have originated from a primitive Cleome allopatric to the northwest. The primitive variety of Wislizenia is allopatric to the west of, and separated by the Sierra Madre Occidental from its putative ancestor in Cleomella. Finally, the most specialized variety of Wislizenia and its offspring Oxystylis have exactly allopatric ranges, the dividing line being the Death Valley basin-Colorado River basin watershe Since the Death Valley basin contained a series of tremendous post-Pleistocene lakes about 15,000 years ago, it seems Yr unlikely that Oxystylis migrated to Death Valley except in the form of a “Wislizenia-like” plant after a desert or semi-desert climate had become established; it is postulated therefore that Oxystylis evolved (1) in the Death Valley area (2) within the extremely short time of 15,000 years or less. That generic evolution at such a tremendous speed is possible under extreme selection is amply illustrated by domesticated organisms, plants an animals. Furthermore, there is a great similarity between the habitat and evolu- tionary rate of cultivated annuals and desert annuals. In Oxystylis, the rapid changes towards greater aridity resulted in extreme selection which, together with other factors characteristic of deserts, is postulated to have been responsible for this rapid evolution. To the four factors (Stebbins, 1952) which help make “aridity a stimulus to plant dioc a fifth, the effect of extreme temperatures on the mutation rate, may The increasing inter-generic discontinuities between the four genera, repte- sented by the arbitrary numbers 3—5—7, point to an accelerative increase in speed of evolution with the development of each new genus. The “accelerative evolution" may be explained by an extension of Simpson’s and Stebbins’ postulates — that in preadapted organisms, the more stringent the environment, the faster evolu- — tion, which in turn would produce better-adapted organisms able to survive in; % ` invade, still more stringent environments with still greater selection. 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 117 The very rapid evolution in deserts may have been of major evolutionary importance in creating radically new morphological types. With a return to mesophytic climates, some of these types may have been highly successful. Per- haps some precursors of the angiosperms, whose ancestors are as yet unknown, had their origin in desert evolution. BIBLIOGRAPHY Abrams, Le Roy (1944). Illustrated gee of = Pacific States. Vol. 2. Stanford, Cal. Anderson, Edgar, p og bbins, G. (1954). Hybridization as an eiiis a mechanism. E Anonymous eee d Valley National Monument. 1lépp. Washington, D. C. Antevs, GE (1937). Age of the Lake Mojave culture. Southwest Museum Papers (Los Angeles) Axelrod, Dania I. (1948). Clima and evolution in western North America during middle Pliocene time. Evolution 2:127- Babcock, Sie B. (1947). Cé Genus Crepis. Univ. Cal. Publ. est e he Bell, C. Ritchie (1954). The Sanicula crassicaulis complex. Ibid. 2 icy owt n and Ellsworth, E. W. (1934). SE lakes Ps Ka eae Basin, Calif. Am. Jou 1:453-463. Wiis, ‘Friedrich (1936). Resedaceae, in Engler and Prantl, Die Natürlichen Pflanzenfam. (ed. 2) F. O. (1923- 1928). The Fer 3 vols. Cambridge, England. been. John ( 1914). Sur Cette et les affinités des Capparidacées à fruits vesiculeux. Engl. Bot. Jah A 50 (Suppl. :435—448. S x Buxbaum, Franz (1951). Smetne und Methoden einer Erneuerung der Systematik der höheren Pflanzen. 22 Cain, piedi A (1 94 43). "The tertiary ipd of Gm reds KC forests of the donat Smoky sins National Park. Bull. Tor t. Club ; Rm Foundations a Plane eco, zn T$ in Camp, W. H. (1950 ). Plant ier In MORE. A edes of Dicotyledons p. 26. Chronica Botanica. Walt : ; : Campbell, D. and Wiggins, T * (942). Origins of the flora of California. Stanford Univ. Publ. Ser Bil Sci. Campbell, E. W Son ee Ae ll, A H. (1937). The Lake Mojave site in the archeology of ci Lake "Mojave geg hwest Museum Papers (Los Angeles) ys 9-44. Chaney, Ralph E. (1947). Ter rtiary desde s and migration routes. l. Monogr. 17:139- — — ——, and Sanborn, Ethel I. (1933). The Goshen Flora of Late central Oregon. eem Inst. Wash. Pu : E Coville, F. V. (1893). Botany of the Death Valley posi M Contr. U. S. Nat. Herb. 4:1 Crum, Ethel K. ana. Geographic ^ stribution and r nships of Isomeris; radi and Whe lizenia. 172 pp. Univ. Cal. (Berkeley), Master’s SES Toopu kiuka]. Dec, Edward S. Jr. (1949). P paatti of the Pleistocene. Bull. Geol. Soc. Amer. 60:1315- Dobz ec nsky, Th. (1941). Genetics and the pp of Species. 446 pp. Rev. ed. New York. Dressler, R. L. (1954). The Genus Tetracoccus. Rhodora 56:45-61. Tex : Fankhaus m Kä (1942). Induction of Ve i in animals by extreme temperature. Biol. Symposia :21— ED AS. 2 al. (1895). Flora of North America 11:184-185 eene, E. L. (19 Á 06). i sies of the genus Wislizenia. Proc Biol. Soc. Wash. 19:127-132. bes Hen nnig, Luise (1929). Beitrige zur Kentniss der ee e und Frucht. perg E :507— Hubbs, Carl L., Tes aen R. R. (1948). Correlation between fish dicite and ~~ aphic hise tory in the desert basins of western United States, a ^h reat Basin, with em iru x agre EI Postglacial Times. Bull. Univ. Utah 38 [No. 20 (Biol. Ser. 10, No. 7)]: Huxley, äm (1943). Evolution: the Modern Syrithesis. 645 pp. New York. [Vor. 44 118 ANNALS OF THE MISSOURI BOTANICAL GARDEN Iltis, quud H. (1950). Morphology Se I wis Geesen? in the Cleomoideae of the New World. p. Master's thesis, Washingt v. St. Louis, Mo. [unpublished » (1952). A revision of t Pe w World scis däer: 335 pp. Ph.D. thesis, Wash- ington nen St. Louis, Mo. [un ublished]. a) can in ds western North American Cleomoideae (abstract). Proc. Ark. :118. , (1955b). Capparidaceae of eee a Archer, W. A. Contributicns to a Flora of Nevada. U. S. Dept. Agr. Bur. Pl. Ind. 3 ; (1956). end in the Capparidaceae m The Mexican species of Cleomella: Taxonomy and ev d j Jaeger, J. C. (1941). Dent wild See 322 pp. inti ed. Stanford, Cal. Jepson, W. L. (1936). A Flora of California. 2:9—16. Berkeley, Cal. Johnston, I. M. (1924). Expedition of the Cité ov me of Sciences to the Gulf of ne x JN. The Botany (The vascular plants). Proc. Cal. Acad. Sci., IV, 12:951—1218 (see p 28). Steeg D. NS and Korovine, E. P. (1942). La vie dans les deserts (French ed. of Th. Monod). 360 Kuntze, Otto (1903). In Post, T. and Kuntze, O. Lexicon uereg Phanerogamarum. p. Lam, H. J. a pede Paleontology, Compt. Rend. 3 Congrès Strat. et Geol. a po nifüre. CN 38 Li; Leg (i332). . Soc. 42 MacDougal D. R. (19 Gë Botanical features of North American deserts. Carn. Inst. Wash. Pub. Hors relationships between eastern Asia and eastern North America. Trans. 71—429. Mange ido Paul C. (1952). sees psu domestication. age Nat. 86: EE Mayr, Ernst (1942). Systematics the in of Species. SH äus (1945). The hes anula. Mua 1:1 e 2. Sitios R R. (1946). Correlation between fish distribution and Pleistocene hydrography in eastern Cali MPs and Seengen Nevada, with a map of the Pleistocene waters. Jour. Geol. 54:43- 534. 2 New us ie ee Rafinesque, d is Meer to Specularia and ; (1948 The Cyprinodont fishes of the Death red —— of eastern California and southwestern Nevada. : Misc. Publ. Mus. Zool. Univ. Mich. 6 —— 0 - Speciation in ihe of. the genera egene? d Em petrichthys, inhabiting the Deat th i Gao dA region. Evolution 4:155—16 Muller, C. H. (1928). "The e measurement of gene utation rate in Drosophila, its high variability, ependence upon ges isch Genet $15: 279—357 94. lation mechanisms, evoluti $a and aptur Biol. Symposia 6:71—125. Mun: A. (193 5). A Manual of Southe ern California Botan d "a 2 pp. apes Cal. Parish, n B. (1903). A sketch of the flora of southern Califor Bot. . 36:20 3-222 , 259—279. uec (a SC Vegetation of the Mojave and Colorado deus of idum California. Ecology 81—49 Pax, F, gé Hoffman, K. (1936). Capparidaceae. In Engler and Prantl, Die Natürlichen Pflanzen- fam. (ed. 2.) 17b:146—2235. Payson, E. B. (1922). A synoptical revision of the genus Cleomella. Univ. Wyoming Publ. Sci. 1:29—46. 3 Pellet, pary (1930). American Honey Plants. Gogo? Bee Jour. H on, Ill. Plough, H. H. (1942). Temperature and spontan mutation. Biol. Sy SN gs. 1-2- Se Romer, Alfred S. (1949 p series and iecit i in animal sivo she e in genetics, pilcontilogy and evolution. pp. 103— we Segen J. ibo the Anand of die Eocene Wilcox Flora to some modern floras. Evolu- Shkvarnikev, P K., and Navashin, M. S. Rice Über in ruhenden Samen unter dem Einflus Chaos des Mutationsvorganges e 0-736. Simpson, G. G. (1944). Tempo ac? ). die n Temperaturerhöhung. d 22:72 eg Gs Mode in Evolution. 237 pp. New York. 6). Jain sinia of evolution in various ra types. Amer. Nat. Stebbins, G. G. Ledyard, Jr. (1942). The genetic approach to problems of rare and ic species E ono 6:24 endemi 947). Evidence on an - evolution from the distributions of existing and fossil plant species. Ecol. Monogr. 17:149-1 , (1949). Rates of evolution in qi In Genetics, Paleontology and Evolution. PP. 229—242. Princeton, N. L » (1950). Mariette and Evolution in Plants. 643 pp. New York. Aridity as a stimulus to plant evolution. Am. Nat. 86:33—44. , (1952). 1957] ILTIS—EVOLUTION AND PHYLOGENY OF CLEOMOIDEAE 119 Thompson, David G. (1929). The Mojave Desert region. U. S. Dept. Int., Geol. Survey, Water Supply Paper 578:1—759 d Timofeeff-Ressovsky, N. W. (1940). Mutations and geographical variation. In J. Huxley ed., The New NETET 73-136. Tryon, R. M. (1955). Sellaginella rupestris and its allies. Ann. Mo. Bot. d 42:1—99. Went, F. W. (1955). The ecology of desert plants. Scien ei Gen 192:68—7 Willis, J. C. (1922). Age and Area. 259 pp. Cambridge Woodson, R. E., Jr. (1935). Observations on the ries ina of the Apocynaceae. Ann. Mo. Bot. ard. 22: 7 Zohary, M. (1948). Carpological studies in Cruciferae. Palestine Jour. Bot. (Jerusalem Ser.) 4:158—165. STYLITES, A NEW GENUS OF ISOETACEAE ERIKA AMSTUTZ! Stylites andicola E. Amstutz, gen. et spec. nov. Isoetacearum. Plantae amphibiae vel subamphibiae in societatibus dense caespitosis supra margines paludum alpinorum frequentes. Caudices maximi crassi apice profunde bilobi usque 7 cm. longi 1.2 cm. lati vel ultra; cortice crassiusculo fusco cicatricibus foliorum vetustorum notato; radicibus per longitudinem uno latere caudicis valde secundis usque fere 5 mm. crassis speciminum nostrorum simplicibus fortasse tan- dem dichotomis. Folia in rosulis densissime imbricata oblongo-lanceolata apice anguste caudato-acuminata ibique valde reflexa ca. 5.0—5.5 cm. longa 5-8 mm lata crassa margine partis basalis aut vaginae late hyalino medio et parte apicale grosse bullato. Sporangia usque 1.5 cm. supra basem foliorum instructa oblonga ca. 5 mm. longa 2 mm. lata e foveola tenue margine indistincto proventa; tra- beculis paucis paxilliformibus; velo nullo; macrosporis albidis laevibus magnitudine aliquantulum dissimilibus ca. 428—643 y latis; microsporis non visis. Ligula anguste cordato-triangularis 3—5 mm. longa. — Nomen e ørvàírys habitore in columna derivatur caudice notabili designans. Peru: Depto. Lima, Prov. Huarochiri, above Casapalca near Caprichosa, in alpine bogs, alt. 4750 m., 2 July, 1956, E. Amstutz 2000 (Herb. Missouri Bot. Gard., HOLOTYPE). These curious plants were found on a moist and limy substrate around the boggy margins of a small glacial lake near Caprichosa, a few kilometers from Casapalca, central Peru, at an elevation of 4750 m. The plants live in extremely crowded colonies forming flat and dense cushions in a belt 2-3 m. broad and about 3—5 m. from the waters of the lake. The largest colonies consist of 40 or more individual plants. Better developed specimens were found in the peripheral parts of the larger cushions (50—60 cm. in diameter). The plants apparently are well adapted to their environment. The leaves and upper portions of the caudices are aerial, but so crowded that considerable pro- tection from the severe environment is effected. Associates of Stylites andicola are sods of peat and swamp bryophyta, convex and very hard cushions of Disticha muscoides (Juncaceae), tufts of Calamagrostis chrysantha (Gramineae), etc. These plants display several innovations which deviate significantly from the Very conservative architecture of all known species of Isoetes. The elongate cau- dices are outstanding; furthermore, in digging my specimens in the frozen soil, I unfortunately failed to obtain unbroken plants, and it seems possible that the —— ! Graduate Student, Henry Shaw School of Botany of Washington University, St. Louis, Mo. Issued March 22, 1957. (121) [Vor. 44 122 ANNALS OF THE MISSOURI BOTANICAL GARDEN dense cushions may segi be formed through branching of the caudices of single plants, at least in part. e fleshy roots are borne along only one side of each caudex. Here, also, my excavation was not sufficiently careful, for all of the roots were cut during the process. However, in all attached roots which remain, some more than a decimeter in length, no branching may be seen. The short broad leaves with thick and recurved tips, to which effective photosynthesis appears to be re- stricted, are most exceptional; but most remarkable of all is the attachment of the sporangia far above the base of the sporophylls. The fovea is extremely shallow and the sporangium is less than half included within it. The trabeculae are few and peg-like. Unfortunately, microsporangia were not observed. 1957] AMSTUTZ—A NEW GENUS OF ISOETACEAE Figure 1. Sfylites andicola (yA fy A ||| Au MI f d ER I I | 123 MISSOURI BOTANICAL GARDEN STAFF Acting Director HucH C. CUTLER EDGAR ANDERSON Grorce B. Van SCHAACK, Curator of Useful Plants Acting Curator of Herbarium CARROLL W. Dopce, < JULIAN A. STEYERMARK, Mycologist Honorary Research Associate Ro S nois E Vc > Ja. FREDERICK G. MEYER, — Dendrologist Henry N. ANDREWS, Paieobotanist ALICE F. TRYON, "Research Associate Rora M. Try ON; E Assistant eé of the Joun D. DwYER, erbarium Research Associate Taror VON SCHRENK Se o ia Assi trator “Museum of Net C. HORNER, A ; nomie Plant ; Bibliographer and Editor Oscan E. GLAESSNER, Controller : BOARD OF TRUSTEES President Jonn S. LEHMANN Maii eii Dante K. CATLIN Sein Vice-President LEICESTER B. FAUST. a kauni a | Dubprzy FRENCH : ape B. riesen Best Hrrcncocx A. Wessex, map cie Bacorrwes Ne Es 2i "EX-OFFICIO MEMBERS — aar uh -ARTHUR C. LiCHYENBERGER, - RM | Srratroap Les Monee, ` Palo of the Diocese of ees ETE DS THAN A H SE Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the Missouri Botanical Garden and the Henry Shaw School of Botany of Washington University in affiliation with the Missouri Botancal Garden. Informaton Ls OF THE MISSOURI BOTANICAL GARDEN appears four times POR the SECH year: February, May, September, and November. Four numbers constitu olume. te a vi Subscription Price _. $10. 00 v volume Single Numbers «—— — — 2.50 ei Contents of previous issues of the ANNALS OF THE MissouRi BOTANICAL GARDEN are listed in the Agricultural Index, published by the H. W. Wilson . Company. ATTENTION WE REGRET TO ANNOUNCE THAT INCREASED COSTS MAKE IT NECES- SARY TO ADVANCE THE PRICE ÓF THE ANNALS TO $12.00 PER YEAR, BEGINNING WITH Vor. 45, rmi Milone BOTANICAL GARDEN Annals of the Missouri Botanical Garden Vol. 44 MAY, 1957 No. 2 A REVISION OF THE FERN GENUS PELLAEA SECTION PELLAEA* ALICE F. TRYON The species included in this section of Pellaea are predominately Cordilleran 8rowing in dry, rocky habitats under semi-desert conditions quite unlike those usually associated with ferns. These and most other ferns of arid regions in the New World belong to the tribe Cheilantheae of the Polypodiaceae. The revision of this section was undertaken as part of a program of study on this tribe. In the "Genera Filicum’, Copeland recognizes the genus Pellaea as a natural one although his remarks on the affinities of the group do not entirely support that view. A more pessimistic. regard toward Pellaea as a genus is evident in the view of Thomas Moore some fifty years earlier for he considered it as a medley of pteroids, cheilan- thoides and platylomas and stated—‘Pellaca can only be kept up as a kind of botanical refuge for the destitute.” The section PELLAEA is a restricted but natural group based on vegetative morphology such as the nearly glabrous, imparipinnate blade and free venation, and on the continuous marginal indusium and terminal sori on discrete vein ends. Relationships with Doryopteris, Notholaena and Cheilanthes are not wholly clear and the final disposition of the group in Pellaea and with other related genera is best delayed until the entire complex can be reviewed. The entities included in section PELLAEA under the above provisions number 15 species and 5 varieties. Within this small group there appears to be two types of speciation. Apomixis, polyploidy and hybridization have influenced speciation in P. glabella, P. sagittata and P. ovata. The series related to P. ternifolia seems to have evolved through gradual quantitative differentiation and through isolation of resulting populations. — Sciences. ashington University and of Doctor of Philosophy. (125) n investigation carried out in the Henry Shaw [Vor. 44 126 ANNALS OF THE MISSOURI BOTANICAL GARDEN GEOGRAPHY AND EcorLocv The species prevail in the southwestern United States and Mexico. P. rufa in South Africa is the only species belonging to this section which occurs in the Old World. Three species are confined to the United States; one is endemic in Chile and two in Mexico. Ten of the species or varieties occur in the southwestern United States and Mexico and three of these, P. ternifolia, P. ovata and P. sagittata, are distributed along the Cordillera to Argentina. The Cordilleran distribution pattern is similar to other ferns as Notholaena sinuata, the N. nivea-incana group, the Cheilanthes pyramidalis-marginata group and to the Umbellifers Osmorhiza mexicana and Sanicula liberta. 'The most extensive distribution among the pellaeas is that of P. ternifolia in the western highlands of the Americas from Texas to Argentina, in the Sierra de la Ventana in eastern Argentina, and it is reported by . H. Hicken! from Minas Geraes although I have seen no specimens from Brazil. Both P. ternifolia and P. ovata occur in Hispaniola but the former is unique in the group in its occurrence in the Hawaiian Islands as well. The Cordilleran-Hawaiian distribution is also common to Asplenium monanthes. The genus occurs in Africa but within this section only the Chilian P. myrtillifolia and Californian P. anat medaefolia have a close relationship with P. rufa of the south African Karroo. Floristic relationships between California, Chile and South Africa occur in other genera as Acaena in Rosaceae and Menodora in Oleaceae. The southwestern United States and adjacent Mexico is the most probable center of dispersal of these species, for the ranges of all but P. Breweri and P. glabella either converge in this region or seem to be derivatives of those that do. In two species complexes it is possible to detect the derived types by the number of spores per sporangium and in these the ranges of the ancestral members converge in this same region. à In the eastern United States the most common species is P. atropurpurea which ranges from Guatemala to Florida, Vermont and Colorado with two isolated stations in British Columbia and in the vicinity of Lake Athabaska in Saskatchewan. : species is reported to be a triploid and apogamous by both Manton? and Britton. Pellaea atropurpurea, P. glabella, P. ovata and P. sagittata var. cordata are calciphilous while P. ternifolia and related species are reported mainly from igneous rock. Both P. /ernifolia var. ternifolia and P. sagittata var. cordata occur on the lava beds in the vicinity of Mexico City. The former grows directly on the lava while the latter occurs in crevices and pockets, where soil has accumulated, along with other calciphiles as Opbioglossum Engelmannii and Selaginella Wrigbtii. P. ternifolia appears to tolerate a variety of habitats and a range in altitude of 200- 4000 meters. Wiggins reports it and also P. sagittata and P. ovata on the dry, 1 Soc. Argent. de Cien. Nat. 13:206. 191 6. ? Ann. Mo. Bot. Gard. 42:101. 1955, * Manton, I. Prob. Cytol. & Evol. of the Pterid. Cambridge Univ. Press. 1950. * Britton, D. M., University of Maryland, Dept. of Horticulture, personal correspondence. ? Amer. Fern Journ. 36:1, 1946. 1957] TRYON-—PELLAEA SECTION PELLAEA 127 earthen walls of Ecuadorian towns. P. ovata is also an epiphyte on trees and generally has a subscandent habit with fronds nearly 114 m. long. MORPHOLOGY These species are not fern-like in aspect and their leaf design is undoubtedly influenced by the arid conditions in which they grow. The spores and gametophyte have remarkable tolerance to desiccation. Herbarium specimens two to five years old bore spores which were germinated in ten days. Prothallia of P. longimucro- nata were kept air-dry for two months at temperatures ranging from 80 to 100° F. and became green when watering was resumed although no sporophytes were produced. Pickett and Manuel" carried on some experiments on désiccation of prothallia and sporelings of P. atropurpurea and P. glabella in which some plants were air-dry for nine and some for 1817 months. When watering was resumed some of the plants became green and continued growth. After five years Pickett’ examined the same cultures and 5 per cent showed whole or partial recovery when satisfactory growing conditions were restored. The rhizomes grow deep in crevices or under rocks and contain an oily sub- stance which is still present in specimens collected some forty years ago. The rhizome has two general forms—one relatively massive, up to 1 cm. in diameter, compact or somewhat elongate, multicipital and ascendent or decumbent—the other slender, cord-like, less than 14 cm. in diameter, creeping and with buds on the older portions. The vascular system of the rhizome is not a solenostele in the strict sense but approaches a dictyostele. There may be more than one foliar gap in a section of the slender and long-creeping stems and there may be several in compact stems with short internodes. The scales of the rhizome are mostly elongate-lanceolate or elongate-triangular with a more or less filiform and tortuous apex and entire or dentate margins. Young scales are usually concolorous. Mature scales may be concolorous or bicolor- ous with a central dark, sclerotic stripe or basal portion. In a few species scales of the older portions of the rhizome are wholly sclerotic, brown and lustrous except for a narrow, clear border. Cells of the scale base are usually small, squarish or irregular and above they are usually 20 or more times longer than broad. Bands of constricted cells are frequent in P. glabella and seem to occur where the scales are appressed. Scales of the stipe base are generally similar to those of the rhizome although in P. sagittata the stipe and bud scales are exceptionally broad and Pseudopeltate. : The plants are generally stiff and erect in habit and the fronds are coriaceous in texture. The smallest mature plants have pinnate fronds 1 cm. long and the largest are quadripinnate, up to 125 cm. long, and may be subscandent. ves are usually monomorphic although several species have the segments of the fertile fronds somewhat contracted. This latter condition is especially evident in — ; Bull, Torr. Bot. Club 53:1-5. 1926. Amer, Fern Journ. 21:49—56. 1931. [Vor. 44 128 ANNALS OF THE MISSOURI BOTANICAL GARDEN P. atropurpurea. ‘The blade is usually elongate-triangular but is linear in P. Breweri and P. ternifolia. The blade of P. ovata is most elaborate with many divisions and strongly flexuous axes. In the series of species from P. ternifolia to P. mucronata a relationship exists in the pinna between the relative position of the terminal seg- ment and the lateral segments adjacent to the rachis. The pinna has a simple, ternate form in P. ternifolia, is longer and has more segments in P. longimucronate | and still more complex in P. mucronata. In each of these species, irrespective of the development of the pinnae, the lateral segments nearest the rachis are inclined toward the terminal segment. P. mucronata having the most complex pinna-form also occupies a peripheral position in the geographic range of the series and it seems possible that this is a derived form in which the blade has been elaborated. The form and indument of the bud and the type of vernation are distinctive in some of the species. P. atropurpurea has lanate buds which are elongate and become uncoiled soon after they arise from the rhizome. The buds of P. glabella var. glabella are round, tightly coiled, sparsely pubescent and slowly uncoil. The stipe may be plane or sulcate on the upper surface or it may be terete. Stipe color is castaneous to black or straw colored to ruddy brown. A few species have stipes essentially light colored with dark patches. In transverse section the stipe has 1 single vascular bundle having the form of a flattened or sharp angled V with in- rolled ends. In P. Breweri and P. glabella the fronds articulate at the stipe base. The blade is imparipinnate but not entirely so in P. Breweri and some specimens of P. glabella. The manner in which the pinnae depart from the rachis is a useful character in distinguishing P. atropurpurea and P. glabella. In the former the pinna-rachises depart abruptly from the rachis and form broad angles while in the latter they are somewhat decurrent or adnate to the rachis below the axis and form smaller, acute angles with the rachis. The length of the pinna-rachises ranges mostly between 5 and 10 cm. but this may be up to 15 cm. in P. ovata and P. andromedae- folia or in some species it is lacking. i Venation is pinnate, free, rarely anastomosing and is usually obscure being sunk in the mesophyll. The sunken stomates give a punctate texture to the lower surface of the segments. The receptacle occupies the vein end usually less than V& of the distance to the midvein and is clavate or cuneate in form. If the sporangia are abundant the sori are superficially coalescent but the vein ends are free or rarely two are fused. The indusium is formed by the reflexed margin of the segment which may be only slightly or strongly inrolled. The tissue which functions as the indusium is usually membranaceous but may be essentially unmodified. ; The sporangia stalks are as long or longer than the capsules except in P. Pringlei and P. brachyptera where they are short. The annulus is vertical, amber colored or rarely yellow and does not extend to the insertion of the stalk. In some apo£- amous species the annulus may be oblique or rarely a cluster of cells at the capsule apex. 19571 TRYON—PELLAEA SECTION PELLAEA 129 Simple capitate trichomes occur in P. brachyptera and P. mucronata and occa- sionally in P. longimucronata and P. ternifolia var. Wrightiana. They are found among the sporangia and secrete a whitish yellow substance. In an old paper the chemical nature of an apparently similar substance in Pityrogramma triangularis is noted by Blasdale® in relation to the material found in Notholaena, Cheilanthes and Primula. In Pityrogramma triangularis the substance is called Ceroptene and consists of yellow, triclinic crystals and whitish amorphous bodies melting at 135° C. and soluble in both alcohol and xylol. Spore color in the species with dark stipes is tan to brownish and the rugae are scarcely prominent with the exception of P. atropurpurea. The latter, along with the light stiped species, has prominently ornamented spores which are generally light yellow. The spores are of special interest in this group for those of apogamous strains are larger, usually monolete with 32 per sporangium in contrast to the smaller, usually trilete 64 spored non-apogamous strains. The spores remain viable for a few years; thus it is possible to germinate them and check the gametophyte state of recent herbarium specimens. A positive indication of apogamy is shown by prothallia on which sporophytes are formed subsequent to the production of antheridia although no archegonia are formed. CyTOLOGY The chromosome counts in this work may be regarded as preliminary until additional material is available to support them. The cytology, however, has been helpful in indicating some of the mechanisms by which these ferns reproduce and has facilitated the taxonomic disposition of several entities. In a survey of the chromosome numbers of ferns of eastern North America, Dr. Donald M. Britton® of the University of Maryland treated P. glabella var. glabella and he kindly agreed to do additional counts on related species of Pellaea. He has studied ten members of six species from aceto-carmine squashes of the spore mother cells and these are reported in the treatment of the species. The chromosome base number of n = 29 as been established and both triploids and tetraploids have been formed on this base. From their morphology the triploids P. atropurpurea and P. sagittata var. sagittata appear to be allopolyploids. P. ternifolia has both diploid and polyploid Strains and it is possible that the latter may be a parent in the triploid P. atropurpurea. PHYLOGENY Since polyploidy is prevalent in the group it is dificult to plot relationships until chromosome counts are reported for all species. There is likewise diffculty .in the establishment of primitive and advanced characters. A sequence of genera on the receptacular condition within the tribe Cheilantheae is discussed by R. M. Tryon!? in his treatment of Doryopteris. He regards this genus as most advanced and derived from either a Cheilanthoid or Pellaeoid stock and Pellaea as ——— *Erythea 1:252. 1892 9 fik Amer. Journ. Bot. 40:575—583. 1953. Contrib. Gray Herb. No. 143. 1942. [Vor. 44 130 ANNALS OF THE MISSOURI BOTANICAL GARDEN parallel to or derived from Cheilanthes or Notholaena and perhaps intermediate to these and Doryopteris. Within the section PELLAEA several natural lines are apparent. Two of these are distinguished on the basis of the color of the stipes. Dark colored stipes are considered advanced since some of the light stiped species have less specialized, creeping rhizomes. Within the dark stiped group there are two distinct lines. One line composed of P. Breweri and P. glabella has evidently emerged from a Cheilanthoid stock as shown in the decurrent pinnae and the pinnate-pinnatifid division of the blade. The other line of dark stiped species is the most closely related group of the section. It appears to have evolved from P. fernifolia and is of such continuity that some difficulty is encountered in delimiting the species. These seem to have evolved through morphological and ecological adaptations such as those discussed by Mason"! in his paper on the role of plant geography in tax- onomy in reference to the tribe Helenieae in the southwestern United States. The lack of distinctive characters in these Compositae is interpreted as an indication of relatively recent elaboration. Among the species of Pellaea with light stipes those with cord-like rhizomes are considered more primitive and of these P. ovata is most specialized by its elaborate frond. Of the light stiped species having stout, multi- cipital rhizomes P. sagittata is regarded more specialized by its dimorphic scales. P. notabilis seems to be intermediate between the two main groups for although the stipes are essentially light colored their bases are dark. It is most closely related to P. atropurpurea on several characters. 7 BRACHYPTERA 6B VAR. CALIFORNICA A 6A VAR. MUCRONATA 15 MYRTILLIFOLI 6 MUCRONATA 14 RUFA CORD RHIZOME 5 LONGIMUCRONATA CONCOLOROUS SCALES 4B VAR. WRIGHTIANA 10 PRINGLE! DROMEDAEFOLIA 2C VAR. SIMPLEX N&2N=116 4A VAR. TERNIFOLIA o 13 ANDROMEDA 2B VAR. GLABELLA N&2N=116 4 TERNIFOLIA , iun em 2A VAR. OCCIDENTALIS N=29 12 2 GLABELLA 3 ATROPURPUREA : TA N&2N*87 € 11 OVATA Kei 1 BREWER! N=29 "n 8 Kaze SCALES CONCOLOROUS SCALES BICOLOROUS STOUT RHIZ CONCOLOROUS SCALES x l DARK STIPE LIGHT STIPE ~ Phyletic chart of Pellaea section PELLAEA with known chromosome numbers. 11 Chronica Botanica 14:154-159. 1953. 1957] $ TRYON—PELLAEA SECTION PELLAEA 131 The phyletic chart shows the relationships discussed above and the chromosome counts that have been reported. BACKGROUND OF THE STUDY From Linnaeus to Desvaux's ‘Prodrome’ of 1827, six of the entities included here were placed in Pferis, one of the major fern genera of earlier botanical works. As interest in ferns increased the botanical gardens of Europe developed their living collections which included many from the American continents and fern classifica- tion advanced. This group was recognized as a natural genus twice within the year 1841 and thus the nomenclature has been confused from the beginning. Link!? proposed Pellaea from the collections at the Berlin Gardens and John Smith!? ap- plied the name Platyloma for the same group from the collections at Kew Gardens. Smith's name, published in August, appears to have preceded that of Link. Alston has noted in the British Museum's copy of Link's work that it was issued between September 3 and 10, fide—Aag. Bibl. f. Deutsch. Lib. No. 37 p. 421. 1841. A committee on nomenclature for Pteridophytes!* reported a unanimous vote to conserve Pellaea and Dr. Pichi-Sermolli,!? Secretary of the Committee, has reviewed the matter. The International Code adopted by the 8th Congress at Paris, 1954, includes Pellaea among the genera to be conserved. Presi? took up a still earlier name, Allosorus of Bernhardi and treated these species in section MONOMORPHI along with species now referred to Jamesonia, Doryopteris and Cheilanthes. Otto Kuntze!” follows Presl in order to avoid con- flict with Pellis of Raddi and Pella of Gaertner and since he believed Presl's work more adequate than Link's. Fée!? distributed these species among four of six groups, based largely on the shape of the blade, he includes in Pellaea. W. J. Hooker!? also considered vegetative characters as diagnostic for he ar- ranged these species in Pellaea according to the division of the blade. His diagnosis of the species, illustrations and general information on the collections he examined are better than previous accounts. ; Brand proposed the section EUPELLAEA for these species as one of eight sec- tions and indicated that these are the Pellaea of Link. He considered the condition of the blade, the sori, the vascular system and scales. Diels?! essentially followed Prantl's scheme. 4 Fil. Sp. Hort. Bot. Berol. Cult. 59. 1841. m Journ. Bot. ed. W. J. Hooker 4:160. 1841. Taxon 3:70. 1954. :428. 1953. Leg Pterid. 151. 1836. ev. I Vor. 44 132 ANNALS OF THE MISSOURI BOTANICAL GARDEN Until studies of the African pellaeas and other genera of the Cheilantheae are completed the disposition of this group as a section of Pellaea seems the most practicable course. UsE AND CULTIVATION These neat, novel plants were popular in English stoves and American con- servatories during the Victorian era. Two of the plants were introduced for greenhouses by Messrs. Veitch & Sons of Chelsea who obtained them from California. They were described as species and the types at Kew are nursery specimens more robust than the native Californian plants. P. brachyptera, P. ovata, P. andromedae- folia and P. mucronata were included in the florist trade of the period and the last is described as a handsome house plant producing as many as 75 fronds. Many of the species are now grown in botanical collections for they are easily cultivated. W. J. Hooker noted that several of the pellaeas were cultivated in the temperate fern houses at Kew more than a hundred years ago and at present the hanging baskets of P. ovata and P. sagittata in these fern houses are most attractive. Sev- eral of the Mexican pellaeas are grown in the University of California greenhouses and P. ternifolia var. ternifolia was volunteering there in pots of cacti and succu- lents. Seven species have been grown in the experimental greenhouse at the Mis- souri Botanical Garden. The plants require little special care other than a rock substrate and good drainage. P. atropurpurea, an attractive rock garden plant, is evergreen and does especially well on native limestone. : There are a few references on the medicinal uses of Pellaea. P. mucronata is reported to make a palatable beverage recommended for the kidneys and as a bl purifier. The vernacular name Tea Fern and the Tejon Indian name Calagula have been applied. Chilian miners use the name Coca or Yerba Coca for the local P myrtillifolia and prepare an infusion of it taken to allay pain. ACKNOWLEDGEMENTS l am indebted to the curators of the herbaria of the following institutions for the generous loan of specimens: Chicago Natural History Museum, Gray Herbarium, Muséum National d'Histoire Naturelle, Paris, Naturhistoriska Riks- museum, Stockholm, Naturhistoriska Riksmuseum Paleobotanical Department, Stockholm, New York Botanical Garden, United States National Museum, Uni- versity of British Columbia, University of California, Cornell University, Univer- sity of Kansas, University of Michigan, University of Minnesota, University of Utah, State College of Washington, Pullman, University of Washington, Seattle, University of Wyoming and Yale University. The abbreviations for these an other herbaria are taken from ‘Index Herbariorum, part I (Regnum Vegetabile vol. 2) compiled by Lanjouw and Stafleu. I wish to thank Mr. F. Ballard of Kew, Mr. J. A. Crabbe of the British Museum, Mme. Tardieu-Blot of Paris, and Mr. C. V. Morton of the United States National Herbarium for courtesies extended during the course of this study. Prof. Maximino Martinez of the Herbario Nacional del Instituto de Biologia, Universidad Nacional 1957] TRYON—PELLAEA SECTION PELLAEA 133 de Mexico, gave kind assistance on Mexican localities, and Prof. Efraim Hernandez- Xolocotzi of the Escuela Nacional de Agricultura, Chapingo, was of much help and good company in collecting Mexican pellaeas. I am especially grateful for collections of living plants which were received from Dr. E. F. Castetter of the University of New Mexico, Dr. Seville Flowers of the University of Utah, Dr. G. B. Ownbey of the University of Minnesota and Mr. J. W. Thompson of Seattle. Dr. Donald M. Britton of the University of Maryland has done the chromo- some studies which are reported in this treatment. The information has been in- corporated in the taxonomy and phylogeny of the group. A more detailed account of the cytology and evolution of these ferns is in preparation by Dr. Britton and myself I am indebted to Mrs. Bernadette Velick who has helped me in the preparation of the drawings. For his patient council and kindly criticism I am most grateful to my husband, Rolla M. Tryon, under whose guidance this study has been made. SYNOPSIS PELLAEA Link, Fil. Sp. Hort. Bot. Berol. 59. 1841, nom. conserv. Section PELLAEA. Eupellaea Prantl, Engl. Bot. Jahrb. 3:415. 1882. The natural characters of the group are best expressed in the imparipinnate, usually coriaceous blade; continuous marginal indusium; terminal sori on discrete and usually clavate vein ends; stipe and rachis tan to black, usually glabrous, smooth, without wings or conspicuous scale scars; costa of the segments mostly Obscure; sporangia usually long stalked; spores yellowish tan to brown, smooth or rugose. Although the section is not defined by a few striking characteristics it is a relatively homogeneous one. Among the species listed below some may not conform in all respects but are most properly placed here on the sum of their characters. Type species: Pellaea atropurpurea (L.) Link. Species: 1. P. Breweri D. C. Eaton, 2. P. glabella Mett. ex Kuhn, 2A. var. occidental Nelson) Butters, 2B. var. glabella, 2C. var. simplex Butters, 3. P.atropurpurea (L.) Link, + ternifolia (Cav.) Link, 4A. var. ternifolia, 4B. var. Wrightiana (Hook.) A. F. sea i on) D. C. Eaton, 6A. var. mucro- nata, 6B. var. californica (Lemmon) Munz & Johnston, 7. P. brachyptera (Moore) Baker, 8. P. sagittata (Cav.) Link, 8A. var. cordata (Cav.) A. F. Tryon, 8B. var. sagittata, bil P. ovata (Desv.) Weatherby, 12. P. inter- Kaulf.) Fée, 14. P. rufa A. F. Tryon, 15. k CE. F. u 14 edia Mett. ex Kuhn, 13. P. andromedaefolia ( P. myrtillifolia Mett. ex Kuhn. The position of Pellaea Bridgesii Hook., which in its general aspect resembles some of the species treated here and occurs mainly in California, remains a problem. l have excluded it on the basis of the conduplicate segments lacking reflexed margins, the short stalked sporangia which are borne on an elongated receptacle One quarter to one half the distance to the costa and which persist in a cup-like I Vor. 44 134 ANNALS OF THE MISSOURI BOTANICAL GARDEN form after dehiscence. It resembles some species of Notholaena in the length of the receptacle and in the abundant waxy indument produced among the sporangia. Key To SPECIES a. Stipe and rachis castaneous to black; spores usually with a sparsely rugose exospore. b. Scales of the rhizome concolorous; stipe and rachis terete or elliptical. c. c. Fronds monomorphic; pinnae sessile or the stalks somewhat decurrent on the rachis and arising at acute angles to the rachis; stipe and rachis glabrous or sparsely pubescent; scales of the rhizome discrete, a uni- form rust-brown color. d d. Rhizome massive with numerous compressed, short, articulated stipe bases; mature fronds with rachises mostly green in the terminal portion of the blade, the apical segments strongly decur- rent; the basal pinnae without persistent stalks often sessile or subsessile.. 1. P. Breweri; p. 1 . Rhizome moderately stout with few stipes persistent, these more or less spreading; mature fronds with rachises mostly brownish in the terminal portion of the blade, the apical segments SA de- current or stalked; the basal pinnae with persistent stalks PRSE ctae UE 2. P, glabella, p p. 1 c. Fronds dimorphic; pinnae stalked, not decurrent and arising at broad angles to the rachis; stipe and rachis scurfy with appressed pubescence; scales of the rhizome matted, rust colored with the young tips light EE Lebe Reid ee e 3. P. atropurpurea, p. b. Scales of the rhizome bicolorous with a sclerotic central stripe or base; stipe and rachis sulcate, convex or plane on the upper surface. c. e. Segments with opaque whitish borders, usually membranaceous; spo- rangia long stalked. Oklahoma to Utah southward to Argentina, Dominican Republic, Hawaiian Islands. f. f. Basal pinnae usually less than twice as long as broad, entire or divided into 3-11 segments; the pinnae-rachises up to 2 cm. long. A. P. ternifolia, p. 149 f. luda pinnae ie nuli more thín twice as long as broad, divided into 9-21 segments; the pinna-rachises up to E cm. long ea eg ees uL CE DAS p. 155 e. Segments entirely green, the border an nbn or narrow, lutescent Ee sporangia short stalked. Southern Oregon, California, Baja California. g. g. Ultimate segments of the basal pinnae shorter than the pinna-rachis; blade ioa m to rhomboid; pinnae ovate-linear. rr e TT De 199. g Üldse s segments of the Ter pinnae E dis the pinna-rachis; blade linear; pinnae semi-circula. 23. P. brachyptera, p. 162 w VA a Léi oo — > > 1957] TRYON—PELLAEA SECTION PELLAEA 135 a. Stipe and rachis straw colored to ruddy brown or darker at the base and mottled above; spores usually with a prominent rugose exospore. h. h. Rhizome moderately stout, compact; scales of the stipe base and buds tawny, broad, cordate or pseudo-peltate; fronds approximate; spores sparsely rugose 8. P. sagittata, p. 164 . Rhizome usually slender, cord-like and creeping, or moderately stout and compact; scales of the stipe base and buds tan, rust colored or brown, slender, acicular to lanceolate-triangular; fronds distant or ap- proximate; spores prominently rugose. i. i. Blades usually once or twice pinnate; pinnae with 1—3 segments; seg- ments elongate-lanceolate or lobed (stellate to ternate). j. j. Veins, except for the midvein, obscure; pinnae elongate-lanceolate, several times longer than broad, subcoriaceous. Nuevo León an Er: — e KT TT EE eee ee 9. P. notabilis, p. 170 j. Veins evident; pinnae stellate to ternate, nearly as long as broad, chartaceous or herbaceous. Central and western Mexico... Gi EE LE P Pcie IS 172 . Blades tri-quadripinnate or more highly divided; pinnae generally with 10—50 segments; segments mainly ovate. k. Rachis and pinna-rachises strongly flexuous.......11. P. ovata, p. k. Rachis and pinna-rachises straight. 1. l Mature rhizome scales bicolorous. Southwestern United States and Mexico. m. m. Veins obscure; segments coriaceous; buds pubescent. Texas, New Mexico, Arizona and northern Mexiecn, - 12. P. intermedia, p. 177 M 174 m. Veins visible on the lower surface of the segments; segments herbaceous; buds paleaceous. Calformng, ———— 13. P. andromedaefolia, p. 17 l Mature rhizome scales concolorous. Chile, South Africa. n. Rhizome cord-like, long-creeping; spores essentially smooth. South Atyca | 3 dee oe p rere 14. P. rufa, p. 184 n. Rhizome moderately stout, compact; spores rugose. Chis... 15. P. myrtillifolia, p. 184 No l. PELLAEA Breweri D.C. Eaton, in Proc. Amer. Acad. 6:555. 1865. (Lecto- type: Brewer 2038 YU!, Isotypes: MO! NY!, Paratypes: Brewer IQIQ, 2102 YU!, Photo of Holotype and Paratypes: GH! US! Fig. 1. Map 1. Allosorus Breweri (D.C. Eaton) O. Ktze. Rev. Gen. Pl. 2:806. 1891. Rhizome compact, ascendent, massive with many compressed bases of articu- lated stipes. Scales of the rhizome entangled or matted, uniform rust brown, these and those of the base of the stipes concolorous, lustrous, acicular, flexuous, the margins sinuate, not dentate, the cells of the scale base short, rectangular or with oblique end walls, the apex attenuate. Fronds 2.5-21.0 cm. long, bent IVor. 44 136 ANNALS OF THE MISSOURI BOTANICAL GARDEN or arcuate, monomorphic, the buds with sparse, reddish brown pubescence. Stipe and rachis terete, often flat when dry, sparsely pubescent, castaneous, lustrous, becoming darker and dull with age, the stipe with prominent lines of articulation, readily breaking at right angles to the axis, the rachis green at the apex. Blade 1.5 cm. long and 0.5 cm. broad to 16.0 cm. long and 3.5 cm. broad, linear-oblong, once pinnate or pinnate-pinnatifid, pure green, subglaucous, the pinnae of nearly the same length, departing at acute angles to the rachis, the upper pinnae entire, sessile, usually strongly decurrent on the rachis, the lower pinnae deeply 2-lobed, unilateral or mitten-shaped, rarely 3-lobed, subsessile or with short, green, rarely brown stalks, decurrent, the axils usually green, without pinna-rachises. Segments 0.5 cm. long and 0.3 cm. broad to 2.5 cm. long and 1.0 cm. broad, linear-ovate to deltoid, usually lobed, herbaceous, the border narrow, whitish, crenulate, the apex muticous. Sporangia sessile or subsessile, the annulus pale yellow, rarely amber colored. Spores 64 per sporangium, tetrahedral- globose, dark to light brown, with a scarcely prominent rugose exospore. Fig. 1. Pellaea Breweri, plant X 14; pinna; rhizome scale. P. Breweri resembles P. glabella more than any other member of the genus and a relationship exists here between Pellaea and several Mexican and South American. species of Cheilanthes. In P. Breweri and P. glabella the division and lobing of the frond, the position and attachment of the pinnae on the rachis and the articulation of the stipes are more like those of Cheilanthes Brande gii, C. allosuroides, C. pyrami- dalis and C. marginata than other pellaeas. The mature fronds of P. Breweri articulate at constricted bands on the stipe No special tissue is developed in the stipe but the hypodermis and cortex are com- 1857] TRYON-—PELLAEA SECTION PELLAEA EEN m € N mw Io. l. Distribution of Pellaea Breweri. : 2A. Distribution of Pellaea glabella var. occidentalis, 2B. var. glabella, 2C. var. simplex. [Vor, 44 138 ANNALS OF THE MISSOURI BOTANICAL GARDEN posed of relatively large, thin-walled cells. Fractures occur from the center of the stipe outward, possibly due to the mechanical strain set up by the drying of the deli- cate central tissue. This is also evident in P. glabella, particularly var. occidentalis. The chromosome number reported by Dr. D. M. Britton is n — 29 from speci- mens collected by Seville Flowers, Salt Lake County, Utah, Big Cottonwood Canyon, Wasatch Mountains, on July 25, 1953. Usually on granite rocks or basalt but also reported from limestone, at 1600- 3800 m. i Representative specimens: U © Pierson 834 (GH, MO, US). mano: Hitchcock & Mulick 11004 (NY, US, WS, WTU); Thompson 14022 (GH, MO, US, WTU). vran: July 25, 1953, S. Flowers j WS). : bron 5 & Leiberg 301 (US); Thompson 13220 (GH, NY, MO, US, WS, Alexander & Kellogg 3103 (MO, US, WS); Bolander 6243 (GH, MO, US, YU); Among 2 2. PELLAEA GLABELLA Mett. ex Kuhn, in Linnaea 36:87. 1869. (Lectotype: Specimen cited from Kimmswick, near St. Louis, Mo. B?). Rhizome moderately stout, compact, decumbent or ascendent, multicipital. Scales of the rhizome discrete and usually not matted, uniform rust brown, these and those at the base of the stipes concolorous, lustrous, linear, flexuous, often with horizontal bands of constricted cells, the margins sinuous or with broad irregular teeth, the cells of the scale base small, squarish, or elongate with oblique end walls, the apex somewhat attenuate, more or less tortuous, discrete, not entangled. Fron 1-36 cm. long, usually lax, arcuate or sinuous, monomorphic, the buds sparsely pubescent. Stipe and rachis terete, glabrous or nearly so, castaneous, lustrous, be- coming darker and dull with age, the stipe breaking at right angles to the axis, occasionally with perpendicular lines of articulation, the rachis castaneous or green- ish at the apex. Blade 0.75 cm. long and 0.5 cm. broad to 21 cm. long and 8 cm. broad, linear or ovate-lanceolate, once to twice pinnate, clear or bluish green, glaucous or subglaucous, the lower pinnae less than twice as long as the upper departing at acute angles to the rachis, the upper ‘pinnae entire or auriculate, sess! or subsessile, the lower pinnae entire, 2~5 lobed or once pinnate with 3—7 segments, stalked, the pinna-stalks somewhat decurrent, the axils colored as the rachis, the pinna-rachises up to 5 cm. long. Segments 0.5 cm. long and 0.3 cm. broad to 35 cm. long and 1.0 cm. broad, oblong to linear-ovate, usually lobed or auriculate, herbaceous to subcoriaceous, the border narrow, whitish, crenulate, the ape muticous to submucronate. Sporangia with long or short stalks, the annulus amber colored, occasionally oblique. Spores 32 or 64 per sporangium ellipsoidal or tetra- hedral- globose, occasionally dumbbell-shaped, rusty brown, the exospore sparsely to moderately rugose. : In his studies of P. glabella, Butters segregated two western species on the 1957] TRYON—PELLAEA SECTION PELLAEA à 139 basis of differences in frond size, shape of the pinnae and vascular bundle and the size of the spores.?? The size and number of spores per sporangium are of interest here for they reflect cytological differences. Britton?’ reports the chromosome number of P. glabella var. glabella from southern Ontario as n — 116 and con- siders this to be a tetraploid on the base number of 29. The plants are apogamous and the sporangia are 32 spored. He reports a tentative count of P. glabella var. simplex, collected by J. W. Thompson 17246 from Grant Co., Washington, as also ın = 116 and the material is 32 spored and apogamous. The monoploid number predicted in the above counts is realized in P. glabella var. occidentalis. Specimens collected by G. B. & F. Ownbey 1676 from Laramie Hills, Albany Co., Wyoming, have 64 spored sporangia, a normal sexual development and the chromosome count was determined as n = 29. Thus the 116 numbered members may be considered as tetraploids derived, at least in part, from P. glabella var. occidentalis. On rocks or cliffs, in crevices, usually on limestone, rarely on granitic rocks. Arizona north to southern British Columbia and northeast across the Uni States to Vermont. Key TO VARIETIES Spores 64 per sporangium; fronds 1-15 cm. long, once pinnate, the pinnae with 2 or 3 lobes, usually sessile or with stalk and rachis up to 0.2 cm. long. Alberta to Wyoming and South Dakota... 2A. var. occidentalis, p. Spores 32 per sporangium; fronds 1—36 cm. long, once to twice pinnate, the pinnae deeply lobed or with 3—7 segments, with stalk and rachis up to 5.0 cm. long. . Fronds 4—36 cm. long, the basal pinnae persistent, with stalk and rachis up to 5.0 cm. long. Vermont to Minnesota and Texas DOR EE om 2B. var. glabella, p. Fronds 1—20 cm. long, the basal pinnae usually withered, with stalk and rachis up to 1.0 cm. long. British Columbia to Washington and Utah tö New Mexicon a ae, lex, 6. 12 -—- vw Mo — A — 2A. PELLAEA GLABELLA Var. OCCIDENTALIS (E. Nelson) Butters, in Amer. Fern Journ. 7:82. Fig. 2A. Map 2A. Pellaea atropurpurea var. occidentalis E. Nelson, in Fern Bull. 7:30. 1899. (Holotype: : Nelspn 1919! RM}, Isotype: MO! NY!). Pellaea ae Rydb. in a N Y. Bot. Gard. 1:4. 1900. (Holotype: Rydberg 1191 !, Isotype: US!). Pellaea occidentalis (E. Nelson) Rydb. loc. cit. 466. 1900. Rhizome slender, somewhat ascendent. Scales of the rhizome and base of en stipes sinuate or the margins with few blunt teeth or erose, the cells of the base o the scales squarish. Fronds 1-15 cm. long. Stipe usually with perpendicular lines of articulation. Blade 0.75 cm. long and 0.5 cm. broad to 10 cm. long and 4 cm. broad, once pinnate or pinnate-pinnatifid, bluish green. Segments ovate to oblong, —Ó — 22 ss Amer. Fern Journ. 7:77. 1917; 11:39, 75. 1921. Amer, Journ. Bot, 40:575. 1953. [Vor. 44 140 ‘ANNALS OF THE MISSOURI BOTANICAL GARDEN mostly entire or 2- or 3-lobed and mitten-shaped, those of the basal pinnae wither- ing early with persistent petiolules, the apex muticuous, rarely submucronate. Sporangia stalks equal to the capsule length. Spores 64 per moop mostly tetrahedral-globose, the exospore with short, moderately prominent rug Generally these are smaller plants than the others of the P. sleid group but not conspicuous in any unique structure. Within the geographic range of the species they occupy a central position, mostly discrete and more restricted than the other varieties. As mentioned in the discussion of P. atropurpurea the possi- bility exists that hybrids were produced between that species and P. glabella var. occidentalis producing the eastern var. glabella and the western var. simplex. ranges of P. atropurpurea and var. occidentalis are contiguous in the Black Hills of South Dakota and nearly so in the vicinity of Banff, Alberta. A few specimens from the former region appear intermediate. P. glabella var. occidentalis is un- doubtedly the source from which the polyploid members of the complex have been derived either through hybridization or possibly autopolyploidy. In narrow fissures of limestone ledges at 1300-2800 m. Banff, Alberta, southeast in the United States to Wyoming and the Black Hills of South Dakota. : Representative specimens: Fig. 2A. Pellaea glabella var. occidentalis, plant X 14; pinna; rhizome scale. 1957] TRYON—PELLAEA SECTION PELLAEA 141 Ownbey 1676 (MO). sourH pakora: Palmer 37126 (GH, MO); Bull Springs, limestone district west of Custer, Black Hills, July 27, 1892, Rydberg 1191 (GH, NY, US). 2B. PELLAEA GLABELLA var. glabella Fig. 2B. Map 2B. Pellaea ee var. Busbii Mack. ex Mack. & Bush, Man. Fl. Jackson Co., Mo., 5. 19 (Holotype: Mackenzie 447 NY}, Isotype: ! Cep heen rpurea var. minima Eggert ex Farw. in Auer. Midl. Nat. 12:281. 1931. ! MAN etc berber us var. minima Eggert e 3 arw. loc. cit. 1931, in synon. Cassebeera atropurpurea var. glabella (Mett. ex Kuhn) Farw. fie its 1931. Pellaea atropurpurea f. glabella (Mett. ex Kuhn) Clute, Our Ferns, 109, 386. 1938. Rhizome moderately stout, decumbent. Scales of the rhizome and base of the stipes sinuate or the margins with irregular, broad teeth, the cells of the base of the scales mostly elongate with oblique end-walls. Fronds 4—36 cm. long. Stipe occasionally with perpendicular lines of articulation. Blade 2 cm. long and 1 cm. broad to 21 cm. long and 8 cm. broad, once to twice pinnate, bluish green. Seg- ments oblong-lanceolate, entire or auriculate or 3—5 lobed, persistent, the apex submucronate. Sporangia stalks as long or shorter than the capsule. Spores 32 per sporangium, ellipsoidal or tetrahedral-globose, the exospore with sparse, scarcely prominent rugae. Variety glabella appears to represent a genotype both vigorous and well adapted to the arid, limestone ledges which it occupies in the central and eastern United States. It is generally distinguished from the western varieties by its larger size. A We ^ aS Fib. 2B. Mp pe var. glabella, plant X V4; pinna; rhizome scale; section of the rachis with pinna- [Vor. 44 142 ANNALS OF THE MISSOURI BOTANICAL GARDEN As indicated in the discussion of P. atropurpurea, P. glabella var. glabella is similar in several respects although the two can be distinguished by several characters as follows: Pellaea glabella var. glabella Pellaea atropurpurea T kt Leaves dimorphic. =: Leaves lax, arcuate or sinuous, 4-36 cm. long. Blades linear to elongate-ovate. pe and rachis glabrous or with few bicolorous Leaves erect, straight, stiff, 6—50 cm. long. Blades elongate-triangular. Stipe and rachis hispid, scurfy. hairs nae at acute angles to the De somewhat Ades departing gradually (fig. 2B). Basal — with 3-7 — less than twice e pinnae ios -15 puer — 2 or as long as the upper pin more times as ids as the upper pinna Scales $ the rhizome E uniformly rust- Seales of the rhizome ap pb bg. dull brown, lustrous, the tips not entangled. wny to rust colored, te: young tips tan a and entangled. Pinnae at broad xi to the rachis and depart- ing abruptly (fig. 3). Margins of the fertile segments usually strong- ] s Margins of the fertile segments plane with y reflexed with relatively few sporangia ex- du ed. abundant sporangia exp On limestone, calcareous sandstone or dolomite cliffs or bluffs in shaded crevices or exposed ledges at less than 400 m. Northwestern Texas to the north shore of Lake Superior, Ontario, and east to Tennessee and Vermont. Representative specimens: CANADA. ONTARIO: Soper & Burcher 2074 (GH, MO); Taylor et al. 115 (US). ATE ) UNITED STATES: VERMONT: 1903, Baxter (MO); Fellows 481 (US). PENNSYLVANIA Aug. 1920, Wherry & We (US). NEw JERsEYv: Mackenzie 4621 (M EST VIR GINIA: Sept. 1937, Spawn (US). viRGiNIA: Adams & Wherry 2257 (GH); Wood 3644 ). orio: Roth et al. 670 (US); Camp 105 ET ckv: Aug. 1938, Richards (US). TENNESSEE: Jamison & Shaver 4161 (MO); Underwood & Sharp 2523 MO). micHican: Hagenah 1157 (US); Au 92, Wheele M, US). INDIANA R. M. & P. F. Tryon 4348 , US). murois: Evers et al. 671 (GH). MINNESOTA: M. F. & J. W. Moore 13563 (MIN, MO). 10wa: Anderson (Pl. Exsicc. Gray. 902) (GH, MO, RM MISSOURI: Bush 2837 (GH, MO, , US); Near St. Louis July, 1878, Eggert (MICH); Limestone rocks at Swope Park Sept 13; 1901, Mackenzie 1923, Buchholz (MO, US). 447 (NY, ay, Williams & Wilcox (US). Kansas: Hor E463 (RM, US, UT, WS). texas: D. S. T H. B. Correll 13085 (MO 2C. PELLAEA GLABELLA Var. SIMPLEX Butters, in Amer. Fern Journ. 7:84, 1917. (Holotype: Heacock 272 GH!, Photo: MOL. Isotypes: MO! NY!). Fig. 2C. Map 2C des ` weng Butters, in Amer. Fern Journ. 11:40. 1921, based on P. glabella vat. ` implex But Pellace oer var. simplex (Butters) Morton, in Leaf. West. Bot. 6:156. 1951. Rhizome moderately stout, ascendent. Scales of the rhizome and base of the stipes sinuate or irregularly dentate, the cells of the base of the scales squarish oF with oblique end walls. Fronds 1-20 cm. long. Stipe rarely with perpendicular lines of articulation. Blade 1.0 cm. long and 0.8 cm. broad to 14.0 cm. long and T 1957] TRYON—PELLAEA SECTION PELLAEA 143 4.5 cm. broad, usually once pinnate or pinnate-pinnatifid, or occasionally twice pinnate, clear green. Segments oblong-lanceolate, 3—5 lobed frequently with 2 or more small auricles, segments of the basal pinnae withering early with persistent petiolules, the apex muticous or submucronate. Sporangia long stalked, the annu- lus often oblique. Spores 32 per sporangium, ellipsoidal or tetrahedral-globose occasionally dumbbell-shaped, the exospore with sparse, scarcely prominent rugae. Fig. 2C. Pellaea glabella var. simplex, plant X V2; pinna; rhizome scale. The distribution of var. sim plex is discontinuous, occurring in southern British Columbia to southern Washington and in northern Utah to central Arizona and northern New Mexico with a gap in the intervening states. In the few specimens available there is no apparent morphological differentiation between collections rom the two isolated areas. In the discussion of P. atropurpurea it is suggested that this disrupted distribution may be due to the independent hybrid origin of P. &labella var. sim plex in the two isolated regions where the ranges of P. atropurpurea and P. glabella var. occidentalis were contiguous. On dry limestone or sandstone cliffs or ledges at 900-3000 m. British Columbia, Washington, Utah, Arizona, New Mexico, Colorado. Representative specimens: ser $ Draw, Selkirk Mts., July 14, 1904, Heacock 272 (GH, M . NY); Taylor & Dm Wh. Univ.B.C.); J. W. & E. M. Thompson 362 (MO, NY, US, WS, WTU). [Vor. 44 144 ANNALS OF THE MISSOURI BOTANICAL GARDEN UNITED STATES. WASHINGTON: Cliffs along the Columbia river, Aug. 11, 1892, Suks- dorf 2038 (MO, NY, US, WTU). uram: Rydberg & Garrett 9421 (NY, RM, US, UT). ARIZONA: Collom LJCr3 (US); Tracy 267 (YU). coromapo: Loew, Wheeler Exped. 1873 (YU). NEW MEXICO: 1883, Mathews (GH). 3. PELLAEA ATROPURPUREA (L.) Link, Fil. Sp. Hort. Bot. Berol. 59. 1841. Fig. 3. Map 3. Pteris atropurpurea L. Sp. Pl. 2:1076. 1753. (Holotype: Clayton 682, Point-look-out on the Rappahanock river, Va. BM). The Clayton collection was not found at the British Museum; a single frond, collected by Kalm, in the Linnean Herbarium is P. atropurpurea. Pteris tbalictroides Muhl. "Ind. Fil. Lancast." in Trans. Amer. Phil. Soc. 3:182. 1793, nom. nud., not P. thalictroides (L.) Sw. 1801. Asplenium atropurpureum (L.) Bernh. in Schrad. Neu. Journ. 17:10. 1806. Pteris spiculata Schkuhr, Krypt. Gewach. 1:92, £. roo. 1809. (Lectotype: f. 100, loc. cit.). Pteris atropurpurea var. venosa Pursh, Fl. Am. Sept. 2:668. 1814. (Holotype: f. IOI, Schkuhr, Krypt. Gewách.). Pteris atropurpurea var. punctata Pursh, Fl. Am. Sept. 2:668. 1814. (Holotype: f. 100, Schkuhr, Krypt. Gewich.). It is apparent from Schkuhr's text and figures that the e numbers are reversed in Pursh's reference to that work 1873. Pellaea atropurpurea var. cristata Trel. in Rep. Mo. Bot. Gard. 12:34. 1901. (Holotype: Pauls, Missouri, in 1899, MO Pellaea atropurpurea var. venosa (Pursh) Gilb. List N. Amer. Pterid. 18. 1901. Pellaea atropurpurea f. bifida Hahne, in Fern Bull. 12:117. 1904, nom. nud. Pellaea atropurpurea f. furcata Hahne, in Fern Bull. 12:117. 1904, nom. nud. Cassebeera atropurpurea (L.) Farw. in Amer. Midl. Nat. 12:281. Rhizome moderately stout, compact, decumbent, multicipital. Scales of the rhizome appressed and matted, rust colored in mass with the young tips light tan, these and those at the base of the stipes concolorous, tawny to rust colored, dull, linear-subulate, usually straight, the margins more or less dentate, the cells of the scale base small, mostly with oblique end walls, the apex attenuate, filiform, some- what tortuous, appressed and entangled. Fronds 6—50 cm. long, erect, straight, stiff, dimorphic, the buds lanate. Stipe and rachis terete or elliptical, strongly pubescent to hispidulous, atropurpureous to black becoming gray with age, t stipe breaking irregularly, without perpendicular lines of articulation, the rachis of the mature frond atropurpureous to the terminal pinna. Blade 5.0 cm. long and 1.5 cm. broad to 25 cm. long and 18 cm. broad, elongate-triangular, once to tripinnate clear to dark green, the lower pinnae usually 2 or more times as long as the upper, departing at broad angles to the rachis, the upper pinnae usually entire, subsessile or short stalked, the lower pinnae entire, deeply 1-3 lobed, or usually once pinnate with 3-15 segments, stalked, the pinna-stalks abruptly departing, not decurrent, the axils tan, the pinna-rachises up to 12 cm. long. Segments 1.0 cm. long and 0.5 cm. broad to 7.5 cm. long and 1.0 cm. broad, lanceolate or linear-oblong, entire or auriculate or ternately lobed, subcoriaceous, the undersurface often pubescent along the veins, the border broad, white or often opaque, crenulate, the fertile segments revolute, the apex muticous or mucronate, the mucro 0.25-1.50 mm. 1957] TRYON—PELLAEA SECTION PELLAEA 145 long, flat, with a whitish margin contiguous with that of the segment. Sporangia with long stalks exceeding the length of the capsule, the annulus amber colored. Spores 32 per sporangium, ellipsoidal, rarely dumbbell shaped, yellow, with a prominent, rugose-lacy exospore. Fig. 3. Pellaeg atropurpurea, plant X, V2; pinna; rhizome scale; section of rachis with pinna-stalks. JN oL. 44 146 ANNALS OF THE MISSOURI BOTANICAL GARDEN 3. Distribution of Pellaea atropur purea — X, literature records, see text. 1957] TRYON-—PELLAEA SECTION PELLAEA 147 P. atropurpurea is widely distributed throughout the eastern United States and wherever it grows it is usually found in abundance. I have observed the species at several stations in New England, Indiana, Minnesota, Texas and Missouri and while much variation exists in frond size the form is essentially uniform. Specimens transplanted from Gray Summit, Missouri, growing in the greenhouse over a period of three years, are more luxuriant than field specimens. These plants produced copious spores which were easily sown, and although only antheridia were produced, sporelings developed. In the eastern United States P. atropurpurea often grows on the same rocks or ledges as P. glabella var. glabella and although the two are readily distinguished when closely observed their relationship has been in question. The cytological and morphological data indicate the possibility of hybrids produced between P. atropurpurea and P. glabella var. occidentalis. The former, apogamous with 87 chromosomes, crossed with the normal sexual P. glabella var. occidentalis, may have given rise to P. glabella var. glabella to the east and P. glabella var. sim- blex to the west. Both of these varieties of P. glabella have chromosome numbers of 116 and are apogamous. Some morphological comparisons of the four entities are made in the following chart. It might be expected that the hybrids would bear a stronger resemblance to P. atropurpurea from which they would receive a triple genome. P. glabella var. P. glabella var. | P. glabella var. simplex i P. atropurpurea glabella Frond length 6-50 cm. 4-36 cm. 1-20 cm. 1-5 cm. Lines of articula- None Occasionally Rarely Abundant tion on the stipe Pubescence on Abundant bi- | Sparse bicolorous | Sparse bicolorous Ser a stipe and rachis colorous hairs, | hairs hairs rarely with bi- ispid colorous hairs Length of pinna- Up to 120 mm. | Up to 50 mm. Up to 10 mm. Usually sessile, stalk and rachis rarely up to 2 mm. Number of seg- m Ka 3-5 1 — this may g 3—15 7 ments per pinna be 2-3 lobed A remark by Butters”! in his studies of the pellaeas in the western United States Pertaining to Pellaea glabella var. occidentalis is appropriate here “ . . . it oer one extreme of a series and P. atropurpurea at the opposite extreme". From a morphological survey of the species in section PELLAEA, P. atropurpures appears to be more closely related to P. ternifolia and P. notébilii w sr Mexico. P. atropurpurea very probably originated from such an austral source and has spread to the north and east across the United States and into Canada. — ^^ Amer. Fern Journ. 7:82. 1917. [Vor. 44 148 ANNALS OF THE MISSOURI BOTANICAL GARDEN Cytological studies on the group seem to support the morphological evidence on the relationships between P. atropurpurea and P. ternifolia. Dr Irene Manton” reports a chromosome count of n & 2n = 87 for P. atropurpurea which is inter- preted as a triploid pending the existence of a monoploid relative with 29 chromo- somes. Dr. Britton confirms this report of 87 for P. atropurpurea from material collected by myself at Gray Summit, Missouri. He also reports counts of both n = 29 and n — 58 for P. ternifolia. On exposed or shaded ledges and crevices of limestone cliffs, masonry walls or rocky loam of forest floor, at 300—3200 m. Guatemala, northward in Mexico along Sierra Oriental, northeast across the United States to Vermont with outlying stations in the Black Hills of South Dakota, Keweenaw Point on Lake Superior, southeastern British Columbia and on the north shore of Lake Athabaska, Saskatchewan. Some of the specimens from the last two localities are depauperate but can be identified on the basis of the branching of the pinna-rachises, the slender, attenuate scales, prominent ornamenta- tion of the spores and the long basal pinnae. Two literature records from Colorado, indicated on the map by X, are taken from W. A. Webber in University of Colorado Studies, Biology Series No. 3:68. 1935. Representative specimens: CANADA. BRITISH E Eastham 11068 (MO, WTU, Univ.B.C.). sASKATCHE- wAN: Raup 6558 (CU, GH). UNITED STATES. vixi T: Fax in 1877 Ke NEBC). MASSACHUSETTS: Bissell r MO). N Eames oot (M US), NEW "ees Crain 4617 (MO). PENNSYLVANIA: Ee 711 (MO, Us DISTRICT OF COLUMBIA: Meso 6393 (GH, MO, NY). 1896, Waters OU VIRGINIA: oe ES we 2265 (MO, US). I UC); F NY, US). ALABAMA: D. E & H. B. cun 8284 (GH, NY, US). onto: Demaree Wo). O). MINNESO : 1899, Eureka, Pauls (MO); Steyermark 65394 (F). ARKANSAS: Palmer 6904 (MO, US). SOUTH DAKOTA: Degener & Peiler 16701 (NY, US); Rydberg 1190 (GH, US). NEBRASKA: Webber Herb. in 1889 (MO). Kansas: Horr E459 (GH, KANU, US). OKLAHOMA NY : IAPAS: Ghiesbreght 216 mp, GH, YU). n HUAHUA: Stewart 2475 (GH). coanura: Stanford et al. 391 (MO, US); Wynd & Mueller 358 (GH, MO, NY): NUEVO LEÓN: C. H. & M. T. Mueller 362 (GH, US). PursLa: Arsène 3548 (GH H, MO, US). oaxaca: Conzatti 1912 B 4279 (US). UON: i ep 357 (GH, NY, US). TAMAULIPAS: Stanford et al. 943 (MO, US). au 3008 (GH, NY). GUATEMALA. ite EC Standley gesot (E, ee $2001 (F). ?? Manton, I. Prob. Cytol. and Evol. of the Pterid. Cambridge Univ. Press, 1950. 1957] TRYON—PELLAEA SECTION PELLAEA 149 4. PELLAEA TERNIFOLIA (Cav.) Link, Fil. Sp. Hort. Bot. Berol. 59. 1841. Rhizome moderately stout, elongate, decumbent, multicipital. Scales of the rhizome appressed, scarcely entangled, brownish or the tips tan, these and those at the base of the stipes usually bicolorous with a narrow sclerotic stripe, subulate, straight or falcate, the margins usually irregularly dentate or erose, the cells of the scale base small, irregular, those above many times longer than broad, the apex attenuate, filiform, appressed and scarcely entangled. Fronds 4—50 cm. long, erect, straight and stiff, monomorphic, the buds sparsely pubescent, rarely lanate. Stipe and rachis convex or plane on the upper surface or sulcate, rarely terete, glabrous or rarely pubescent, glaucous, castaneous becoming darker with age, to ebony black, the stipe breaking irregularly without articulation lines, the rachis of the mature frond black or castaneous to the terminal pinna. Blade 3.0 cm. long and 0.5 cm. broad to 32 cm. long and 6 cm. broad, linear, lanceolate or elongate- triangular, once or bipinnate, clear or grayish green, the pinnae of nearly the same length or the lower longer, departing at broad angles to the rachis, entire, ternate, or with 3-11 segments, subsessile or stalked, pinna-stalks not decurrent, the axils colored as the rachis or tannish, the pinna-rachises up to 2 cm. long. Segments 0.5 cm. long and 0.5 cm. broad to 4 cm. long and 1.0 cm. broad, lanceolate to narrowly oblong, entire or ternately divided, subcoriaceous, the borders whitish, crenulate, mucronate, the mucro 0.5—1.5 mm. long, flat or reflexed, with a whitish margin contiguous with that of the segment. Sporangia long stalked, the annulus amber colored; glands with ceraceous indument may occur among the sporangia. Spores 64 per sporangium, mostly tetrahedral-globose, yellow-tan, with a sparse, scarcely prominent, rugose exospore. P. ternifolia is a critical species for it bears relationships to several others and problems relevant to them. The series from this species through P. brachyptera shows greater cohesion than any other in the genus. The entities differ by quanti- tative characters and their ranges for the most part are geographically distinct. The pinnae of P. ternifolia var. Wrightiana are longer and more dissected than those of P. ternifolia var. ternifolia and are further expanded in the following two species. Two chromosome numbers have been reported for P. ternifolia var. ternifolia by Dr. Britton. A collection from Cuernavaca, State of Morelos in Mexico, R. M. & A. F. Tryon 5105, and one from El Pedregal on the south edge of Mexico City, R. M. & A. F. Tryon 5146, have chromosome numbers of n = 29. A collection from Valle de Bravo, State of Mexico, was reported as n = 58 and another from Cuernavaca was a probable polyploid although the number was not definitely established. It is unfortunate that both of the polyploid numbered collections have been lost. ) P. ternifolia is also suspected of entering into crosses which have produced P. atropurpurea and possibly P. Pringlei. The range of P. ternifolia, though more extensive, resembles that of P. ovata but there is no evidence from spore counts, as exists in the latter, of a wide-ranging apogamous strain. [ VoL. 44 150 ANNALS OF THE MISSOURI BOTANICAL GARDEN KEY TO VARIETIES Blade linear, pinnate or pinnate-pinnatifid; pinnae ternate or entire, without pinna-rachises; southern Texas and Arizona southward along the Cordil- lera to central Chile, Dominican Republic, Hawaiian Islands... ———- 4A. var. fernifolia, p. 150 Blade linear to usually narrowly triangular, bipinnate; pinnae lobed or divided into 3—11 segments, pinna-rachises 4—20 mm. long; southern Oklahoma to northern Mexico. — 4B. var. Wrigbtiana, p. 153 4A. PELLAEA TERNIFOLIA var. ternifolia. Fig. 4A. Map 4A. Pteris ternifolia Gar, Dier PI. 266. 1802. (Lectotype by C. Christensen in Dansk Bot. Ark : Née, Peru MA; Isotype: Pteris Kile age se das in Lam. Ency. Meth. Bot. 5: Hn. 1804. (Holotype: Joseph Jussieu, P!, Photo: GH!). feris subverticillata Sw. Syn. Fil. 103. 1806, based on fiti ternifolia Cav Pteris triphylla P in Merc. Chil. 16:748. 1829, nom. nud.; ex Colla, Herb. Pedem. 6:199. 1836. (Holo type: Bertero, in 1828, TO; ; lioyper. GH! SGO). STE as to the type from Lou er, in Rev. Univ. (Univ. Ce? Gigs, 25:111. 1940; as to the Colla reference, Pichi-Sermolli,in Webbia 8:135. 19 Allosorus subverticillatus (Sw.) Presl, Tent. Pterid. 153. RT Platyloma ternifolia (Cav.) J. Sm. in Bot. Mag. 72: Comp 21. 1846. Allosorus ternifolius (C av.) Kze. ex Kl. in Linnaea 20:339. 1847. Pellaea Weddelliana Fée, in Mém. Soc. Mus. Hist. Nat. Strasb. 5:74. 1857. (Mém. Fam. oug. 8). (Lectotype: Weddell d K!, Lsoty pes: PI TUD. Pellaea heuse arte Fée, in Mém. Soc. Hist. Nat. Strasb. 5:114. 1857. (Mém. Fam. Foug. om. provis. (Holotype: Schaffner 321 Bd mucronata Fée, Cat. CMS da Lycop. Me ex. 8. 1857. (Mém. Fam. Foug. 9) nom. nud., not (D. C. Eaton) D. C. Eaton, in Torr. U. S. & Mex. Bound. Bot. 233. 1859. (Isotype: Scheffer. aoe YU!). — regu (F j^ e Ind. eier 255. 1861; not putes ex Hook. Syn. Fil. 18 2:99. 8, Mart. & Gal. in Mém. Acad. Brux. 15:75. Cheilant js. Veddeliana Se Ke Ind. Ti 255. 1861. Cheilanthes ternifolia v.) Moore, loc. den ternifolia (Car) Keyserl. nbs Cyath. Herb. Bung. 29. 73. Adiant ge vt d ernatum Sessé & Mes Pl. Nov. Hisp. 182. 1887—90; ed. i 169. 1893, not & Bonpl, ex Willd. Pellaea derifoli var. eile e vehi: in Ark. Bot. 20: 18. 1926. (Holotype: Asplund 30 Cassebeera ieren? ee ^ Ree in Amer. Midl. Nat. 12:281. — tags Se Hall, in Amer. Fern Journ. 37:111. b bd Brandegee, o UC!, hot t USD. Se of the ere and base of the stipes bicolorous with a slender central sclerotic stripe usually narrower than the border, the margins entire at the base, dentate above. Stipe and rachis plane on the upper surface or sulcate, atropur- pureous or black. Blade linear to narrowly lanceolate, once pinna ternate or entire, sessile or subsessile, without pinna-rachises. Without glands among the sporangia. Several segregates of P. ternifolia var. ternifolia have been made on differences in the form of the pinnae, length of the pinna-rachises and serration and color of the scales. Such variations are not sufficiently clear for taxonomic recognition although they are of interest in the dynamics of the species. The segregates, P- 1957] TRYON—PELLAEA SECTION PELLAEA 151 Fig. 4A. Pellaea ternifolia var. ternifolia, plant X Vs; pinna; rhizome scale. Brandegeci at the southern end of Baja California and P. Weddelliana in Bolivia and northern Argentina show slight differences which would be expected in popu- lations isolated from the main geographic range of the species. Collections from several localities in central Mexico which have entire pinnae, pubescent rachises and lax fronds are presumably Fée's P. lanuginosa. This may be cytologically distinct for a pubescent specimen from Valle de Bravo is reported to have a chromosome number of n — 58. Two other collections of P. ternifolia var. ternifolia from the Vicinity of Mexico City are reported as n = 29. In crevices or among igneous rocks, on dry earthen walls in Peru, in sun or semi-shade, at 200-4000 m. [I Vor. 44 192 ANNALS OF THE MISSOURI BOTANICAL GARDEN H ore 4A. Distribution of Pellaea ternifolia var. ternifolia, 4B. var. Wrightiana. 1957] TRYON—PELLAEA SECTION PELLAEA 153 Southern Arizona and Texas; Mexico; Guatemala; Nicaragua; Hispaniola; Vene- zuela, Colombia to Argentina; Hawaiian Islands. Representative specimens: UNITED States. arizona: Phillips et al. 2565 (US). vTexas: E. J. E prO (GH, MO, NY); R. M. & A. F. Tryon 5097 (GH, K, MO, MEXU, U, UC, US, UT). UATO: Sept. ; (F, MO, NY, US). uimparco: Clokey 1892 (MO). Jausco: . McVaugb 12998 (US); 5d MEXIC "e H 54 (US). P Palmer 982 (F, MO, NY, US, YU). stnatoa: Gentry 6434 (M sonora: White 4221 (GH). TLAXCALA: Hernández- Le ub 334 (US); en & Hernández 44445 (US). veracruz: Seaton 279 (F, GH, US). zacatecas: Rose 2792 (US). GuATEMALa: Skutch 806 (GH, ard. Standley 58350 (F, US). NICARAGUA: get 11005 (F). Dominican REPUBLIC: Ekman 11706 (NY, US); R. H. & E. S. Howard 9137 (GH, MO, NY, US). VENEZUELA. MERIDA: Gebriger 251 (US). CoLoMBiA. Boyacá: Fosberg 22210 (US). cunpinamarca: Pennell 2548 (F, GH, MO, NY, US). VALLE: re 20829 (F, US). cauca: Killip 6882 (GH, NY, US). MAGDALENA: ifriz 471 (US). NORTE DE SANTANDER: Garganta 1225 (F). SANTANDER: Killip & Smith UT (F, GH, NY, US). Ecuapor. azuay: Rose ef al. 22872 (NY, US). imMBABURA: Mexia 7403 (UC, US). Loja: Fosberg & Giler Ge (US). picnincna: Heilborn 177 (GH, US). TUNGURA- Hua: Pachano 114 (U Peru. Near Se Guarimaya Valley, Née (F). AREQUIPA: nell T3073 TA GH, NY, US). cuzco: Herrera 3172 (F). Junin: Killip © Smith 21817 (NY, US). ra LIBERTAD: Stork & Horton 9997 (F, = a Macbride & Feath stone 286 (F, US). puno: R. S. Williams 2633 (GH, N S ecd CHUQUISACA: Tomina, 1845—46, Weddell spi ecu YU). eie uchtien 1139 (US). porosi: Asplund 3071 (UPS, US), 4904 LA PAZ: Ban Ü H, MO, NY. ds US); Giele 3 (GH, US). mama: Fiebrig 3138 (GH, P, UC, GENTINA. BUENOS AIRES: Eyerdam et al. 23403 (GH, UC). CATAMARCA: 4 weg sen 1251 (GH, MO, UC, US). compóma: Hieronymus 475 (NY, US). Jujuy: Se e C HD West 6249 (US). LA RIOJA: Honisiler 5278 (MO). sata: Venturi 9003 ). TUCUMAN: Venturi jn ud (6B. O.) E : : os (GH, US). cauríw: Claude- jid 2658 ANTOFAGASTA: Jobnsto TA 3 d vatpivia: Hollermayer 1898 (MO, US). HILE. US). JINS: (US) oe May 1828, n ( MO, US). Maur Degener Hawauan IstANDs. Hawa: Mann & Brigham 262 (GH, $ Wiebke. fer (GH, MO, US). oaHu: Degener 10419 (MO, US). 4B. PELLAEA TERNIFOLIA var. Wrightiana (Hook.) A. F. Tryon, comb nov. Fig. 4B. Map 4B. Pellaea Wrightiana Hook. Sp. Fil. 2:142. 1858. (Holotype: Wright 2130 K!, Isotypes: X ! NY! US! YU!).- Cheilanthes mucronata var. Wrightiana (Hook.) Moore, Ind. Fil. 248. 1861. Scales of the rhizome and base of the stipes bicolorous, the central sclerotic stripe as broad or broader than the border, the margins dentate to the base. Stipe [Vor. 44 154 ANNALS OF THE MISSOURI BOTANICAL GARDEN Fig. 4B. Pellaea ternifolia var. Wrigbtiana, plant X V4; pinna; rhizome scale. and rachis sulcate, castaneous, becoming darker with age. Blade narrowly tri- angular, bipinnate. Pinnae lobed or divided into 3-11 segments, usually sta the pinna-rachises 0.4-2.0 cm. long. Glands with ceraceous indument rarely occur among the sporangia. "vr P. ternifolia var. Wrigbtiana has formerly been treated as a distinct species. n is essentially similar to P, ternifolia var. ternifolia in the architecture of the bla and is considered a northern extension of the species having quantitative differences not sufficient for specific separation. "ET In crevices of igneous or limestone rocks or on rocky or gravelly hillsides, sun or partial shade, at 1400-2300 m. Oklahoma to Arizona; northern Mexico. Representative specimens: 1957] TRYON—PELLAEA SECTION PELLAEA 155 UNITED STATES. OKLAHOMA: Waterfall 6456 (GH). Exas: Moore & Steyermark 3010 (GH, MO, NY, US); Mueller 8265 (GH, MO, US); E. J. Palmer 34260 (F, MO, NY); Reverchon 1216 (F, GH, MO, NY, US); R. M. & A. F. Tryon 5041 (MO, US) 5096 (GH, K, MO, U, US). New mexico: Eastwood 8266 (GH, US); Goodding 766 (GH, NY, US) ; Dec. 1880, Rusby (MO, US, YU); 1851, Wright 2130 (GH, K, NY, US, : ARIZONA: Blumer 1867 (F, NY, US); Sept. 1880, Engelmann (MO); B. & R. R. Maguire 12027 (US) ; Phillips & Reynolds 2943 (GH, US). MEXICO. BAJA CALIFORNIA: April 1889, Brandegee (GH, US); Gentry 4283 (GH, MO). cHmuaHua: Johnston 8090 (GH); Pennell 19209 (US). con A: Johnston & Muller 1309 (GH). sonora: 1895, Druery (US). 5. PELLAEA LONGIMUCRONATA Hook. Sp. Fil. 2:143. 1858. (Holotype: Wright 2131 K!, Photo: GH!, Isotypes: GH! US! YU!). Fig. 5. Map 5 Pellaea Wrightiana var. longimucronata (Hook.) Davenp. Cat. Davenp. Herb. Suppl. 46. 83. Pellaea truncata Goodding, in Muhlenb. 8:94. 1912. (Holotype: Goodding 977 US!). Cassebeera ternifolia var. longimucronata (Hook.) Farw. in Amer. Midl. Nat. 12:282. 1931. Rhizome moderately stout, elongate, decumbent, multicipital. Scales of the rhizome and base of the stipes appressed, brown, bicolorous, with a central sclerotic stripe usually broader than the border, subulate, straight or falcate, the margin erose-dentate, the cells of the border trapezoidal, usually not more than five times longer than broad, the apex dentate, not attenuate. Fronds 12-38 cm. long, erect, straight and stiff, monomorphic, the lower pinnae often sterile, the upper fertile, the buds sparsely pubescent. Stipe and rachis sulcate, glabrous or nearly so, usually glaucous, castaneous becoming darker with age, the stipe breaking irregularly without articulation lines, the rachis of the mature frond castaneous to the termina Pinna. Blade 8 cm. long and 4 cm. broad to 22 cm. long and 12 cm. broad, tri- angular, acuminate, bipinnate, rarely tripinnate, grayish green, the lower pinnae longer, departing at broad angles to the rachis, once pinnate with 9-21 segments, rarely bipinnate, sessile or short stalked, the pinna-stalks not decurrent, the axils colored as the rachis, the pinna-rachises up to 7 cm. long. Segments 0.3 cm. long and 0.1 cm. broad to 1.5 cm. long and 1.0 cm. broad, ‘narrowly oblong to oval, entire, subcoriaceous, the border narrow, white, crenulate or dentate, mucronate, the mucro 0.5-2.0 mm. long, flat or reflexed, green at the base, with a white Margin extending into the attenuate apex and contiguous with the borders of the segment. Sporangia usually long stalked, the annulus amber colored, glands with Céraceous indument occurring among the sporangia. Spores 64 per sporangium, mostly tetrahedral-globose, yellow-tan, the exospore of short, scarcely prominent gae. During field studies made in July and August of 1950 in southern Texas, New Mexico and Arizona six species of the section were collected and P. longimucronata Was most frequently encountered. It grows in extremely xeric habitats often devoid of other vegetation. I In clefts of igneous rocks or among boulders or gravelly soil, in open sun, at 1250-2000 m. Texas to Colorado, Arizona, Utah, Nevada; northern Mexico. [Vor. 44 156 ANNALS OF THE MISSOURI BOTANICAL GARDEN WU e 00 SZ ( NU fa SHAA N W nf dar ZB Fig. 5. Pellaea longimucronata, plant X 14; ; pinna; rhizome scale. Representative specimens: RADO: April 24, 1874, Brandegee (F, MO); es 25, 1874, aen pe Nev MEXICO: Sept. 10, 1880, Rusby (F, GH, MO, NY, US, iun R. M. & A. F. Try a (GH, K, MEXU, MO, U, US); 1897, Ze 99 (MO, NY, US); Gare Wright T ( ; NY, US). uran: Gould I fo y, US); [one 497 (NY). ARIZONA Blumer 1970 (GH, US); Mule Mts., (Ser? 19 Goodding 977 PS Moi 649 ix MO, U); March 11, 1884, peer (F, US); Fh ier 464 (F, MO); R e (MO). Nevapa: Clokey 8257 (F, MO. NY); Bl Pos 737 (F, LA : NY, US); Jones 5055 (MO, NY). EXICO. BAJA CALIPORN NIA: April 20, 1889, Brandegee (GH, NY); Palmer 554 (GH). sonora: F. A. Phillips 511 (GH). 1957] TRYON-—PELLAEA SECTION PELLAEA 157 n . PELLAEA MUCRONATA (D.C. Eaton) D. C. Eaton, in Torr. U. S. & Mex. Bound. Bot. 233. 1859, not Fée, Cat. Meth. Foug. Lycop. Mex. 8. 1857 (Mém. Fam. Foug. 9) nom. nud. Rhizome moderately stout, compact or somewhat elongate, decumbent, multi- cipital Scales of the rhizome and base of the stipes appressed, brown, bicolorous with a central sclerotic stripe as broad or broader than the border, straight or falcate, subulate, sparsely dentate to entire at the base, dentate to erose above, the cells of the border with oblique end walls, usually not more than five times longer than broad, the apex filiform, tortuous, entangled. Fronds 6.5—43.0 cm. long, erect, stiff, monomorphic, the lower pinnae sometimes sterile, the upper fertile, the buds strongly pubescent or lanate. Stipe and rachis sulcate or plane on the upper surface, glabrous or nearly so, glaucous, castaneous, the stipe often bent or curled, breaking irregularly without articulation lines, the rachis of the mature frond Castaneous to the terminal pinna. Blade 3.5 cm. long and 1.0 cm. broad to 25 cm. long and 18 cm. broad, ovate-triangular to rhomboid, acuminate, often abruptly $0, bi-to tripinnate, rarely quadripinnate, grayish green, the central or lower pinnae longest, departing at broad or acute angles to the rachis, once pinnate or bipinnate with few to 40 segments, subsessile or short stalked, the pinna-stalks not decurrent, the axils colored as the rachis, the pinna-rachises up to 10 cm. long, exceeding the length of the segments several times. Segments 0.2 cm. long and 0.1 cm. broad to 1.2 cm. long and 0.6 cm. broad, narrowly oblong, entire or ternate, subcori- aceous, the border undifferentiated, crenate, mucronate, the mucro 0.25-1.50 mm. long, flat, thickened, green at the base, the apex lutescent, distinct from the revo- lute borders of the segment. Sporangia short stalked, the annulus amber colored, glands with ceraceous indument occurring among the sporangia. Spores 64 per sporangium, mostly tetrahedral-globose, yellow-tan, with a sparse, scarcely prom- inent rugose exospore. KEY TO VARIETIES Pinnae distant, not imbricate, at broad angles to the rachis; pinnules few to 40 Per pinna, usually ternate; northern California to Baja California and coastal islands, usually belów 1800n$ — ——— — 6A. var. mucronate, P. 158 Pinnae imbricate, especially on the apical portion of the frond, ascending at acute angles to the rachis; pinnules fewer than 20 per pinna, usually entire; Sierra Nevada and southern California, usually above 1800 m... eis i ...6B. var. californica, p. 161 This Californian species exhibits a plasticity of form as great bed that of the wider ranging P ternifolia. It occurs among igneous rocks in arid or semi-arid Montane regions of the state and is undoubtedly influenced by the topographic and climatic diversities. Mason’s2® observations, especially pertaining to this region, —— %8 Madroño 8:209, 241. 1946. [Vor. 44 158 ANNALS OF THE MISSOURI BOTANICAL GARDEN on the effect of topography and climate on the distribution of restricted species helps to interpret the variations of this fern. In southern California Ewan?" has studied P. mucronata, particularly those plants occurring at the higher altitudes. He considers P. mucronata var. californica to be derived from P. mucronata var. mucronata through isolation and consequent morphological changes in extreme environments. Less distinct variations in collections of var. mucronata from Kern, Tulare and Fresno counties have small, closely set pinnules and have been noted by Maxon.?® Other specimens from northern California are distinctive in having com- pound pinnules with 3-11 segments. A few specimens, two of which are cited in the treatment of the next species, have entire pinnules and lax pinnae intermediate between P. mucronata var. mucronata and P. brachyptera and are possibly hybrids. 6A. PELLAEA MUCRONATA var. mucronata Fig. 6A. Map 6A. M mucronatus D. C. Eat r. Jou 2:138. SEU (Holotype: . Eaton, near Bay d Sin. riera Calif Se Mee Ain Pelleae longimueronata var. minor Hook. Sp. Pi, :143. 1858, nom. mis based on Maj pe Calif. K, Isotype o Pellec ortho H ook. Sp. Fil. 2:143. SY (Holotype: Hartweg 2042 K!, Photo: Hi, Isotypes: GH! Pl). Civiles thes Decal: (D.C. Eaton) Moore, Ind. Fil. 45, 248. 1861. Allosorus ornithopus (Hook.) O. Ktze. Rev. Gen. Pl. 2:806. 1891 T Cassebeera ternifolia var. ssi geo d Farw. in Amer. Mil. Ims 12:282. 19 Cassebeera ternifolia var. mucronata (D.C. Eaton) Farw. loc. ! Cassebeera ternifolia var. stipitate Far. See cit. (Holot Me ueri in 1881, MICH!). | Fronds 6.5—43.0 cm. long. Blade 3.5 cm. long and 1.0 cm. broad to 25 cm. long and 18 cm. broad, ovate to rhomboid, bi- or tripinnate, rarely Sect more or less plane, the apex abrupt or acuminate. Pinnae linear to ovate, d imbricate, at wide angles to and widely spaced on the rachis, those of the e region often longer than the basal, with few to 40 segments per pinna. Segmen widely spaced on the pinna-rachises, usually ternate. A collection of Ezra C. Knopf from Santa Catalina Island bears the accompany ing information — infusion of the dried leaves of E E is much prized by the Spanish, pager’ © and some cat the white residents, as a tea. e drunk it and found it quite palatable. go said to be good for the kidneys and as Se purifier. fter the rainy season is pu d e plant dries and can be picked any time dining the summer. The Indian name is Calagu'2- gë In crevices of igneous rocks or among boulders, full sun or semishade, at 2 2000 m. California; Baja California. Representative specimens: UNITED STATES, CALIFORNIA: Alexander & Kellogg 1993 (MO); Austin Fe ar D US); Baker 2613 (GH, MO, US); San ER 1881, Blaisdell (MICH); H. E. B YU); (F, MO, US); Clokey 4840 Get P); Bay of San Francisco, Major A. B. Baton C (1775 Oct. 1880, Engelmann (MO); F osberg n XE, MO, US); Hansen 657 (M pers Hartweg 2042 (GH, K, P); Heller 7816 (GH, MO, NY, P, US); Kellogg "3 H 27 Journ. Wash. Acad. Sci. 25: . i 1935. ?8 Proc. Biol. Soc. Wash. 30:179. x 1957] TRYON—PELLAEA SECTION PELLAEA 159 1170 (MO, US, YU); McKelvey 5053 (GH, US); Palmer 424 (GH, MO); July 14, s Ce (GH, MO, US); Torrey 595 (GH, NY, US); R. M. & A. F. Tryon 5058 (GH K, MEXU, MO, U, UC, US, UT). MEXICO. BAJA CALIFORNIA: Wiggins 9821 (US); Wiggins & Gillespie 4151 (GH, MEXU, MO, US). GUADALUPE ISLAND: Palmer 100 (GH, MO, NY YU). oh aK A a gu GR D yos IA = + Gei RY ih REH BEN 3 (XS (7 X X Ai Yr N N m- KR EK Fig, 6A. Pellaea mucronata var. mucronata, plant X Y2; pinna; rhizome scale. [Vor. 44 160 ANNALS OF THE MISSOURI BOTANICAL GARDEN NI - L D ES 5 i ' n dr I : I | "3 es S E e zn eui. ce 5. Distribution of Pellaea longimucronata 6B. var. californica, ta, “esha vege of Pellaea mucronata var. mucron4 7. Distribution of Se brachypter 1957] TRYON—PELLAEA SECTION PELLAEA 161 6B. PELLAEA MUCRONATA var. CALIFORNICA (Lemmon) Munz & Johnston, in Amer. Fern Journ. 12:106. 1922. Fig. 6B. Map 6B. Ages els cii var. PEE Lemmon, Ferns dni Coast, 10. 1882. on, May 30, 1876, UC!, Isotypes: F! GH! ow aiia bella ed in Se Chron. 215. 1873. from California K!, Photo: GH!). Pellaea bella (Moore) gs So Fil. 477. spes oe var. compacta pu. Ss Judi Herb. Suppl. 46. 1883. (Holo- Wright i in 1879, Allora Ke (Moore) O. Ktze. "pu Gen. Pl. 2:8 Pellaea compacta (Davenp.) Maxon, in Proc. Biol. Soc "Wash. 30:183. 1917. (Lectotype: y nee Hort. Veitch. Chelsea, Gë KÉ EAS , CIS. y7 D AY E zn CS 3 EPA AK n Sek s SONS d SS Ny du E le Cj 1154 ee ed SS Riv ems LES g PR TK) n f P ATA TS « £> NEAN W bs eg uf TAS Ay WEEK NAS ei VAS e (| D SE Ges Wwe E p S W Da DU" KN > (2 ZS UN. BS wl -O CK DE AN = Zi? ` d 7 A S CN e H = (— eL. (2 W Cé MAC eA) Gs Eë ZA fo Sat = N tE Key, [S2 Gei LO IN. e Le, e « Sy : WN SZ Sg oH [LS WAS KW UA Fig. 6B. Pellaea mucronata var. californica, plant X V4; pinna; rhizome scale. IVoL. 44 162 ANNALS OF THE MISSOURI BOTANICAL GARDEN Fronds 9-33 cm. long. Blade 5.5 cm. long and 2.0 cm. broad to 18 cm. long and 8 cm. broad, narrowly deltoid, bipinnate, often conduplicate, acuminate. Pinnae linear, imbricate, usually closely appressed and at acute angles to the rachis, those of the mid-region as long as the basal, with few to 20 segments per pinna. Segments closely placed on pinna-rachises, usually entire. Some of the earliest collections of this variety from the San Bernardino moun- tains were made by W. G. Wright. Specimens which he sent to Eaton and Gray are accompanied by letters remarking on the characteristics of the plants which distinguish them from P. mucronata var. mucronata. His allusion to the pinnae standing back together much as the wings of a bird is a striking one and the char- acter is often apparent in herbarium specimens. The stipe may be curled or bent at the base and proportionately longer than the blade for the plants often grow from the base of boulders or among rocks. In the absence of a specimen with sufficient data to unquestionably indicate the type I am choosing the above-cited Lemmon specimen. This sheet from the Lem- mon Herbarium, collected prior to 1882, is not labeled var. californica but bears two tickets—one with locality data and the other labeled " Belleg Wrightiana var. compacta Dav." in Lemmon's handwriting. He apparently added the name after 1883. A collection of 1888 from the same locality, which has been named by Lemmon var. californica, matches the earlier one and would indicate his concept of the variety. On talus, among loose rocks or in rocky woods, open sun or shade, at 1800- 3000 m. California. pantat specimens: NITED STATES. CALIFORNIA: Abrams 2684 (GH, NY, US); Alexander 9 viet ME (UO); Dudley. 8 Lamb 4648 (F, uC); wan "9828 (US); — 1504 (UC, US); Mt. San Bernardino, Ped 30, 1876, Le on (F, GH, UC ,US) B V.P PIN srr (F, GH, US); R. M. & A. F. Tryon e (GH, k K, MEXU, MO, U, vc US, UT); San Bernardino Mts., 1879, W. G. Wright (GH 7. PELLAEA BRACHYPTERA (Moore) Baker, Syn. Fil. 477. 1874 Fig. 7. Map 7 Platyloma brachyptera Moore, in Gard. Chron. 141. 1873. (Holotype: Hort. Veitch. Chelsea, from Calif. K!, Photo: GH! US! Pellaea ornithopus var. brachyptere (Hook.) Dc. Eaton, in Bull. Torr. Bot. Club 4: dl Allosorus brachypterus (Moore) O. Ktze. Rev. Gen. Pl. 2:806. 1891. Rhizome moderately stout, elongate, decumbent, multicipital. Scales of the rhizome and base of the stipes appressed, ruddy brown, bicolorous, the central sclerotic stripe narrower than the border, often not extending into the scale ape linear, straight or tortuous, strongly dentate, the cells of the border with oblique end walls, 5-10 times longer than broad, the apex filiform, tortuous, entang Fronds 9-39 cm. long, stiff, bent or straight, monomorphic, the buds pubescent Stipe and rachis sulcate, glabrous or nearly so, glaucous, castaneous, the stipe wiry» often curled, breaking irregularly without articulation inet the rachis of the 1957] TRYON—PELLAEA SECTION PELLAEA 163 mature frond castaneous to the terminal pinna. Blade 5 cm. long and 1 cm. broad to 20 cm. long and 4 cm. broad, linear, bipinnate, rarely tripinnate, grayish green, the pinnae of nearly the same length, ascending at acute angles and appressed to the rachis, pinnate with 5—13 segments, subsessile or short stalked, the pinna-stalks not decurrent, the axils colored as the rachis, the pinna-rachises up to 2 cm. long, usually shorter than the segments. Segments 0.5 cm. long and 0.2 cm. broad to 2.0 cm. long and 0.4 cm. broad, linear, subcoriaceous, the border strongly revolute, undifferentiated, crenate, mucronate, the mucro 0.25-1.0 mm. long, flat, thick- ened, green at the base, the apex lutescent, distinct from the revolute borders of the segment. Sporangia short stalked, the annulus amber colored, glands with ceraceous indument occurring among the sporangia. Spores 64 per sporangium, Fig. 7. Pellaea brachyptera, plant X V2; pinna; rhizome scale. [ Vor. 44 164 ANNALS OF THE MISSOURI BOTANICAL GARDEN mostly tetrahedral-globose, yellow-tan, with a sparse, scarcely prominent, rugose exospore. In the search for unusual ferns for English stoves this species was brought from California by Veitch & Sons of Chelsea. The type material grown in the greenhouse in England is more robust than any of the collections from the field. Specimens at the Gray Herbarium and at Yale include some field notes of H. N. Bolander in which he observes that plants from rocks moistened by spray have larger, more numerous segments than those growing among dry rocks. From these observations on the effect of the environment D. C. Eaton concluded that P. brachyptera be given only varietal status. Bolander's collection from the Sierras at 4,000 feet, in 1869 (MO, YU) and that of Kellogg & Harford 1169 (GH, US) are difficult to place as they are inter- mediate between P. brachyptera and P. mucronata var. mucronata. These two are obviously closely related for their differences are mainly quantitative. The former has shorter pinnae, longer segments, more tortuous rhizome scales as well as a range farther north in California than P. mucronata var. mucronata. Morphologically the two entities are better defined than in the case of P. ternifolia var. ternifolia and var. Wrigbtiana. On talus slopes or basalt or serpentine rock outcrops, at 800-2000 m. California, Oregon. Representative specimens: UwiTED STATES. CALIFORNIA: Constance & Rollins 2933 (GH, MO, US); Copeland 401 (F, GH, MO, NY, P, US); Aug. 1880, Engelmann (MO); July 15, 1907, Heller 8 Kennedy (F, MO, NY, US); Ownbey & Brown 2424 (GH 2198 (MO, NY, US); Wheeler 2744 (GH, MO, NY, US). GO (Us) D jur 5, 1887, Howell (F, GH, MO, OSC, US, YU); Thompson 13066 (GH, MO; , MO, NY); Ownbey 8 Meyer 8. PELLAEA sAGITTATA (Cav.) Link, Fil. Sp. Hort. Bot. Berol. 60. 1841. Rhizome moderately stout, compact, decumbent, multicipital. Scales of the rhizome appressed and matted, tan to rust colored, concolorous, dull, elongate lanceolate-triangular, usually cordate, straight or nearly so, the margins irregularly dentate, the cells short rectangular or with oblique end-walls or elongate, the apex filiform, more or less tortuous, scales of the base of the stipes and buds appressed, tawny, concolorous, dull, broadly ovate-lanceolate, cordate or pseudopeltate, straight, the margin erose, the cells irregular, sinuous or curved, the apex filiform, tortuous. Fronds 16-78 cm. long, erect, straight, stiff, approximate, dimorphic, the sterile frond shorter and usually with larger segments than the fertile, the buds paleaceous. Stipe and rachis convex or plane on the upper surface, glabrous OF puberulous, straw colored to ruddy tan, or mottled, becoming darker with age; the stipe breaking irregularly without articulation lines, the rachis straight or somewhat flexuous, often green in the juvenile frond. Blade 10.0 cm. long and 5.5 cm. b to 46 cm. long and 32 cm. broad, elongate ovate-triangular or rhomboid, once pinnate or bipinnate, rarely tripinnate, light green, the pinnae at a broad angle to the rachis or ascending, entire or divided into 3-18 segments, long stalked, the ` pinna-rachises up to 10 cm. long, usually flexuous. Segments 0.5 cm. long and 0.5 1957] 165 TRYON—PELLAEA SECTION PELLAEA pura A9 WC emt tt Lë eme m m assess mm Ee uw Y D ^ į KÉ À ' " Keck : 7 à H D £e lx Y Lage Dag Zë $ fi 4 e Se DA "gé "m H P i ez 1 "P r D ; o ~ - em d ra Seen. - dag X ap Rm uq 1 1 E Le ; 1 ' A D j . 8A. Distribution of Pellaea sagittata var. cordata, 8B. var. sagittata, I Vor. 44 166 ANNALS OF THE MISSOURI BOTANICAL GARDEN m. broad to 5.5 cm. long and 2.75 cm. broad, rotundate-cordate to oblong- sagittate, entire, herbaceous to subcoriaceous, glabrous or puberulent on the under surface and along the margin of the upper surface, veins evident, the border white or lutescent, crenulate, the apex obtuse. Sporangia with short stalks rarely ex- ceeding one half the capsule length, the annulus occasionally irregular-oblique or with a cluster of cells at the apex of the capsule. Spores 64 or 32 per sporangium, ellipsoidal or tetrahedral-globose, yellow or light brown, the exospore with sparse, prominent rugae. The varieties of P. sagittata—cordata and sagittata—are appropriately named from the form of the segments but are often confused and the matter is further complicated by mixtures of the two under the same number as in the abundant Pringle collections. In addition to the differences in segment form they can be distinguished on several other characters, most easily seen in living plants, as noted in the discussions of the varieties. The range of the species is extended from Mexico southward to Bolivia by the apogamous variety. In the highlands of central Mexico both types occur in abun- dance and appear equally vigorous. The apogamous plants do not appear to have any advantage over the normal type except that they, of course, would not require the moisture necessary for fertilization. The two varieties usually grow a few miles apart in colonies of a few to 25 independent plants but on calcareous soils near Cuernavaca, Mexico, I have found them growing together. Preliminary cytology work indicates that the apogamous var. sagittata is a triploid with 87 chromosomes. The collections which have been cytologically examined by Dr. Britton are from Mexico. One near Morelia in the state of Michoacan collected by R. M. & A. F. Tryon 5153 had both n and 2n chromosome numbers of 87. Col- lections of var. cordata 5147 and 5149 from El Pedregal in the Distrito Federal were both reported as probably n = 29 by Dr. Britton. It seems possible that the latter and P. ovata, which has both normal and apogamous strains, are the most likely parents of P. sagittata var. sagittata. KEY To VARIETIES Segments rotundate-cordate; rachis and segment stalks glabrous; spores 64 per sporangium, tetrahedral-globose, 37-53 at broadest diameter; southe: Texas southward in Mexico to Oaxaca.___ 8A. P. sagittata var. cordata, p- 166 Segments ovate-triangular to sagittate; rachis and segment stalks usually puberulous; spores 32 per sporangium, ellipsoidal, 53-106 p at broadest diameter; Chihuahua, Mexico, southward to Bolivia. e 8B. P. sagittata var. sagittate, P: 168 8A. PELLAEA sAGITTATA var. cordata (Cav.) comb. nov. Fig 8A. Map 8^ Pteris cordate Cav. Descr. Pl. 267. 1802. (Holotype: Née, Chapultepec, Mexico MA, Isotype Alacer? E (Cav.) Pe Tent. Pterid. 153. 1836. Platyloma cordata (Cav.) J. S in Journ. Bot. ed. W. J. Hooker 4:160. 1841. Pellaea cordata (Cav.) J. Sm. Cat: Kew Ferns, 4. 1856, not Fée, 1850-52. 1957} TRYON—PELLAEA SECTION PELLAEA 167 Nothochlaena cordata (Cav.) Keyserl. Polyp. Cyath. Hari» Bung. 29. 18 Adianthum cordifolium Sessé & Moc. Pl. Nov. Hisp. 182. 1887—90; ed. 2, et? 1895. Chicbiere cordata (Cav.) Farw. in Amer. Midl. Nat. 12: Hi. Pellaea cardiomorpha Weath. in Journ. Arn. Arb. 24:309. 1943, based on Pteris cordata Cav Rachis and segment stalks glabrous. Blade rather lax in habit, the pinnae at broad angles to the rachis. Segments 0.5 cm. long and 0.5 cm. broad to 3.0 cm. long and 2.75 cm. broad, rotundate-cordate. Spores 35—53 y at greatest diameter, 64 per sporangium, mostly tetrahedral-globose with trilete commissural ridges. rhizome scale, right, stipe scale. Fig. 8A. Pellaea sagittata var. cordata, plant X oi pinna; left, [Vor. 44 168 ANNALS OF THE MISSOURI BOTANICAL GARDEN Variety cordata is distinguished from var. sagittata in the lax habit of the fronds, the pinnae at broad angles to the rachis, glabrous stipes and rachises and sporangia with 64 spores, in addition to the cordate form of the segments. Cavanilles cites Née collections from Chalma and Chapultepec, Mexico, under Pteris cordata and the holotype at Madrid is indicated by Christensen?? as the Chapultepec specimen. Two fragments, one of var. cordata and the other of var. sagittata from the Née collection, on a single sheet in the Chicago Natural History Museum herbarium are labeled Chalma & Chapultepec. The larger specimen is var. cordata from Chapultepec and the other is var. sagittata from Chalma. n open or shade, moist or dry rocks, andesitic, porphyritic, lava or limestone, at 1900—2900 m. Southwestern Texas to Oaxaca. Representative specimens: Mexico. Chapultepec, Née (F). CHIHUAHUA: LeSueur II 39 (UC, GH, US); Pringle 448 (F, GH, MEXU, MO, NY, UC, US, YU). coanuma: Wynd & Mueller 590 (GH, MO, NY, US). DISTRITO FEDERAL: Matuda 18828 (US); R. M. & A. F. Tryon 5103 (MO), 5147 (MO), 5149 (GH, MEXU, MO, US). cuanayuato: Kenoyer 1787 (US). mrmarco: Rose & Painter 6741 (GH, NY, US); R. M. & A. F. Tryon 5124 (GH, MEXU, MO, US). Mexico: Bourgeau 685 (GH, NY, P, US); Lemmon 50 (F, GH, UC, C. & E. Seler 1233 (GH, US). moretos: R. M. . F. Tryon 5121 (MO). NUEVO LEÓN: C. H. & M. T. Mueller 970 (F, GH, MEXU). oaxaca: Conzatti & Gonzdles 521 (GH), 521a (GH). PuEBLA: Arsène 9961 (US); Purpus 2712a (F, GH, , UC, US). QUERÉTARO: J. N. & J. S. Rose III (NY, US). san Luts porosi: Parry & Palmer 980 (F, MO, NY, US, YU). ZACATECAS: Lloyd 247 (US). 8B. PELLAEA SAGITTATA var. sagittata. Fig. 8B. Map 8B. Pteris sagittata Cav. Descr. Pl. 267. 1802. (Holotype: Née, Cerro de Guadeloupe, Mex- ico : Allosorus sagittatus (Cav.) Presl, Tent. Pterid. 153. 1836. Platyloma sagittata (Cav.) J. Sm. in Journ. Bot. ed. W. J. Hooker 4:160. 1841. Pellaea cordata £. sagittata (Cav.) Davenp. in Bot. Gaz. 21:261. 1896. Rachis and segment stalks usually puberulous. Blade stiff and erect in habit, the pinnae ascending at acute angles to the rachis. Segments 0.5 cm. long and 0.5 cm. broad to 5.5 cm. long and 2.0 cm. broad, ovate-triangular to oblong-sagittate- Spores 53-106 y at the greatest diameter, 32 per sporangium, mostly ellipsoidal with monolete commissural ri ge. Variety sagittata is distinguished from var. cordata in the erect, rigid habit of the fronds, the pinnae ascending at acute angles to the rachis, puberulous stipes and rachises and sporangia with 32 spores, in addition to the sagittate form of segments. The tawny scales of the base of the stipes and buds are fewer and smaller in var. sagittata than in var. cordata. The presence of pubescence, pat ticularly on the rachises, appears to be correlated with the apogamous condition. In P. ovata and P. andromedaefolia, as well as in this variety, it is a convenient "7 Dansk Bot. Ark. 9:23. 1937. 1957] TRYON- —PELLAEA SECTION PELLAEA 169 . Fig. 8B. Pellaea* sagittata var. sagittata, plant X Y2; pinna; left, rhizome scale, right, stipe scale. clue for detecting specimens with 32 spored sporangia and apparently apogamous. dry banks, in open or shade, among limestone rocks, on adobe walls or Stone fences, at 1700-3000 m. orthern Mexico to Guatemala; Colombia to Bolivia. Representative specimens: Mexico, cHiapas: Ghiesbreght 303 (F, GH, YU). CHIHUAHUA: Knobloch 5040 (GH, MSC, US). mere aie Malade 26234 (US). wmarco: Mexia 2760 (UC). . [Vor. 44 170 ANNALS OF THE MISSOURI BOTANICAL GARDEN Jausco: McVaugh 129074 (US). Mexico: Bourgeau 1163 (GH, P); Rose et al. 9464 (F, US); R. M. & A. F. Tryon 5138 (GH, MEXU, MO, US). micHoacAn: Arsène 6546 (MEXU, MO, US); R. M. & A. F. Tryon 5153 (MO). Mwonrros: Hitchcock & Stanford 7089 (UC); R. M. & A. F. Tryon 5122 (MEXU, MO). oaxaca: Conzatti 8 Gonzáles 428 (GH, MEXU, P); Pringle 4887 (GH, MEXU, MO, NY, P, UC, US). PUEBLA: Ar: séne 7027 (GH, MO, US). TLAaxcaLa: Arsène 9971 (US). saw ruis Potosi: Schaffner 059 (GH, YU). zacatecas: Rose 2705 (GH, US). UATEMALA. QUEZALTENANGO: Skutch 805 (F, GH, US). santa rosá: Heyde 8 Lux 6282 (F, GH, MO, NY, US). OLOMBIA. CAUCA: Killip 6877 (GH, NY, US); Lehmann 5710 (F, P, US). SAN- TANDER: Killip & Smith 17440 (NY, US). ECUADOR. IMBABURA: Mexia 7428 (UC, US); Wiggins 10272 (US). PICHINCHA: Asplund 6297 (US); Couthouy 39 (GH, YU); Mille 130 (GH, MO, US). Peru. apurmac: Stork et al. 10648 (F, UC UZC Herrera 1195 (F, US); Vargas 11049 (F, UC). nuaNcavrLicA: Tovar 317 (US). Lima: Macbride & Feather- stone 422 (F, GH, US). Bora, La Paz: Mandon 1554 (GH). 9. PELLAEA NOTABILIS Maxon, in Contrib. U. S. Nat. Herb. 10:500. 1908. (Holotype: Palmer 234 US!). Fig. 9. Map. 9. Rhizome moderately stout, compact, decumbent, multicipital. Scales of the rhizome appressed and matted, rust colored, the young tips light tan, these and those at the base of the stipes concolorous, subulate, straight or nearly so, the margins pectinate-serrulate, the cells short rectangular or with oblique end-walls or elongate, the apex long attenuate. Fronds 12—43 cm. long, erect, straight, stiff, approximate, monomorphic, the buds pubescent. Stipe and rachis terete or ellipti- cal, pubescent, the rachis and upper portion of the stipe straw to buff colored, the stipe base and segment stalks atropurpureous, the stipe breaking irregularly without articulation lines, the rachis straight. Blade 3.5 cm. long and 3.5 cm. broad iei 23 cm. long and 14 cm. broad, deltoid or elongate-triangular, once pinnate, grayish green, the pinnae at broad angles to the rachis or somewhat ascending, entire, short stalked, without pinnae-rachises. Segments 3.0 cm. long and 0.75 cm. broad to 7 cm. long and 1 cm. broad, elongate-lanceolate, entire, subcoriaceous, glabrous, veins obscure, the border narrow, lutescent, slightly crenulate, the apex more or less mucronate. Sporangia with short stalks not exceeding the capsule length, the annulus amber colored. Spores 64 per sporangium, tetrahedral-globose, yellow or light brown with prominent, rugose and somewhat lacy exospore. Although clearly unique, the position of this rare fern is intermediate between the dark and light stiped groups. It resembles P. atropurpurea in having COn- colorous scales, terete rachises and atropurpureous color at the base of the stipes and petiolules. Similarities to P. ovata and P. sagittata occur in the predominant straw color of the rachises and stipes, the prominent spore ornamentation and pectinate-serrulate scales. The subulate form of the scales is similar to P. ovat and both of these species have been collected in Tamaulipas. Since there are m few collections of P. notabilis I searched for it at some length in the vicinity of Monterrey—but with no success. It appears to be a rare fern possibly restric to select ecological niches which are explored with some difficulty in that region. 19571 TRYON—PELLAEA SECTION PELLAEA 171 Id Fig. 9. Pellaea notabilis, plant X V; pinna; rhizome scale. [Vor. 44 172 ANNALS OF THE MISSOURI BOTANICAL GARDEN Among rocks, at 320-1100 m. Tamaulipas and Nuevo León. XICO. NUEVO LEÓN: Sanchez 316 (US). rAMAUrLIPAs: Bartlett 10352 (US); 1902, Kemp (NY); among rocks in cañon, vicinity of Victoria, alt. about 320 m., Feb. 1- April 9, 1907, E. Palmer 234 (US); Stanford et al. 2129 (US). 10. PELLAEA PRINGLEI Davenp. in Gard. & For. 4:555. 1891. (Holotype: Pringle 2591 GH!, Isotypes: F! MO! US!, Paratype: Palmer 543 US!). Fig. 10. Map 10. Allosorus Pringlei (Davenp.) O. Ktze. Rev. Gen. Pl. 2:806. 1891. Rhizome moderately stout, compact, multicipital, decumbent. Scales of the rhizome appressed, tan to ruddy brown, these and those of the base of the stipes concolorous, the scale base often sclerotic and adherent to the stipe, acicular, straight or nearly so, the margins entire or sparsely dentate, the cells short- rectangular or elongate with oblique end-walls, the apex long-attenuate. Fron 6—47 cm. long, erect, straight or nearly so, approximate, the sterile fronds shorter, the buds covered with subulate scales. Stipe and rachis convex or plane on the Fig. 10. Pellaea Pringlei, plant X 12; pinna; rhizome scale. 1957] TRYON—PELLAEA SECTION PELLAEA 173 upper surface, or channeled, more or less paleaceous, the rachis and upper portion of the stipe straw-yellow or mottled, the stipe base darker, usually atropurpureous, the stipe breaking irregularly without articulation lines, the rachis straight. Blade 4.0 cm. long and 2.5 cm. broad to 20 cm. long and 13 cm. broad, elongate- triangular, once pinnate or bipinnate, bluish green, the pinnae ascending, entire or with 2—3 segments, usually long stalked, the pinna-rachises less than 0.5 cm. poly o. TI d E een. A. ds H 22 ( | . k Lë - n | am tte | | f 9. Distribution of Pellaea notabilis, 10. Distribution of Pellaea Pringlei, 11. Distribution o Ni [Vor. 44 174 ANNALS OF THE MISSOURI BOTANICAL GARDEN long, usually absent. Segments 0.75 cm. long and 1.0 cm. broad to 4.0 cm. long and 4.5 cm. broad, deltoid-hastate or stellate, entire, ternate or with 3—5 irregular lobes, chartaceous or herbaceous, glabrous, the veins evident, the border whitish crenulate, often opaque, bluish on the upper surface. Sporangia stalks short, one fourth the capsule length, the annulus amber colored, occasionally irregular- oblique or with a cluster of cells at the apex of the capsule. Spores 64 per sporan- gium, tetrahedral-globose, yellow or brownish, with prominent rugose-reticulate exospore. The complete specimens and ample number of collections of these pellaeas made by C. G. Pringle have made possible a better understanding of the group. From his carefully prepared specimens one receives the impression that he admired these ferns. It is appropriate that this attractive species bears his name for the original as well as the most complete suites of specimens are his. It is uncommon and few others have collected it. The stipes are largely straw colored, the pinnae star-shaped and bluish green. Some specimens are irregularly lobed rather than stellate and have incompletely formed spores. It is possible that the species may be a hybrid of P. ternifolia and P. ovata although it has distinctive acicular and largely concolorous rhizome scales. It occurs on moist limestone along the banks of streams partially within the ranges, but much more restricted, than those two species. Uncommon, on wet calcareous banks, at 1000—2200 m. Sinaloa south to Guerrero and Morelos. Representative specimens: Mexico. cotta: Reko 4846 (US). GUERRERO: Hinton 9475 (F, NY, US). J^ Lisco: Río Blanco, June-Oct. 1886, Palmer 543 (US, YU); Cool mossy banks near Gua- dalajara, Oct. 19, 1889, Pringle 2591 (F, GH, MEXU, NY, P, US, YU). MEXICO: Hinton 1537 (GH, MEXU), 4227 (GH, NY). Moretos: Pringle 7949 (F, GH, MO, Mo, US); Sanchez 97 (US). Nayarit: Mexia 709 (US). stnaLtoa: Gentry 6523 (GH, ll. PELLAEA ovata (Desv.) Weatherby, in Contr. Gray Herb. 114:34. 1936 Fig. 11. Map 11. Pteris ovata Desv. in Mém. Soc. Linn. Paris (Prodrome de la famille des Fougères) 6:301. 1827. (Holotype: Peru, Herb. Desvaux P!, Photo: GH! UC!). : Pteris flexuosa Kaulf. ex Schlect. & Cham. in Linnaea 5:614. 1830, excl. synon. (Holotype: Schiede 8 Deppe, Aug. 1828, near Jalapa, Mexico, B). Allosorus flexuosus (Kaulf. ex Schlect. & Cham.) Kze. in Linnaea 13:136. 1839. Pellaea flexuosa (Kaulf. ex Schlect. & Cham.) Link, Fil. Sp. Hort. Bot. Berol. 60. 1841. 46 Platyloma flexuosa (Kaulf. ex Schlect. & Cham.) J. Sm. in Bot. Mag. 72: Comp. 21. He Rhizome slender, cord-like, dichotomously branched, creeping or compact and enveloped by matted roots. Scales of the rhizome appressed, tan to dark brow? se and those at the base of the stipes bicolorous, elongate, lanceolate-triangulat, cordate, the sclerotic central portion lustrous (old scales sclerotic except for , narrow margin), straight or somewhat falcate, the margins pectinate to € serrulate, the cells short-rectangular or with oblique end-walls, the apex filiform, more or less tortuous. Fronds 10-125 cm. long, straight or frequently twining, 19574 TRYON—PELLAEA SECTION PELLAEA 175 subscandent, approximate, monomorphic, the buds on old portions of the rhizome as well as the apex, lanate to pubescent, with a few scales. Stipe and rachis convex or plane on the upper surface, glabrous or pubescent, straw colored to ruddy brown becoming gray with age, the stipe breaking irregularly without articulation lines, Tig. 11. Pellaea ovata, plant X V4; pinna; rhizome scale. [Vor. 44 176 ANNALS OF THE MISSOURI BOTANICAL GARDEN the rachis strongly flexuous. Blade 15 cm. long and 2 cm. broad to 84 cm. long and 30 cm. broad, elongate-triangular, bi- to quadripinnate usually tripinnate, clear or grayish green, the pinnae descending at broad angles to the rachis, divide into 5 or more than 60 segments, long stalked, the pinna-rachises up to 14 cm. pes strongly flexuous. Segments 0.5 cm. long and 0.3 cm. broad to 4.5 cm. long 5 cm. broad, ovate-triangular, sagittate or cordate, coriaceous, glabrous or get pubescent, veins obscure, the border narrow, whitish lutescent, the apex obtuse or somewhat mucronate. Sporangia with stalks as long or slightly longer than the capsule, the annulus amber colored. Spores 32, occasionally 64 per sporan- gium, ellipsoidal or tetrahedral-globose, yellow with prominent light brown, rugose exospore. These are the largest plants of the section, usually subscandent and unique in the retrorse position of the pinnae and strongly flexuous rachises. The species 1s relatively uniform throughout most of its range from central Texas to northern Argentina and appears to be the counterpart of P. ternifolia in the dark stiped group. Two species, P. intermedia and P. andromedaefolia at the northern end of its range and the Chilian P. myrtillifolia at the southern end, are related. As in P. glabella and P. sagittata the center of origin and direction of migration of the species can be determined by the spore number. In southern Texas and adjacent Mexico the specimens are s spored and produce the normal type of prothallia. A collection of R. M. & A. F. Tryon from Marble Falls, Texas, was reported by Dr. Britton to have a chromosome count of n — 29. South of this area extending to northern Argentina and in the Caribbean region the specimens are 52 spored and apogamous and evidently account for the general uniformity of the species. Subscandent, in open sun, brushy savanna, among limestone rocks, on shaded banks, in oak forest, or growing on adobe walls, stone fences or crotches of trees, at 800—2850 m Southern P southward in Mexico and Central America to Nicaragua; His- paniola; Venezuela southward to northern Argentina. E specimens: TATES. TEXAS: Correll 1 sio (MO); E. Palmer 1428 (GH, MO, US, YU); ed me (F, GH, MO, US); R. M. & A. F. Tryon 5029 (GH, K, MEXU, MO, US). MEXICO. BAJA CALIFORNIA: ger 14 5 (US). cuiaPAs: ” Ghiesbreght 204 (GH, US); 1 Little & Sharp 9907 ns CHIHUAHUA: Gentry 1538 (F, GH, MO, UC). COAHUILA: pe H & Mueller 318 (GH, M ). DISTRITO FEDERAL Lyonnet 861 ( , , US) ANGO: Ortega 5313 wie GUANAJUATO: 1880, Duges (NY). GUERRERO eg 11305 (F, GH, MO, co: Chase 7217 (F, MO) Jib ve n Wée : e 72 Pringle (F, UC), 5408 (MO, US). mexico: Hinton 7206 (F, MO); Tryon 5134 (GH, K, MEXU, MO, US). MICHOACÁN: Arsène 3645 “Cex, US) ook R (US). , MO, U N EON: C. : Mueller 1130 (F, GH). oaxaca: Conzatti M Makrinius 3030 D US); P ring 5051 (F, GH). PUEBLA: Arsène 1477 (US); Copeland 108 (MEXU, P, UC). 9 Taro: J. N. 8 J. S. Rose 11195 (US). san Luis Potosi: Schaffner 53 (P), yd E YU); R. M. & A. F. Tryon 5154 (MO). sonora: Rose et al. 13113 (US). PAs: Bartlett TN SS VERACRUZ: Bore 3067 (GH, NY, P, UC, Us, S YU) GUATEMALA: ” 1862, Godman & Salvin (BM); Hatch & Wilson 330 (UC, US); Standley f 77995 (F, US); Steyermark 47282 (F, US). 1957] TRYON—PELLAEA SECTION PELLAEA 177 HoNDURAs: — 56496 (F, US ^ Nicaracua: Standley 9721 (F), IOI Costa ee Brade 199 (NY, UC), 545 (P, US); Valerio 169 (US). HisPANIOLA. nam: Ekman H6153 (NY, US); Leonard 4804 (GH, NY, UC). DOMINICAN REPUBLIC: -e & E. S. Howard 8823 (GH, och NY, US). ENEZUEL 1341 7 (F, US); Pittier 7040 (GH, US). DISTRITO FEDER Jahn 230 (US); 1929, Vogl (S-PA, UC). mériwa: Fendler 89 (GH, MO, NY); Jahn SCH (US). CoLomBIA. VALLE: Cuatrecasas 20467 (F, US); Killip t$ Smith 16382 (GH, NY, T I9090 (F, GH, US). CUADOR. Sipe. 1918, Mille (S- ej 131 (GH, US). azuay: Haught 3346 (GH, US). CHIMBORAZO: Camp 3167 (MO, US). tmBaBura: Mexia 7404 p UC, US); Wiggins 10270 (US). PICHINCHA: SEET 6572 (US); Firmin 236 (NY, U Peru. apurimac: Herrera 1497 (F, GH); Stork & Horton 10712 (F, UC). " cuzco: Suda Pro (F 254 (US); — Pp g Vargas 1051 (GH, MO). HUÁNUCO: Livia. April 1892, Kuntze eon. Rusby 142 (NY, US). cocHABAMBA: Cardenas ou ©), DRE (US); ; Janssen 70 (S-PA). A. CATAMARCA: Oct. 1910, Castillon (GH). Jujuy: Aug. 1925, Cocker- ell (US); Sg Ze EI "Beetle 22417 (UC). TucuMAN: Borsini 13 (GH); erc 10367 (GH, MO, S-PA, US). 12. PELLAEA INTERMEDIA Mett. ex Kuhn, in Linnaea 38:84. 1869. (Holotype: "Mexico, Herb. Fournier" B?). Fig. 12. Map 12. Pellaea oo M var. pubescens Mett. ex Kuhn, loc. cit. 1869. (Holotype: Wright Field no. 7190, Label no 82 5 B, Isotypes: roe MO! Ce ER a (Mett. ex Kuhn) Farw. in Amer. Midl. Nat. 12:281. 1931. Pellaea intermedia f. E ded ex Kuhn) Broun, Ind. N. Amer. Ferns, 132. Rhizome slender, cord-like, dichotomously branched, long-creeping. Scales of the rhizome appressed, brownish, these and those of the base of the stipes bicolor- ous, elongate lanceolate-triangular, usually cordate, the sclerotic central portion usually lustrous (old scales sclerotic except for a narrow margin), straight or nearly so, the margins pectinate-serrulate, the cells short rectangular or with oblique end-walls, the apex filiform. Fronds 10 to more than 65 cm. long (largest speci- mens examined were incomplete), erect, straight and stiff, distant, monomorphic, the buds pubescent and with a few scales on the older portions of the rhizome as well as at the apex. Stipe and rachis convex or plane on the upper surface, more or less puberulent, straw colored to ruddy brown, becoming gray with age, the stipe breaking irregularly without articulation lines, the rachis straight or somewhat flexuous. Blade 8.0 cm. long and 3.5 cm. broad to 25 cm. long and 18 cm. broad, elongate-triangular, once pinnate to tripinnate, usually bipinnate, clear or grayish green, the pinnae at a broad angle to the rachis or slightly ascending, divided into 3-25 segments, long stalked, the pinna-rachises up to 10 cm. long, usually straight. Segments 0.5 cm. long and 0.3 cm. broad to 2.0 cm. long and 1.8 cm. broad, ovate to ellipsoidal, entire or ternate, coriaceous, glabrous, the veins obscure the borders narrow, whitish, the apex obtuse or somewhat mucronate. Sporangia with stalks as long or longer than the capsule, the annulus amber colored. Spores 32 = 1€ sionally 64 per sporangium, ellipsoidal or tetrahedral-globose, yellowish wi Prominent tan or light brown, rugose exospore. 1938. I Vor. 44 178 ANNALS OF THE MISSOURI BOTANICAL GARDEN The choice of the name P. intermedia by Mettenius is of interest for he con- sidered the species to be a hybrid between P. sagittata and P. ovata. This observa- tion of the occurrence of a hybrid in nature is exceptional for that time. The closest relationship of P. intermedia as expressed in the form of the rhizome and leaves appears to be with P. andromedaefolia and P. ovata. Although the type of P. intermedia was not examined the pubescent variety is not maintained. All of the collections seen were somewhat pubescent and the Fig. 12. Pellaea intermedia, plant X V4; pinna; rhizome scale. 1957] TRYON—PELLAEA SECTION PELLAEA 179 plants seem to become less so with age. Even though the type might be com- pletely glabrous the single specimen would not validate the recognition of two varieties. On shaded hillsides in oak humus or among limestone, porphyritic, or sandstone rocks, at 650-2400 m Trans-Pecos Texas to Arizona southward in Mexico to Zacatecas. Representative specimens: UNITED STATES. ‘TEXAS: rei 8260 (GH, MO, NY, US); E. J. Palmer 34078 f NY, US, YU); Aug. 16, 1895, Wooton (GH, YU). arizona: Blumer 1528 (F, G MO, NY, US); Goodding 770 (GH, US); Nov. 1, 1880, Greene (F, NY); Rothrock 494 (F, US), 674 (GH). 354 (GH. ; MO, NY US), Noxvo Cep "Metteg 1079 (GH); Taylor 84 (US). so- ue White 330 (GH), 4743 (GH). zacatecas: Lloyd & Kirkwood 22 (MO), 135 ). 13. PELLAEA ANDROMEDAEFOLIA (Kaulf.) Fée, Gen. Fil. 129. 1850-52. 3. Map 13. Fig. 1 ap Pteris iA abn Kaulf. Enum. Fil. 188. 1824. Hewes Chamisso, Calif. P!, Photo: GH! U Allosorus eege (Kaulf.) Kze. in Linnaea 9:56. 1834 Platyloma gone de ador (Kaulf.) J. Sm. in Journ. Bot. ed. W. J Hooker 4:160. 1841. Cry pteris divaricata Nutt x Hook. Sp. Fil. 2:149. A in synon FA, cedit pubescens Nutt. ex Hook. loc. cit. 1858, in Nothochlaena Are Dat ) Keyserl. Foy. fach. deu Bung. 29. Pellaea andromedaefolia var. pubescens D.C. Eaton, Ferns N. mer. 1: 205. 1878, not properly of Baker, SÉ ES 150. 1868, eie lists “6 C. pube P ven CN edaefolia var. rubens D.C. Eaton, n re. Zo o Mrs. S. P. Cooper, Santa Barbara, Ca Pellaea Gët bci in yo Fern Journ. 5:107. 1915. (Holotype: Moxley 214 ,P assebeera da Ms éi olia (Kaulf.) Farw. in Amer. Midl. Nat. 12:280. 19 931. Gerten d era oie lia var. rade Summers ex Farw. loc. cit. 281. 1931. (Holo- wW. Summers, San Luis Obispo az Calif. MICH) ^ d. b d anie andromedaefolia var. pubescens Farw. loc. cit. 281. 1931, nom. e ” a D.C. SE eilantbes pubescens Nutt."; evidently an attempted transfer EE Chet Club 6:360. 1879. Cancers andromedaefolia var. pe (D. C. Eaton) Farw. loc. cit. 281. rue thes pubescens Nutt. arw. loc. cit. 281. 1931, in synon., e an error P dis Rem ec Summers ex Farw.) Broun, Ind. N. Amer. Ferns 129. Pelee andromedaefolia f. pubescens (Eaton) Broun, loc. cit. 1938. Rhizome slender, cord-like, dichotomously branched, long-creeping, rarely Compact, Scales of the rhizome appressed, ruddy brown, these and those of a of the stipes bicolorous, elongate lanceolate-triangular, usually cordate, sclerotic central portion lustrous (old scales sclerotic except for a narrow margin), Straight or nearly so, the margins pectinate-serrulate, the cells short-rectangular or I Vor. 44 180 ANNALS OF THE MISSOURI BOTANICAL GARDEN Nu S istri- 12. Distribution of Pellaea intermedia, 13. Distribution of Pellaea andromedaefolia, 14. om bution of Pellaea rufa, 15. Distribution of Pellaea myrtillifolia. 1957] TRYON—PELLAEA SECTION PELLAEA 181 elongate with oblique end-walls, the apex filiform. Fronds 10-60 cm. long, erect, straight or nearly so, distant, rarely approximate, monomorphic, the buds palea- ceous on older portions of the rhizome as well as the apex. Stipe and rachis convex or plane on the upper surface, glabrous, pubescent or paleaceous, straw colored to ruddy brown becoming gray with age, the stipe breaking irregularly without articulation lines, the rachis straight or somewhat flexuous. Blade 6 cm. long and Ww — LI n (paz e EPA v Fig. 13. Pellaea andromedaefolia, plant X V2; pinna; rhizome scale. (Vo 44 182 ANNALS OF THE MISSOURI BOTANICAL GARDEN 3 cm. broad to 40 cm. long and 20 cm. broad, elongate-triangular, rhomboid or deltoid, bi- to quadripinnate, usually tripinnate, clear green or occasionally reddish, the pinnae usually ascending at a broad angle to the rachis, divided into 8 to more than 50 segments, long stalked, the pinna-rachises up to 14 cm. long, usually straight. Segments 0.2 cm. long and 0.1 cm. broad to 1.8 cm. long and 1.2 cm. broad, ellipsoidal or ovate, retuse, entire or ternate, herbaceous, glabrous or pubes- cent, veins evident, the borders narrow, whitish lutescent. Sporangia with short stalks usually less than one half the capsule length, the annulus amber colored. Spores 64 or 32 per sporangium, tetrahedral-globose or ellipsoidal, yellowish, with prominent tan or light brown, rugose exospore. This species is closely allied to P. intermedia, P. ovata and an African species, P. rufa. They form a unique group characterized by reddish segments and the slender, dichotomously branching rhizome. A Chilian species is related to these on the form and color of the leaves and the form of the rhizome scales. The geographic pattern of the species of this group in the Americas is similar to that of other members of the section having a Cordilleran range. The alliance with P. rufa in the Karroo of South Africa, however, forms the only apparent link between the two main geographical centers of the genus Pellaea. Most collections of P. andromedaefolia are glabrous and have 64-spored sporan- gia but plants from the southern part of the range are pubescent and have 32- spored sporangia. Both the 64 and 32 spored types have migrated to the coastal islands from the adjacent mainland. Cedrus Island off Baja California has the 32-spored plants while collections from more northerly islands, as Santa Catalina, Santa Cruz and San Clemente off the California coast, have 64 spores. The species is reported from southern Oregon on the basis of a single Howell collection from Roseburg, Douglas County. Since this collection is out of range and the spore number is not consistent with specimens from northern California the validity of the label may be questioned. 100 California and Baja California. Representative specimens. In dry, rocky ravines or ledges or along moist, shaded stream banks, at 60- 0 m. UNITED STATES. CALIFORNIA: Abrams 3121 (GH, MO, NY, P, US); Abrams E) wi Katz 30 (GH, NY, US); Abrams & Wiggins 350 (F, GH); Santa Barbara, Mrs. S. F. ooper (YU); Copeland 2750 (GH, MO, NY); Eastwood 6378 (GH, NY, US); Fosberg R83 (F, MO, ^ (F, MO); Grant 918 (F, GH, MO, US) ; Hansen 656 (MO, P, US); Heller 4998 (US), go14 (US); M. E. Jones 3227 (MO, NY, UC, US); Meyer 137% (GH, MO); Nuttall 90 (F, US), 346 (F, US); E. Palmer 427 (F, GH, MO, NY); $- W. F. Parisb 513 (F, US); i Obispo Co., Mrs. R. W. Summers (MICH); Tore? ër (GH, NY); R. M. & A. F. Tryon 5064 (MO); Wiegand &$ Upton 20544 (F, MO: MEXICO. BAJA CALIFORNIA: April 16, 1936, Epling & Stewart (NY, US); E- Palmer 730 (NY, P, US), 749 (NY, US); Wiggins 5152 (US), 9967 (US). 1957] TRYON-——PELLAEA SECTION PELLAEA Fig. 14. pope ites A, plant X 34; B, segment with vein-ends exposed; C, rhizome scale: D, s 183 I VoL. 44 184 ANNALS OF THE MISSOURI BOTANICAL GARDEN 14. PELLAEA RUFA A. F. Tryon, Ann. Mo. Bot. Gard. 42:101. 1955. (Type: Compton 16402 US!). Fig. 14. Map 14. Rhizome slender, cord-like, dichotomously branched, long creeping. Scales of the rhizome, appressed, ruddy tan, these and those of the base of the stipes con- colorous, elongate, lanceolate-triangular, cordate, the margins sparsely dentate, the cells short-rectangular or elongate with oblique end-walls, the apex filiform. Fronds 12-40 cm. long, bent or straight, approximate, monomorphic, the buds paleaceous. Stipe and rachis convex or plane on the upper surface, nearly glabrous, ruddy tan to red becoming darker and gray with age, the stipe breaking irregularly without articulation lines, the upper portion of the rachis usually flexuous. Blade 10 cm. long and 2 cm. broad to 25 cm. long and 7 cm. broad, elongate-triangular, bi- to tripinnate, the pinnae ascending at a broad angle to the rachis, divided into 6 to 15 segments, long stalked, the pinna-rachises up to 4 cm. long, somewhat flexuous. Segments 0.3 cm. long and 0.2 cm. broad to 1.0 cm. long and 0.6 cm. broad, elliptical or oval or a few ternate, retuse, coriaceous, the veins immersed and obscure, glabrous, the border lutescent, crenulate. Sporangia with short stalks less than one-fourth the capsule length, the annulus amber colored. Spores 64 per sporangium, tetrahedral-globose, pale yellowish brown, with smooth exospore. This species forms the only link between the two main geographic centers of the genus Pellaea in Africa and the Americas. It is most closely allied to P. myr- tillifolia in central Chile and to P. andromedaefolia and P. intermedia in the western United States. The Cordilleran species P. ovata is also related. A characteristic reddish color of the stipes, rachises and segments especially on immature fronds in these species is unique in the section. Relationships are also expressed in the elongate-triangular form and generally tripinnate division of the blade, in the elliptical, retuse form of the segments and elongate basally cordate scales. P. rufa has concolorous rhizome scales similar to those of P. myrtillifolia but has a cord- like, creeping rhizome similar to that of P. andromedaefolia. = The range of the species is largely within a radius of 40. miles in the vicinity of Whitehill. In crevices on rocky hillsides, at 1000 m. Cape Province, South Africa. Specimens examined: UNION oF SouTH AFRICA. CAPE PROVINCE: Rock crevices, Ngaap, Laingsburg Dist» Nov. 6, 1944, Compton 16402 (US), 2061 (K), 12619 (US); Dickson, Lady Barkly, recd. 5/75 (K); 1838, Drége (BM, K, MO, P); Rodin 3342 (UC, K, MO, US); Schelpe 4922 (MO, US), 4939 (US). 15. PELLAEA MYRTILLIFOLIA Mett. ex Kuhn in Linnaea 36:85. 1869. (Lecto- type: Péppig, Chile, Decbr. 1827, (Diar. 603) cited by Kunze, Linnaea 9:56. 1834, as Allosorus andromedaefolius B, Isotype: P!). Fig. 15. Map 15. Pellaea andromedaefolia of authors not Kaulf., for example Gay, Hist. Fis. y Pol. Chile Bot. 24. 7493. Allosorus andromedaefolius of authors not (Kaulf.) Kze., for example Sturm, Enum. 85 ant. Crypt. Vasc. Chile, 15, e Allosorus myrtillifolius (Mett. ex Kuhn) O. Ktze. Rev. Gen. Pl. 2:806. 1891. 1957] TRYON—PELLAEA SECTION PELLAEA Fig. 15. Pellaea myrtillifolia, plant X V2; pinna; rhizome scale. 185 [Vor. 44 186 ANNALS OF THE MISSOURI BOTANICAL GARDEN Rhizome moderately stout to slender, compact or somewhat creeping. Scales of the rhizome spreading, tawny, ruddy brown, these and those of the base of the stipes concolorous, elongate lanceolate-triangular, cordate, the margins pectinate- serrulate with the base irregularly lobed, the cells short rectangular or elongate with oblique end-walls, the apex filiform, long attenuate. Fronds 14—50 cm. long, bent or straight, approximate, monomorphic, the buds pubescent with a few scales or nearly glabrous, evident only at the rhizome apex. Stipe and rachis convex or plane on the upper surface, sparsely pubescent, buff colored to reddish becoming gray with age, the stipe breaking irregularly without articulation lines, the rachis straight or especially the upper portion flexuous. Blade 9 cm. long and 4 e broad to 31 cm. long and 10 cm. broad, elongate-triangular or rhomboid, tri- or quadripinnate, clear green or reddish, the pinnae ascending, divided into 1} to more than 50 segments, long stalked, the pinna-rachises up to 8 cm. long, straight or slightly flexuous. Segments 0.2 cm. long and 0.1 cm. broad to 0.8 cm. long and 0.5 cm. broad, oval or ellipsoidal, retuse, usually ternate, coriaceous, glabrous, veins obscure, the border narrow lutescent. Sporangia with short stalks less than one fourth the capsule length, the annulus amber colored. Spores 64 per sporan- gium, tetrahedral-globose, yellow with prominent brown, rugose exospore. Although endemic to Chile, the species is closely related to the group discussed under P. rufa. Geographically, it is closest to P, ovata in Peru and Argentina but morphological affinities as expressed in the structure and division of the leaves and the reddish color of the segments are closer to P. andromedaefolia in California. An infusion of the plant is used in some parts of Chile to allay pain and thus the vernacular names Coca and Yerba Coca have been applied. In rock fissures, at 1000-1400 m. Central Chile. Representative specimens: CHILE. Cumings Herb. 184 (GH, K); Gay (CU, GH). aconcacua: Bridges 557 (K). anne Pearce (K); Wagenknecht 18536 (MO, U Wagenknecht & Looser D C); ag US); 45 ). sANTIAGO: Dec. 7, 1923, Bebn (F); Claude-Joseph 632 (US), 807 (U5); July 1934, Grandjot (MO); "Les Cruzes” Páppig 84 (P). INDEX To ExsiccATAE Abrams 2379 (6A); 2684 (6B); 3727 Applegate 2606 (7). (13); 4530, 4692 (6A). Ariste-Joseph A337 (4A). fa Abrams & McGregor 30 (13); 145 (6A). Arsène 207, 530, 1477, 1620 (11); 1 5 Abrams & Wiggiris 350 (13). 1837 (4A); 1004 (3); 1905 (4A) ; 2045 (11) ; ; Alexander & Kellogg 13234 (6A); 1879 ` 3620 (8B); 3645 (11); sort, ug (13); 1003, 2070 (6A); 2333, 2409 (8A); 5496 (11); 5966, 6012 6 5, (12): 2585 (1); 2637 (6B); 2048 (6A); ^ 6645, 6546 (8B); 6567 (11); (ës 3703 (1): 3094 (6A); 3765 (6B); 4361, ^ 6690, 7027 (8B); 7055 (8A); pe 5049 (1); 5227, 5386 (7) ; 586r (1 4A); » 9950, 9957 (11); Anderson, E. G. 537. CES TERI ng SE $8); E (8A); 9962, Anderson, L. C. 648 (2A). 9963, 9064, 0065, 9966, 7> je Anderso A. (Gray Pl. Exsice, 902) 9969, 9070, 9071, (8B); 9974 Sa (2 0076, 0077, 0078, 9070, 9080, 99°1 1957] TRYON—PELLAEA 9982 (4A); a 9985 (11); 9086, 9987, 9988 (8B); 0090, 10654 (4A); 17380 (4B). M 3071, 6074 (4A); 6297 (8B); 5/2 (11). Austin I99 (13); 221 (6A); 356, 989 e 1138, 1367 (6A); 1638 (7); 1864 3). H Bartlett BER (11); 10352 (9). Basile 3 (8B). Biltmore Herb. 755, 7554, 755b, 7550, 7556, 7558, 755i (3). Bishop 7 (3 Blankenship 638 (2 (2A Blumer 1528 (12); 1860 (3); 1867 (4B); 1061 $E (3); 1970, 2105 (5); 2106 (4B); 3274 (5 "neni 2. Ce (6A). 6243 (1). 11). unas 208 (8A); 457 (4A); 685 (8A); 922, de ab; 3067 (11); 3068 (3). (6 Bowman A). Brade ; E xig (11) Brandegee 328 (1). Penes (11); 303 (4B); 371, 371a, Brenckle 51042, 51055 (3); 51126 (6A); AEN = 3). 3 (6A); Sc 2038, 2102 (1). Bridges $3 (6A); 5 7 (15). Brown, H. E Ea o n, Y. Y. 0, 2I 6); D (4B). Brown & Heller lee Bryan 198 (4A Buchtien 77 39, 2800, 9139 (4A). Burkart 4857 (4A Push 470 (2B); 549 (2B, 3); Boss 1139 (3); 1145 (2B); 2809 (3); 4035 (3); 4144 (2B); 4205, 4135, 4447, 050, P 15004, 15- i 296, 15435, 15602. (3). utters & Rosendahl 3370 (2B). Cabrera 8327 (4 (4A Camp 1053 (2B); 1369 (3); 1924 un um 2487 (11); 2659 (4A); pa ox SCH E 20467 D Cuatrecasas SECTION PELLAEA 187 DIE 3313, 4798 (11). Clark 6772 Sech Cep (4B); 9001 (12); 9200 (3); 9201 (4B); 9218 (3); 9219, 0207, 9208, 9300 (4B); 9706, 10034 (5); 12623 (3). Claude-Joseph 632, 897 (15); 2658, 5844 1881 (8A); 1892 (4A); 4830 (13); 4840, 5161 (6A); 5376, 7021 (Di 7814 (5); 7815 (1); 8248, 8257, 689 (5 Ss & Templeton 4536 (13); 4537 (6A Collins & Kempton 54 (13); 62 (6A). Constance & Rollins 2933 (7). Conzatti (8B); 1401 prd (8B); 1912 (3); 2002 (8B); 4216 74 (11); 521, 521a 605 (11); Conzatti & ‘Gonkales (8A); 428 (8B). Cook & Gilbert 254 (8B, 1D; p (8B); ; 5 (11 Cooper 416 (13); 441 (6A). eng 108 (11); 109 (4A); 401 (7); 2750 (13). Correll 13640, 1 13604 (12); 13799» 15049 (5); e (3); 15230 (1 Dd Correll, D. 8253, 82 82 (3); 12762 mo 13085 (2B); 15227 (4B). Cory 2818, 8600 (11); 9070, 9330 (4B) ; Q5 11); 29058 (4B); 30302 (12); 30330 (4A); 30466 (4B); 30467 (3); 30468 (12); 30968 (5); 31115 (4B); 31481 (3); 40197, 41134 (4B). Cottam 3751, 4321, 4537» 8487 (1). ville & een 1149 (13); 1349 (6A); I501, ville & Gilm wi 50, 71 (0; 476 (6A). Coi t tubes -9 OI 09 (3). 20829 (4A). Cumings Herb. 184 (15 188 cereis 2562 (4B); 2592a (4A); 2502b soe & Haskell 2022, 2308 (12). Davis 1529, 1621, 6071 (3); 6494, 6503 (2B) ; 6504, 6509 (3). Deam 17351, 23317 (3); 57092 (2B). Degener 9229, 10419, 14623, 14952 (4A). Wiebke 3060, 307 A). ene 3001, 3031, 326 ufa Ed (2B, 3); 10327A GC 11472, 23493 (3). Dickson 5/75 ( Drouet, et al. 3 : Dudley & Lamb 4648 (6B). Duncan 2511, 8130, 13315 (3). Duran 2811 (1). Eastham 11068, 15840 (3). Eastwood 5$ .(1 2002. (27); 3 (6A); 6378 (13); 6771 (6A); 3234 (5); 8266, 8321 (4B); $413 n II54I Sa 12563, 12807 (6A); I (5). ood & Howell r907 (13 11). 9905 (1); 10243 (2C); 10903 (4B); 12181, 12225, 16- 289, 16658, 16735, 17401, 17444, 17613, (3); 19533, 19702, 19769 (6A); 19909 Ekman 1195, H6153 (11); 11706 (4A). T 2024 (6A); 3034, 3400 (13); 3401 Evers 54 (2B). Ewan 2702, ie (6B) ; 0828, 1009 Eyerdam & eile 11 8671 (6A); ` 22163 (4A); 22417 Eyerdam; etal. _ 23403 (4A). Fellows 481 (2B). Fendler 89 (11). Fernald & Long 12513, 12514 (3). Fernald & Pease phe Mur 3035 (3). Ferre, LH. 43 Ferris, R. S. p sy, 1636 (6A); 2054 (13); 2097 (6A). Fiebrig A 38, 3260 (4A Firmin 58 (4A); 236 (11); 246 (8B). Fisher Gan 25 (8B); 33 (6A); 42 (4A); 46, 64 (11); 110 (13); 35414 (4A); 35418 (8 (8B). Fosberg R82, 64 (6A); R83 (13); R84 (4A); 676 (13); 20280, 22210, 23198 g & Giler ng (4, 4A). Fries [L2 2 Frye, T & E. i d UD. [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN Fuertes 7008 (4A). -— 6558 Cas nta 1225 (4A 1610, ec 5606, 5753, 6057 ( Gehriger 251 (4A). 1538 (11); 4283 (4B); 6434 iai 1 Gentry (4A); 6523 (10). Ciber d (11); 276 (3); 303 Gilman, F. G. 1793 (1); 2073 (6A). Gilman, M. F. 7, 8 "x ): 9 (4B); 28, 42 (13) 3 43.3) 1 45 Goodding 183 (2A); EC (5); 766 (4B); 770 (12); 977, 1052 (5); 1114 (1); t 1 & Hitchcock 1692 (1). Ge F. C. (4A). Grinnell, J. & H. W. 215 (6B). Günther 887 (8B, 11). Hansen fu (13); 657 (6A). Happ 71i (8B). Hinr 535 353, 2034, 2231 (3). Harrison 7734, 7735, 7780 (5). 3 dud ds EI n et 2B 2 D ES B sea nd ress " Hatch & Wilson 249 (8 Ce 290 (4A); 291, 293, 330 (11). Hough: 3346 (11); 5946 (4A); 6053 664 5 we Heinrichs 7 (4A). Heller 717 6: 4908 (13); 7816 (6A) (13); vi, 9548, 11103 (1); ib orp en 475 (4A). inckley Bel (8A); 385, 386 (4B); 514 “G2 ); 616 (3); ro10 (8A); 2143 2 2585 (5); 2721 (4B); 2732 (12); 3 (8A); 3453 (12). 1957] TRYON—PELLAEA Hinckley & Warnock 46871 4B). 1); 1537, 1538 (10); (10); 10490 (4A); 11305 (11); 15497 (4A); 16700 H 2). Hitchcock, A. 14240, 14481, 15560 (4A); 30332 (10); 22550 (4A Hitchcock, CL 16235, 18029 (2A). Hitchcock & Martin 5350 (7); 5589, 5708 Hitchcock & Muhlick 8855, rooro, 10378, 10589, 11004, 11354 (1); e (2A). Hitchcock & Stanford 7098, 7181 (8B). Beta 971 (11); 1467 (4A); 1973 NE 1563 (6B); 4230 (13); 4396 (6A). Horr E459 (3); E463 (2 p. Hor r & McGregor & or Howard ES. Sa (4A). 3 (11); 9137 Howell, A. H. 4 : Howell, J. ei e Oo 1). Howell, T. 1329 (7); 2070 (6A). 4102, 10208, "10554, 11932 (3); 15862 pn) 17730, 42012 (3). Hunziker 5278 (4A). pee 2733 (4B); 2734 (8A nnes & Warnock 559 CN » ig (6B); ro er 476 (3); 642, 859, AS MC UE 1309 (4B) ; 9040 (3). M. E. , T 2B); 3227 (13); 4807 (4A, (5); 6633 (1); 24556 ( (3); 260 26045 (5); 27013 3 (4A Jorgensen 1251 (4A). uen (6A); Kearney & Peebles 108 6 (5 Kick et ab pene : 55, IIIIÓ (5). 305 Kellogg 108 (2B); 808, 1636 (3). sg & Harford 1170 eg, 1171 (13). (2 Kennedy & Goodding 757 (5 Kenoyer 1787 (8 A). GC Os SECTION PELLAEA 189 Killip 6877 (8B); 6882 (4A); 36314 Killip & bp AI (11); Ea 17357 i, o (8B, 11); (4A); a di 21817 (4A). Kimball 7240a (13). Knobloch 24, 35 (4A); 1051 (8A); 1178, (4A); 5052 (8B); 5092 (4A); 5940, 5983 (8B); 6002 (4A Knopf 124 (13); 282 (6A). Knowlton 257 (5). Koristan & Baker 29082 (2C). Kurtz 8346 (4A). > edness is 3 (11); 200 (4A). eavenworth & Hoogstraal 1234 (8B). Lacas 54, 52 3 (1 da Lehmann 5045, 5047 (4A); 5710 (8B). Leiberg 3153 Foe fs (13); 5353 (1)- Lemmon 50 (BA); 145, 333 (11)- Leonard, E. C. 1986 (3); 4450, 4584, 4804 (11). LeSueur 1139 AAA 2020 (4A). ‘Lewis 685 (6A); $). Lindheimer pr aed o (11). Little, EL 2212 (12); et (4B). Little & Cas 9907 (11 Little, R. R. 7 (4A). Lloyd, F. E. or RH as C. E.) 22 (12); 247 (8A); 495 (3 oyd & S 22, 135 (12). Lorentz 10,95 (4A). 14, A ) Lyonnet 26 (4A); 227 (8A); 245 (8B); 246 (8A); 86r (11); 2563 (10). Macbride 3142, 3691 (11 Macbride & Drouet 4323, 4540 (6A). Macbride oe 286 (4A); 422 (8B); 1778 (4A). Macbride x fre 3433, 3758 (1). Maguire 11070 (5): 13024 (1): 16317 $3. pe B. & R. R. 12027 (4B); 12080 yi Maguire & Holmgren 22078, 22427, 25- (1). " Maguire et al. 17653 (3); 13153; 14136, i: 262 (4A). M ; 251 (8A); 258, 680 (12); 1888 (3). Mason 3778 (6B). 190 Matuda 18828, 19622 (8A); 26234 (8B); 26520 (4A); 26719 (8B); 26737, 26823 (4A); 27204 (8B); 27687 (8A); IIQIOI (4A). Maxon e 547, 6393 (3). McCall 2432 (2A). McDougal 54 (2C). McGregor & Horr 1311, 3299, 3301 (3). McIntosh 201,905 (2A). McKeever 3 (6A). McKelvey 71858 (4B); 2132 (5); 5053 McMurphy 444 (13); 445 (6A). McVaugh 7451 EM 7922 (12); 12974 (8B); 12908 (4 Mearns 25 ve 219 (5); 236 (12); 1745, rg 2190 (3); 2604 (5); 3493, 3791 Metcalfe eu ay 1457 (3). Meyer, F. G. Meyer, F. G. & di E. 2296, 2436 (1). Meyer & Rogers 2636 ( Mexia 709 (10); ES AS 2760 (8B); 7403 ipa 7404, 7426 (11); 7428 Mile 128 (4A); 130 (8B); 131 (11). Millspaugh 4407 ( ass 4559 (13); 4655 (6A); 4677 ( Moore, J. A. S (IT 3018 (3); Lee (4B); 3580 (3 Moore, J. W. & M. F. 10617, 13563 (2B). Morus m A). Moxley 214 (13); 228, 400 (6A); 936 Mueller, C. H. $260 (12); $265 (4B); 8266 (3). Mueller, C. H. , : W. 14625 (13). Man 5734, 6156 (6B); 6634 ssh eng 6B); S070, 12393 (6A); 12399 (1); 13048 (6A); 13723 (6B); 14185, 14849 band 17070 (6B). nz, et al. 4018 (6A). Mutis 3103 (4A). Neos, A IQIQ (2A). Nelso N Xu ie o "xt S 4 Norton 964 (2B diei Orcutt 1057 (12); 3485 ( (8A); 3637, 3827 (4A); de (8B); 6oro (11). Ortega 5313 (11 Over 1680, 13841 (3); 17604 (2A). [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN Ownbey, M. & R. P. 1878 (1). Ownbey & Brown 2424 (7). Ownbey & Meyer 2108 (7). Pachano 114 (4A). P 100 (6A); 234 (9); 269 (5); 31974A (4B); 319 ; 32106 (4B); 32857 (3); 32895 (2B); 33436 (4B); 34078 Q2 34165 (4A); 34394 (4B); irt 35556, 35647 (3); Lë 019 (3); 37126 99 (2B); 39276, y (t yi Palmer & Steyermark 40681 (2B); deed 41405 (3); 41409 (2B); 41638, 41075 (3). Parish, S. B. 277 (3); 512 (6A); 4697, 4744 (13); 4770, 5050 (6B); 6337 (13); 10308 (6A Parish, S. B. & W. F. ont (6B); 512 (6A); 513 (13): 11948 (6B). PEH 94, 1659 (1). ry & Palmer ọ80 (6A); 982 (4A). & L. B. Mei E. B 2708, 3066, 3475 49 (1) Pedersen 157 (4A). ee E D AD TA Aa 1: 3961 (3 8 E Peebles, p al. 2 (3); ps (5); e H (12); 2G » Penilea Sch on 507 (6A). Penland 765 (4A Pennell 2548, 13075 (4A); 16054 Ca T7151, 17166 (3); 18258 (4A); 197 2600 (5); H 27554 000 (4A). : illips & Darrow 3016 (12); 3018 (4B); 021 (3). Phillips & Reynolds 2943 (4B). 2565 (4A). 7040 (11). Pollard & ER SÉ 392, 417 (3)- Póppig $4 (15 1957] TRYON—PELLAEA Porter, E js MAT "B. (8A); 26r (12); Porter Eel ay ST (Di 4497 (2A). 667 (2A). 256 (8A, 8B) ; 15711 (10 Pultz & Phillips 2724, 2732 (5). Purer 7148 (13 Purpus 559 (4A); II52 (11); 1582 (8A, 8B, 11); 2712, pe pies 4032 (4A); 4035 (11); 7248 (8B Raup 6558 (3). Redfield 476, S us (3). Reed 976 (6B). Reko Reverchon uc AE 5225 (4A). 5 (3); 1216 (4B); 1628 )- ose, J. N. 325 (1); 2792 (4A); 2795 EM 3403 can. se, J. N. & J. S. 5 (8A); ed NO 11195 (11); ec? 11509 t Rose 6464 (8A); 6492 (4A); 6809, 6825 (1 Ye e et al. 9464 (8B) ; 9466 (8A) ; IOI E 9 3 o7 (11); 22872 (4A). E / S. 34524 (13); 41083 (6A); 70 (2B SS i; 494, 674 (12). 142 E 323 (4A); ; 10423 (5). IIQI 1 (2). di e erg & Garrett 9421 (2C). SE 664, Si 995 ds ée? (6A). 2 (11 seem P si 7 (10 = es (8B); 613 (10); 3 ied ^5 Aus Miis 2). ner 14 (4A); 53, 89 (11); 90 (8A, 8B); 9I (8A); 114, 150, 321, 924 (4A); 958 (11); 959 (8B). SECTION PELLAEA 191 SR 4922 Seaton 39 (11); e (4A). Seifriz 471 (4A). Seler, C. & E. 1233 (8A). essé, et al. 3850 (11); 3927 (4A). J. 1683 (3); 4575 (2B); 44445 Shepard. 72 (4A). Shinners 2380 (2B). Skutch 805 (8B); 1596 (3). 7 (12). we ^). Sperry 179 (2); 183 (4A); T415 (8A); T735 (4B). Standley 5362, 8767, 9231 (3); 9721, 4 (11). Stanford, et al. 76I, GH (12); 301 (3); 576 (12); 943 (3); 2129 (9). Steere apo t A). eege 06 II ADI ); III Stewart xe 2, 1005, II ; (12); 1467 (11); 1698, Zeck 1813 (3); 1821 (12); 2475 (3) 2525 (8A). SE 1404 (2B); 1836, 4536 (3); (2B) ; 20336, 20377, 20 46 SH & Horton 9997 (4A); 10712 (11); 10720, arr s db Stork, et al. T AU à 256 (13); f (2A); 2083 Taylor, T. M. C. & Lewis 172 (2C). 192 Taylor, T. M. C. et = 41, I15 (2B). Thompson, J. W ZE Geen, — (1 mg (7); 1 13336, pe 14023, 14063 iw A (2 Td Sue OW .M. 362 (2C). Lei e Mosischer 7685 (3). Thurber 614 (13). Tidestrom 731 (4B); 2284 (2C); 3142 Topping 3634, 3833 (4A). Torrey 595 on 595a (13). Tovar 317 (8B Tracy; J. P. 4328, I5516 (7); 15812 (6A); vd Rd ). rask 13); Trelease QI A oeh 7130 (3). syn, 1 5000 AS 5028 5147, 5149 ed, 5150, S151. (4A); R. M. & P. F. 4348, 4350, 4474 ; 4698 (3); 46083 (2B). Tweedy 20 (1); 110, 2653, 3546 (2A). Ulke gor (2A). Underwood róro (1). Valerio 169 (11). Vargas 1050 (8B); 1051 (11); 1059, 3152 (4A); 8774 (11); 11049 (8B). Vasey 554 (4B); 679 (6A); 685 (13). [ VoL. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN Venturi 1246 (11); 1658, 3197, 4187, 5918, 9983 (4A); 10367 (11); 10428 (4A). oras e 18536 (15 nknecht & Looser 9 (12); Bes gë: 4 (3); Wage Wa MER e ÓI SC Waterfall 4568, 4572 (5): 4924 (12); 6196 (3); 6456 (4B); 6695 (12); 6744 6746 (4B). (7) (39€. 330 Ge 4221 (4A); 4743 Whitehouse 7827 (4B). Whiting 752 (8A). odas (13); 2955 (6A); 2956a (4B). os 3152 (13); 6533A (5); 75494, 9054, (6A); 13); 10270 (11); paf (8B); 10364 (4A); 11185 (1). Wiggins & Demaree 4721 (6A); 5024 (4B). Wiggins & Ferris. 9355 (13 viene & Gillespie 4151 ps Wiggins & Rollins 492 (1) Wilcox 127,256 (4B). Williams, L. 2532 (2A); 14301 (4A)- Williams, L. & Pierson 34 (1 Williams, R. S. 241 (2A); 2632, 2033 Wilson, E. 1087 (2C). coo 12). Wright des ur 12); 1190 (12); 213? (4B) ; 2 Mns (5). Wynd & Mueller 378 (11); 358. G): 574 (4A); 590 (8A). INDEX TO NAMES Accepted names are in Roman type; synonyms and rejected names in italic. m COVE OUR Lo ues i a IE. trhoinm S S S Rem PN Allosorus éiere heien E eg A RT atropurpureu. Ratt E 144 bellus E OR um Ge GE y] wen uuu E ME Jette oso o s vc de ide + m fessi — 05 1 V on 158 mucronat s eR 184 yrlibfolist — = —— 158 Ormitbobet . 2. T 172 Panela oo e 46 ittal E i 1957] TRYON—PELLAEA SECTION PELLAEA ternifolia _ Shc eae v var. dee bubesc EE rdat atropurpur E f. bif var. Liane EE var. californica var. com c MICI SRM var. longimucronata ——— [77 Ee eg MISSOURI BOTANICAL GARDEN STAFF - Acting Director Heng C. Cu EDGAR ANDERSON, GEORGE e Geo SCHAACK, Curator of Useful Plants Acting Curator of Herbarium CARROLL W. DODGE : ? ULIAN A, STEYERMARK, Mycologist J Honorary Research Associate RonERT E. Woopson Jr : : on. FREDERICK G. MEYER, Senior Taxonomist Dendroloniat Henry N. ANDREWS, Paleobotanist AricE F. TRYON, . esearch Associate Rorta M. Try ; Assistant Soret of the Jonn D. Dees, Herbarium ` Research Associate Triron VON ScHRE sistant CUm M t Museum of NELL C. HORNER, : Economic Plan Bibliographer and Editor Oscar E. GLAESSNER, Controller BOARD OF TRUSTEES President Jonn S. Vice-President / Dant K. CATLIN SEN Second Vice-President n o Eesseg Perrys : ge Leicester B. Faust J. Lockwoop | Duprey FRENCH Henry B. PrrtacER — A. WESSEL SHAPLEIGH RORERT BROOKINGS SMITH | : |. EX-OFFICIO MEMBERS _ UR C. LICHTENBERGER, e SraaTFORD ee UE Mem Die SS Science of St SE tous Volume XLIV Number 3 Annals of the Missouri Botanical Garden SEPTEMBER, 1957 ` A Revisio ‘of Gentianella ia Moench. Besibn of the Noris American Specie ot por red 195-269 Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the Missouri Botanical Garden and the Henry Shaw School of Botany of Washington University in affiliation with the Missouri Botancal Garden. Information The ANNALs oF THE Missouri BOTANICAL GARDEN appears four times during the calendar year: Macr May, September, and November. Four numbers constitute a volume. Beginning with Vol. 45, 1958: Subscription Price $12.00 per volume Single Numbers nn 3.00 each : Contents of previous issues of the ANNALS OF THE Missouri BOTANICAL GARDEN are listed in the Agricultural Index, published by the H. W. Wilson Company. . Annals of the Missouri Botanical Garden Vol. 44 SEPTEMBER, 1957 No. 3 A REVISION OF THE NORTH AMERICAN SPECIES OF GENTIANELLA MOENCH ^? JOHN M. GILLETT When the genus Gentianella was segregated from the Linnaean Gentiana by Moench? in 1794 it contained but one species, Gentianella tetrandra Moench (= Gentiana campestris L.) one of the amarella group of gentians. In 1796 Borck- usen* published independently his own genus Gentianella to include the fringed Bentians. These attempts at segregation were generally ignored, however, until 1894 when Kusnezow^ brought together several of the present sections to con- stitute his subgenus GENTIANELLA of inclusive Gentiana, but without revision of the included species. Since then, only in occasional floras has Gentianella been returned to generic rank with the name usually credited to Borckhausen rather than to Moench. Thus the fundamental taxonomy of Gentianella has lain more or less dormant. It is astonishing that such showy and rather familiar plants should have been neglected until now. At the present time, particularly in the Americas, very few tanists even give recognition to the genus. This situation probably is due in Part to the reluctance of taxonomists to undertake the many nomenclatural trans- fers required. Perhaps also there has been a tendency to avoid a genus with an obviously extreme degree of variability and with a range so wide that a regional study is difficult. Nevertheless this study, which includes but three of the possibly eight sections of the genus, has been attempted in an effort both better to indicate relationships and to bring order to the nomenclature. ————— for the degree of Doctor of Philosophy. : : tribution No. 1518, inas éi Plant Pathology Laboratory, Science Service, Canada Depart- ment of Agriculture, Ottawa, Ontario. 4 Moench, C., Meth. Pl. 482. 1794. : : i Borckhausen, M. B., in Roemer’s Archiv für Botanik 1:23-32. 1796. 8 Kusnezow, N. J., in Acta Hort. Petrop. 15:1-507. 1896-1904. cf. Schustler, Fr., in Vestnik Sjezdu Cesko Bot. v. Praze, 32-34. 1923. 7. in Issued October 9, 195 (195) [Vor. 44 196 ANNALS OF THE MISSOURI BOTANICAL GARDEN Although a natural classification ideally should be based on a study of a genus in its entirety, this is not always feasible. For practical reasons it has been necessary to select regions that provide abundant study material and opportunities to study the plants at first hand in the field. However, in order to maintain perspective, representative material for the remainder of the genus throughout its range has been examined where it has been readily available. While Gentianella is not as large a genus as Gentiana (sensu stricto), neverthe- less the total number of described species would have been unwieldy for the time available. Gilg, in his monograph,’ described 182 South American species.? The number of species within the genus is not definitely established because many of those described still are included within Gentiana. Too often their descriptions are inadequate, so that a decision regarding their generic position is impossible without authentic material. : For this study nearly 4000 specimens have been examined from the major her- baria of North America and Europe. In addition, opportunity was afforded to study and to collect Canadian and northern United States species in the field. This study has been based primarily on standard herbarium techniques and on field observations in natural habitats in the provinces of Alberta, British Columbia, Manitoba, Ontario, Quebec, Saskatchewan, in the Yukon Territory, and in the states of Idaho, Indiana and Wyoming. In general, seeds of North American material have been difficult to procure, and plants were difficult to maintain in cultivation. Seed exchange provides material frequently of dubious origin or consists primarily of European and Asian perennial species of Gentiana. More definitive studies of Gentianella in the future will require better facilities for observation under cultivation. GENERIC PosrrioN Gilg,® who made the most recent and comprehensive survey of the family, separated the Gentianaceae!Ü into two subfamilies, the Gentianoideae and the Menyanthoideae. His key to the subfamilies and tribes is based primarily on pollen characters. The morphological and anatomical differences between these two subfamilies are evidently considerable as well, and although no phylogenetic relationship with other groups has been proposed it seems possible that these are two separate, perhaps scarcely related, families. Gilg further divided his Gentianoideae into five tribes: the Gentianeae, Rusby- antheae, Helieae, Voyrieae, and Leiphaimeae. 'The Gentianeae, in which we are primarily interested, he divided into the subtribes Exacinae, Erythraeinae, Chironi- inae, Gentianinae and Tachiinae. , Gilg, E., in Engler's Bot. Lee 54 (Beibl. 118):4—89. 1916. nly one species of true Gentiana (G. trat i ? Gilg, E., in Engl. & Prantl, Nat. i Lie: ù arri den biu Pflan . ed. 1, 42; 10 Dumortier, B. C., Analyse Fam, Pl. 20, 1829. E 1957] GILLETT—REVISION OF GENTIANELLA 197 Within the subtribe Gentianinae, Gilg recognized seven genera: Crawfurdia, Jaeschkea, Gentiana (sensu lato), Ixanthus, Pleurogyne (= Lomotogonium), Swertia, and Halenia. Since that time, as indicated in the historical survey w ich follows, Crawfurdia has been united in part with Gentiana and in part with Gentianella. Gentianella is more closely related to Swertia, Lomatogonium, and Halenia than to Gentiana. Two groups can clearly be distinguished on the basis of the position of nectaries: Gentiana, with glands at the base of the ovary; Gentianella, Swertia, Lomatogonium, and Halenia, with glands, foveae, or spurs borne on the surface of the corolla alternate with the stamens. The corolla of Gentiana, too, bears plicae or folds between the lobes, and the calyx has a membrane or rim extending com- pletely around the interior of the tube. The calyx lobes, which arise from im- mediately below the edge of this membranaceous rim, may be foliaceous or they may be reduced to teeth. Kusnezow! has shown that this type of calyx has developed as an invagination of the inner epidermis of the tube. The intra- calycular membrane is not consistently present in species of Gentiana; for example, G. douglasiana Bongard, a west-coast species, bears no membrane, but does exhibit corolla plicae accompanied by glands at the base of the ovary. In Gentianella and in the other three related genera, no such membrane exists. The small membrane that extends across the interior of the sinuses of the calyx in subgenus EUBLEPHIS (the fringed gentians) appears to be of different origin, for it is never continuous around the interior of the tube but is restricted to the sinus, and extends towards the center of the outer lobes only. _ Again, in Gentiana, the corolla lobes have 3 principal vascular bundles, while in Gentianella there are 5-9. I have seen too few species of the other genera in the group to comment on this character. Borodin’? reported calcium oxalate crystals in the leaf mesophyll of all the species of Gentiana (subgenus EUGENTI- ANA) that he examined but did not find them in the species of Kusnezow's subgenus GENTIANELLA. Although not all species have been examined cytologically, and such informa- i d interspecific relationships, some ussed. It seems evident that Gentiana a development series of glands to fovae, fovae accompanie development of the squamellae themselves, ated by the development of spurs. Allen Haleniae are more primitive than the spurred ones, n SMOD M Kusnezow, loc. ci , loc. cit. 1896-1904. S i, Borodin, J., in Trav. Soc. Imp. Nat. St. Petersb. 22:131-137. 1892. (Russian). Allen, C., in Ann. Mo. Bot. Gard. 20:135. 1933. [Vor. 44 198 ANNALS OF THE MISSOURI BOTANICAL GARDEN reasonable. Swertia was shown by St. John" to be divisible into two groups, those bearing a single fovea on each petal, and those bearing two. In his revision, he used this difference as a major key character, although it appears to be of such rime importance as to suggest two separate evolutionary lines. Lomatogonium shows considerable affinity to Swertia because of the rotate corolla and the presence of squamellae. Chromosome data available have contributed little towards the problem of generic delimitation. Table 1 provides calculated base numbers, gametic — zygotic numbers and probable level of ploidy for a number of selected species.’”?* From these data we glean that more than one series may exist within a genus; that the majority of the true gentians (genus Gentiana) of North America are diploids on a base of 13 (§ PNEUMONANTHE and § CHONDROPHYLLA), that the European $ APTERA has base numbers of 7 and 13, that the Old World § THyLAactTEs has a base of 9, and that $ cvcrosriGMA has a base of 7; that the genus Gentianella has a base 9 in both series AMARELLA and series ARCTOPHILAE of § AMARELLA, and that subgenus EUBLEPHIs (— § CROSSOPETALUM) has two series each of 11 and 13. The diploid subgenus comastoma has a base of 5 in common with the genus Lomatogonium. * St. John, H., in Amer. Midl. Nat. 26:1—29. 1941. 15 Rork, C. in AM x dees tor 36:687—701. 1949. ie 16 Favarger, e. Bot. Suisse 59:62—86. 1949; 62:244—257. 1952. Y Lóve, D., in ra 39: 225—235. 53. TABLE I* Calculated Probable ase Gametic | Zygotic level of Author and year number number | number ploidy Gentiana COELANTHE G. lutea Sor7 | ca. 17-18 ord 1935 21 42 Stolt, 1921 20 40 Favarger, 1949, 1952 G. purpurea 5 20 Favarger, 1949 soos desi s 13 13 26 Dipl. Rork, 1 G. ris 11 22 Dipl. "ech? ger 44 Rork, G. cherokeensis 13 13 26 Dipl Rork, 1949 G. clausa 13 13 26 Dipl Rork, 1949 G. decora. 13 13 26 Dipl. Rork, 1949 G. gebleri 13 13 Dipl Sokolovskaja : à trelkova, 1938 G. linearis 13 13 26 Dipl Rork, 1949 G. maki 13 26 Dipl. Sakai, 1934 G. pneumonantbe 13 13 Dipl. Scheerer, 1939 G. porpbyri 13 13 26 Dipl. Rork, 1949 G. puberula 13 13 Dipl. Rork, 1949 Son 13 13 26 Dipl. Rork, 1949 os eqs "- 13 13 26 Dipl. Rork, 1949 G. frigida 13 13 Dipl. Sokol. & Strel., 1938. 1957] GILLETT—REVISION OF GENTIANELLA TABLE I (Continued) 199 Calculated Probable base Gametic | Zygotic level of Author and year number | number | number ploidy APTERA G. cruciata 13 26 52 Tetra. Lii 1949 52 Rork, 1949 G. daburica 13 13 26 Dipl. Rork, 1949 G. gracilipes 13 13 26 Dipl. Rork, 1949 G. macrophylla 7 21 42 Hexa. Rork, 1949 d bblogifolia 13 52 Tetra. Rork, 1949 G. straminea 13 52 Tetra. Rork, 1949 CHONDROPHYLLA G. altaic 13 13 Sokol. & Strel., 1938 THYLACITES G. acaulis 9 36 Tetra , 194 G. alpina 9 18 Tetra. Favarger, 1949 G. clusii 9 18 Tetra avarger, 1949 wk na 9 18 Tetra Favarger, 1949 CYCLOSTIGMA G. nivalis dÉ ipl Favarger, 1949 G. verna 7 14 Tetra Favarger, 1949 ianella ‘ibe GENTIANELLA ect. AMARELLA AM "e. endet 9 36 Tetra. Favarger, 1949, 1952 Skalinska, 1952 36 öve, G. campestris 9 18 36 Tetra. ri iio rni ve, ser. tbls Sg rea pis 9 36 Tetra. Rork, Zë Lóve, 1953 i G. quinquefolia 9 18 36 Tetra. Rork, 1 Subg. EUBLEPHI G. Sege 11 22 sale cb api 1949 G. detonsa 11 44 etra , G. crinita 13 78 a pr 1 up G. 9 Lie z ca. 80 Denniston, 1913 subg, os : enella 5 10 Dipl. Sot i abe 10 Love, 1953 Lomatogonium Se L. rotatum 5 5 10 Dipl. Lóve, 1953 Werlig Sakai, 1940 ipee So 9 18 Ae S. berenn 12 Woycicki, 1937 Zem Rork, 1949 F. caroliniensis 13 39 78 Hexa ork, OMG 1. EE retorted of the level of ploidy is, in many cases, that of the a information uthor. For more complete IVor. 44 200 ANNALS OF THE MISSOURI BOTANICAL GARDEN Tetraploidy seems to be the rule in North American subgenus GENTIANELLA, and tetraploidy and hexaploidy the rule in subgenus EUBLEPHIS. The diploid sub- genus COMASTOMA is left provisionally in Gentianella. Of the seventy or more genera of Gentianaceae only the herbaceous members of the north temperate zone seem to have been investigated cytologically and these in a general manner only; ie., selected species rather than all species. Probably investigations have been dependent upon the availability of seed. Gentiana, then should be restricted to include those plants with glands at the base of the ovary and plicae between the lobes of the corolla, whether or not an intracalycular membrane is present. Halenia, with its spurs and dextrorse convolu- tion of the corolla lobes in the bud, is evidently closely related to Gentianella subgenus GENTIANELLA. Late-flowering plants of Halenia deflexa occasionally form spurless flowers that are nearly indistinguishable from Gentianella amarella. Nevertheless, in spite of these occasional irregularities, Halenia appears to be quite a distinct group worthy of generic rank. Subgenus EUBLEPHIS of Gentianella, on the other hand, while distinct from subgenus GENTIANELLA in ovule arrangement, constant tetramery, calyx formation, papillose seeds, etc., has been retained at the subgeneric rank to indicate close relationship with the amarellas and for reasons of usage. I should again emphasize that this study of the North American species of Gentianella includes but three of the eight sections assigned to this genus (other sections include: DAsvsTEPHANA Griseb. (not Adans.), ANDICOLA Griseb., IMAI- cota Griseb., ANTARCTOPHILA Griseb., and PrERYcocALvyx Maxim.), and but a minor number of the total species involved. The segregation of Gentianella roughly divides the gentians into two phylo- genetically distinct and quite convenient units. To separate further the fringed group as did Yu-chuan Ma,! chiefly on the ovule arrangement, sets a precedent that might make it necessary at a later date to segregate sections of Gentiana on this character. Finally we should achieve a larger number of segregates, a larger num- ber of nomenclatural changes, and add very little more to an understanding of the group. This understanding can best be expressed by raising or lowering the r of the sections without unduly disturbing the generic concepts. Within Gentian- ella it is sufficient to show that the fringed gentians are distinct from the amarellas but no more so than is G. tenella, and that the amarellas and arctophilas are far more closely related. In order to illustrate the position of Gentianella with respect to the other gener? of the Gentianinae, I have prepared a key to the genera of the subtribe, including the genus Megacodon H. Smith of which unfortunately no material was available to me. In this instance characters employed in the key are taken from Hemsley's original description and from his extremely fine plate. 15 Ma, Yu-chuan, in Acta Phytotax. 1:1-19. 1951 1957] GILLETT—REVISION OF GENTIANELLA 201 KEY TO THE GENERA OF THE GENTIANINAE Shed et: of our glands at the base of the ovary, never appearing upon the base of the corolla t b. Coro tle without interlobal plicae; calyx tubular without an intracalycular membrane, or the sepals f » Ovary Soa a a bilocular condition by a lamellate intrusion of the placentae into the locular ph i sepals united Ixanthus cc. Ovary dioe ular o e» restricted to the sutures and with an additional row between them; Megacodon bb. Corolla vii intsriobil pics: calyx tubular, generally with an intracalycular mem- rane, or spathifo: Gentiana aa, Whorl of nectaries upon the base of the corolla tube da with the stamens. d. Corolla tube spurred, foveate, or ere ge at = e. Stigmas fnis nt along the carpel sutur Lomatogonium e. Stigmas f. Co La Tn sinistrorsely E in the bud, spurless, with one or two fimbriate foveae at the base of the ertia"? ff. Geile Ze me usb id convo vat in ven bud, each with a non-fimbriate spur : at of de corolla tube, or rarely spurless Halenia dd. Corolla Lexa mou glands at the dise between the stamens, never spurred, foveate, nor mbriate at the g. prameny eot. at the middle of the corolla tube or below; corolla imbricate in iem bu ds generally many in eac capsule entianella gg. Stamens Gre at the sinus of the corolla lobes; corolla valvate in the bud; dn few ach cipale. See Jaeschkea HISTORICAL SURVEY The genus Gentianella is a segregate of the much larger Gentiana; hence their histories are closely interwoven. Although rather well treated by Kusnezow?" in the monograph of his subgenus EUGENTIANA, a review of the history up to modern times and with stress on that of Gentianella will assist in an understanding of the generic problems existing in this portion of the Gentianaceae today. Gentiana is generally accredited to Tournefort?! who used a Dioscoridean?? nàme commemorating legendary King Gentius of Illyria. Ruppius?? recognized two genera, Gentiana and Gentianella, both of which are included within the modern Gentiana (sensu lato). In his ‘Genera Plantarum’ of 1737, Linnaeus took up Tournefort’s Gentiana, dividing the genus into seven groups, some of which correspond to contemporary sections. Later, in his ‘Species Plantarum’ of 1753, he redivided Gentiana less successfully into three groups as follows: 1 EE G. p ue G. pneumonant die? ge rg G. villos S deg 2. +*+ : KE vi? ; ormibus includes PPS Corollis D wenn t G. spicata, a quinquefolia, G. G. nivalis, G. aquatica, G. ut G. centauria, es s ` ili G iata 3. ***Corollis non es dE G: campestris, G. ciliata, G. cruciate, G; E G. - filiformis, G. perfoliata —— N° attempt is made here to distinguish between Swertia and Frasera. ee loc. cit. 1896-19 d I J. P., Inst. Rei Herb. 80. 7. GË og Gunther SN Is The Greek Herbal of Dioscorides. "Oxford d, 1934. ° Ruppius, H. B., Fl. Jen. ed. 2. 17. fieret ^ Hall. 21. 1745. [Vor. 44 202 ANNALS OF THE MISSOURI BOTANICAL GARDEN Necker,** in 1790, included most of the Gentianoideae in one of his all-inclusive "genera". His "genus" Darinyphytum not only included most of the Gentianoideae but also a great variety of incongruous groups from a number of modern families. Within this "genus" he recognized a number of "species". Some of these, “Pneu- monantbe", v Anibojo gon", "Spiragyne", and "Tbyropbora", are equivalent in extent to modern genera. In most cases his descriptions are so inadequate that these "species" cannot be identified readily with modern taxa. Plates published in his ‘Elementa Botanica’ demonstrate characters used in the diagnosis of his super-genera. This ambiguity of description, failure to include names on which his "species" were based, and his usage of a species concept quite at variance with either modern practice or that of other authors of his day, make these names nomina invalida. Significantly, perusal of the 1950 International Code reveals that fifty-seven Neckerian names have been rejected for one reason or another, against six conserved and three proposed for conservation. In 1794, Moench” described the genus Gentianella, basing it on Gentiana campestris L. which he renamed Gentianella tetrandra. No other species were assigned by him to the genus. In 1796, Schmidt? divided Gentiana of Linnaeus into six segregate genera: Pneumonantbe, Hippion, Chironia, Gentiana, Swertia and Chlora. Hippion in- cluded sections cRossoPETALUM Froelich, AMARELLA Griseb Se ARCTOPHILA Griseb., now included within Gentianella. Borckhausen,2" in a paper following 1n the same publication, divided Linnaeus’ genus into thirteen genera of which Gen- tianella corresponded to section cRossoPETALUM of Froelich. ich*® monographed Gentiana, also in 1796, describing 47 species which he divided into four undenominated groups: COELANTHE, CALATHIANAE, ENDOTRI- CHAE, and CROSSOPETALAE. Many later authors have assumed these to be sections. In 1829, Bunge” described 80 species of Gentiana but maintained a classification somewhat influenced by Borckhausen. n CROSSOPETALAE corresponde to Grisebach's ckossoPETALUM. Endlicher?? essentially adopted Bunge's system in his ‘Genera Plantarum’ and maintained section cRossoPETALUM Froel. Grisebach's monographic study of the Gentianaceae?! is one of the landmark works of the family. His genus Gentiana contains some 125 species, which are divided into seven sections. In his treatment of the Gentianaceae in De Candolle's ‘Prodromus’ in 1845,?? the number of species was increased to 153. In 1838 the *4 Necker, N. J., Elem. Bot. ^ 11-14. 1790. 25 Moench, loc. cit. 482. 17 chmidt, F. W., in Roemer ‘Archiv f. Bot. 1:3-23. 1 ?? Borckhause: sen, loc cit. fie unge, A., in Nouv. Mem. Soc. Imp. gen Moscou 1:197—256. 1829. we indlicher, S. L., Co Pl. em 1836—4 vo ; A. H. R., Gen. et Lens Cane 1839. Prodr. $: rina 1845 1957] GILLETT—REVISION OF GENTIANELLA 203 North American species were treated separately by Grisebach in Hooker’s ‘Flora Boreali- Americana’. Huxley?* in 1888 suggested a rearrangement of the family based on the position of the corolla glands. His MESOMELITAE includes the plants having glands at the ase of the ovary, while his PERIMELITAE includes those with epipetalous glands. In 1860, Turczaninow?? monographed Gentiana in his ‘Monographische Beschreib- ung der Enzianen’. In 1892, Murbeck?9 published a small study of the European ENDOTRICHAE which includes several new species with distribution maps of some entities. He failed to include keys to the species. In 1894, Kusnezow?' divided the Linnaean genus into two subgenera, EUGEN- TIANA and GENTIANELLA. His voluminous monograph covers his subgenus EU- GENTIANA only. He made no attempt to revise subgenus GENTIANELLA beyond establishing characters to distinguish it. This monograph was later translated from the Russian into German and was published in parts from 1896 to 1904. Kus- nezow?9 again maintained the subgenus GENTIANELLA in his treatment in Engler and Prantl’s ‘Die Natürlichen Pflanzenfamilien’ in 1895. In 1896, Wettstein?? monographed the European species of the section ENDO- TRICHA, recognizing 27 species; and in 1900, he revised the North American species of the same section 20 In 1896, Wettstein“! segregated Gentiana tenella Rottb. from the section ENDOTRICHA, erecting a new section COMASTOMA. In 1901, Th. Holm* presented a short revision of the Canadian species of the section CROSSOPETALUM of which he described three new species. ; In 1917, Gilg?? monographed the South American species of Gentiana, recog- nizing 182 species of subgenus GENTIANELLA and one species of the subgenus EUGENTIANA. In this work he created a series of divisions to which he assigned no rank. Schustlerf* returned Gentianella to t ascribing it to Borckhausen rather than to Moench. In 1936, H. Smith? raised Hemsley's section MEGACODON of Gentiana to the rank of genus, and in 1937, Marquand*® united the genus Crawfurdia to Gen- tiana, but relegated the section PTERYGOCALYX Maxim. of Crawfurdia to Gentian- —— he rank of genus in 1923, incorrectly a Hooker, W. J., Fl. Bor-Amer. 2:54. 1838. 35 Huxley, T. H., in Jour. Linn. Soc. Bot. 24:101—124. 1888. inow, N.. in Nat. Sci. Meetings, 34-35. 1860 (Not seen). 8 sa Kusnezow, loc. cit. 1896-1904 (German; des 1,42:85. 1895. : n ae Wettstein, R. V., in Denkschr. K. Akad. Wiss. Wien, math.-nat. Cl. 64: ` ` A terr, Bot. Zeitschr. 50:168—173, 189—195, 290—293. 1900 ss Oesterr. Bot. Zeitschr. 46:121—128. 2 Ottawa Nat. 15:175-183. S 44 loc. cit. 17. i loc. cit. 1923. Smith, Harald, in Handel-Mazzetti, Symb. Sin. 7:950. 1936. *5 Marquand, C. V. B., in Kew Bull. 134-180. 1937. [Vor. 44 204 ANNALS OF THE MISSOURI BOTANICAL GARDEN ella. In an earlier paper?! he had merely excluded rrERvcocarvx from Gentiana. Yu-chuan Ma,*® in 1951, raised the section CROSSOPETALUM to generic rank, naming the genus Gentianopsis, but discussed no synonymy. However, whether by design or accident, no earlier generic name would have been valid for this group. GENERAL MORPHOLOGY The North American species of Gentianella are glabrous herbs. In habit the plants are branched or simple. Of the North American species, only G. barbellata is a true perennial; the others are either annual, biennial or perhaps winter annuals. Actual growth tests are needed to determine their life cycle. he primary root is rather small in comparison with the size of the entire plant and is slightly thickened, forming a slender tap-root. As a rule there are few secondary roots. In G. barbellata there is a horizontal rhizome which apparently grows a few inches below the surface of the soil and sends vertical branches upward to the surface. At the surface level these branches thicken considerably and are enveloped in the basal rosette. These horizontal rhizomes seldom are collected and appear on only the occasional dried specimen. Probably collectors as a rule do not take the trouble to dig into the stony ground of alpine tundras! : The stems frequently become woody at the base and, in coarse plants such as G. crinita, they become hollow in the lower part. Variously developed stem-wings occur in many of the plants, or the stems may be prominently angled or striate. The wings, however, vary to such a degree that the character is of little diagnostic value. In some plants the wings are twisted spirally about the lower part of the stem. This feature greatly influences the symmetry of the phyllotaxy, so that the leaves, instead of being strictly decussate, show a distinct shift in their position at each successive node. The leaves in Gentianella are glabrous as are those of the entire family. Basal leaves are spatulate to lingulate, and may be sessile or the blade may be attenuated rather sharply to a narrow, decurrent, connate, or clasping petiole. These leaves may wither early in the season. The shape of the median leaves is sufficiently diagnostic to be useful as a Supporting character. In the fringed gentians the leaves usually have a single mid- EA vein; in the amarellas, arctophilas and comastomas the venation is palmate with three to five nearly parallel veinlets. The mid-vein is generally fairly prominent, while the laterals may or may not be distinct. Leaf margins are entire throughout the genus, occasionally minutely papillose or ciliate under magnification. If a leaf be carefully pulled downwards and torn from the stem so that a frag- ment of the stem epidermis is removed with it, several minute processes or squamel- åt Kew Bull. 66-68. 1931. 48 loc. cit. 1951. 1957] M GILLETT—REVISION OF GENTIANELLA 205 lae may be found attached to the base. These structures have been reported in other families of the Contortae, and may represent reduced stipules.*? The inflorescence is basically cymose, consisting of a simple cyme, an aggregate umbelliform cyme borne in a compact head, or may be reduced to a single terminal flower on a slender pedicel. Often flowers are borne singly or in pairs in the axils of the cauline leaves (dichasium). The upper leaves may frequently become bract-like and envelop the inflorescence, as in G. aurea. The calyx is tubular in all species except G. wislizemi where the tube is split down one side to form a spathe, with the lobes reduced to minute teeth at the apical margin. Grossheim?? described this condition for the European G. ciliata but I have never seen this type of calyx in any of the North American fringed gentians. In subgenus EUBLEPHIS (the fringed gentians) there are four calyx lobes, Probably representing an inner and an outer cycle. The lobes of this group are thinly membranaceous, usually hyaline-margined, the outer lobes frequently plicate and carinate at the tip. The sinuses of the calyx lobes have a small internal hyaline membrane which bears a row of minute blunt processes. The venation of the lobes of subgenus EUBLEPHIS is usually reticulate. In subgenera GENTIANELLA and comasToMA there may be four calyx lobes or five forming a quincunx. The lobes vary in shape and length with respect to the tube and furnish useful characters. The calyx lobes of these two subgenera are somewhat fleshy. A mid-vein and two lateral veins occur and the laterals either unite below the sinus or at the base of the tube. Squamellae are usually found on the inner surface of the calyx tube near the base. These structures, however, are of little taxonomic value. The corolla may be tubular, funnelform, or salverform when the lobes are fully expanded. In North American species its color varies from white to yellow, ue, and violet. South American species are often brilliant yellow or red. The lobes are oblong, obovate, ovate, to triangular, with acute, apiculate, mucronate, or obtuse tips. At the base of the lobes in many species there is a horizontal row of briae (the crown) which may be free or may be united at the base to form a fringed scale. In section AMARELLA these fimbriae usually are continuous across the base of the lobe, sometimes irregularly scattered, with a vascular trace in each bria. In the coMAsTOMAE, on the other hand, there are two separate fimbri- ate scales at the base of each corolla lobe and there are no vascular strands within the fimbriae. As in certain apocynaceous corollas, these scales may be of stipular Origin,51 At the base of the corolla tube and alternate with the stamens are small glands 9r nectaries of various shapes. These never appear foveate, but are more frequently swollen. The glands are never accompanied by fimbriae. — 50 G den W., in Ann. Mo. Bot. Gard. 37:484. 1950. rossheim, A. A., in Fl. U. S. S. R. 18:591. 1952. Woodson, R. E. pare r J. A., in Bull. Torr. Bot. Club 65:135—165. 1938. [Vor. 44 206 ANNALS OF THE MISSOURI BOTANICAL GARDEN The stamens are equal in number to the corolla lobes and are alternate with them. Occasionally in G. wislizeni and G. microcalyx the anthers abort, with the staminodia elongate or reduced to small protuberances. The stamens are inserted near the center of the tube or below. e filaments usually are winged at the base, the wings curving inwards to form a canaliculate filament. The anthers are ob- longoid or ovoid, and versatile, appearing introrse before dehiscence, then rotating and becoming extrorse following dehiscence. The three-pored pollen grains are essentially uniform in the genus. The size and shape of the pistil are dependent upon the age of the flower. Prior to fertilization the style frequently is distinct; after fertilization it may gradually disappear with the maturation of the ovary and the formation of the capsule. The gynophore may also vary in length and distinctness with age. In some species the pistil is sessile or subsessile; in others the gynophore may be minute in young flowers, elongating considerably as the capsule matures. In some the gynophore always is elongate. The two stigmas vary considerably in size and shape. They may be flabelli- form and broad; or oblong, elliptic, or reniform and narrow. Their margins may be entire, sinuous or toothed—all features that vary within a population. Seeds are of very great value to distinguish sections, but in most instances fail to separate species. Those of subgenus EUBLEPHIs are oblong and frequently angular, with reticulate or papillose surfaces. In two species, however, G. simplex and G. ciliata of Europe, they are reticulate and pronouncedly caudate. Those of subgenus GENTIANELLA series AMARELLAE and ARCTOPHILAE, as well as those of subgenus COMASTOMA, are round to slightly flattened. The color almost invariably is light brown. GEOGRAPHICAL DisTRIBUTION The genus Gentianella has an almost global distribution. In North America the two subgenera EUBLEPHIS and GENTIANELLA, have roughly a similar dis- tribution but apparently different histories. Both groups have species that occur in Europe and Asia and connect with North America through either Greenland or Alaska. The variability and complexity found in the North American populations also occur in Asia and Europe. The South American species of the genus are placed chiefly in sections IMAICOLA and ANDICOLA. In subgenus coMAsTOMA only one widespread species, G. tenella, is represented in North America. This species occurs in Colorado south to Arizona and New Mexico, west to California, and in a few scattered stations across Alaska, Canada; and Greenland. The same species is common in Iceland, northern Europe and in the high mountains of southern Europe and Asia. The fringed gentians (EUBLEPHIS) exhibit interesting distributional pat- terns. G. simplex and G. barbellata both occupy rather limited areas below the margin of maximum glaciation and probably have had quite an independent and 1957] GILLETT—REVISION OF GENTIANELLA 207 an older history than the detonsa-crinita complex. The seed of G. simplex of Cali- fornia bears a striking resemblance to that of G. ciliata of Europe. This seed similarity between these two species prompted Engelmann?? to set up two groups or series of undesignated rank to indicate this relationship: a. Lepidospermae (G. crinita, G. lanceolata, G. barbellata), b. Pterospermae (G. ciliata, G. simplex). Seed similarity is the only major character that these two species have in common, so the classification of Engelmann may not represent true relationships. Recently Grossheim in the ‘U.S.S.R. Flora’ employed a similar division of the species (Series CILIATAE and BARBATAE) but based it on floral characters, apparently failing to note the distinct nature of the seed of G. ciliata. If we hypothesize a Eurasian origin, there may have been several “invasions” of the detonsa-crinita complex into North America. The meagre, but no less valuable, chromosome data at hand indicate an n = 13 series represented by G. crinita and its ssp. procera, and an n = 11 series represented by G. detonsa (prob- ably ssp. detonsa), hexaploid and tetraploid species respectively. The most stable portion of the complex is that occurring south of glaciation. Apparently the crinita group entered North America either from the north and reached the Appala- chians, or arose independently there. The detonsa group perhaps entered through Alaska and migrated southwards as far as central Mexico with a branch into Cali- fornia (ssp. holopetala). Glaciation probably eradicated the group in the north except for a few isolated refugia. This may have been followed by a remigration back into the deglaciated area from Alaska -Yukon by means of the Yukon and Mackenzie River systems. The southern portion of the detonsa population was unable to move north perhaps through lack of suitable high altitude/high latitude habitat requirements. The crinita group, finding suitable low altitude/low latitude habitats, moved rapidly northwestward into the Great Lakes region and across the Plains as far as the Rocky Mountains to close the population gap. This broad conjectural outline may aid in understanding the subspecific treat- ment afforded this complex. The various populations are more or less imbricate in their distribution—or in the case of the crinita group, clinal—indicating a differ- entiating process accompanying the migration flow. Further discussion follows in the taxonomic treatment. Stupy MATERIAL stitutions have been available Specimens or photographs from the following in prei by visits to the herbaria. For for the pur pose of this study, either through loans or ege 5 D : Engelmann, G., in Trans. Acad. Sci. St. Louis 2:216. 1868. [Vor. 44 208 ANNALS OF THE MISSOURI BOTANICAL GARDEN the sake of consistency, the abbreviations used are those listed by Lanjouw and Stafleu. A—Arnold Arboretum, Cambridge, M C—Universitets Botaniske Museum, Copenhagen, Denmark. CAN—National Museum of Canada, Ottaw DAO—Botany and Plant Pathology positos Science Service, Canada Department of i Ont et DH griculture, O ; nt. F—Chicago Natural History — seum, Chicago, Ill. GH—Harvard University Herbarium, Cambridge, Mass agp oes of the Thiereracy yr. Idaho, ion Ida. (This includes the Botany ium and the SCH nge d ean of the Department of Forestry). JBM— Cé Botanique de Montréal, Montréal, P. MR Nacional del Sage de Biologia de la Universidad Nacional de México, o, D. MO Misouri Botanical Cie? St. ten Mo. Sg Minüeipólis; Minn. NY—New York Botanical Garden, "New Y, ork, E dtum, University of Notre Dist (Greene-Nieuwland Herbarium), Notre Dam PH Academy E Natural os Philadelphia, Pa. Naturhistoriska Riksm SW C—Dominion i WS- State College of Washingt an ocky Meester? Herba Sd ech of Wyoming, Laramie, Wyo. Citation of Specimens: For reasons of economy, citations have been consider- ably reduced. Specimens of species represented by little material have been cited in full. Those of common and widely distributed species have been reduced. One specimen from each county or, lacking counties, places or districts is cited. Indi- vidual collections have been selected for citation on the basis of breadth of distribu- tion of duplicates or of historical importance. cknowledgements: The writer would like to express his appreciation to Dr. Robert E. Woodson, Jr., under whom this work was carried on, for his help ful guidance and criticism; to the curators of the herbaria listed above for loans or use of their herbaria and facilities; to Dr. H. A. Senn who enabled field observations to be carried out; to Dr. Amy Skallerup who sketched the flower dissections; an to the many other workers who contributed advice, material or discussion. TAXONOMIC TREATMENT GENTIANELLA Moench, Meth. Pl. 482. 1794, emend. Schustler, in Vestnik Sjezdu Cesko Bot. v. Praze, 32-34. 1923. (T.: Gentianella tetrandra Moench based on Gentiana campestris L. Sp. Pl. ed. 1, 231. 1753 = Gentianella campestris (L.) Bórner, Fl. deut. Volk, 543. 13153 53 Lanjouw, J., and Stafleu, F. A., in Regnum Vegetabile 2. Index Herbariorum, p. 131. ed. 2, 1954. 1957] GILLETT—REVISION OF GENTIANELLA 209 eis MNA Elem. Bot. 2:12. 1790, ut species; Raf. Fl. Tellur. 3:25. 1837, ut genus, t. Gen. N. eg Pl. 1:81. 1818. ori rene Nec cit. 12.717 ue nom. ambig. Otsenthe Delarb. Fl. CSC Ei 1800. Gentiana L. subg. Gentianella Kee i "Eagles & Prantl, Nat. Pflanzenfam. ed. 1, 47:85. 1895. Caulescent or acaulescent annual, biennial, winter annual, or perennial herbs, rare- ly shrubs (in South America), usually with tap-roots, rarely with slender rhizomes. Leaves opposite, membranaceous or fleshy, sessile or petiolate, usually palmately 3- to 5-veined, the veins prominent or inconspicuous. Inflorescence a terminal or axillary umbelliform dichasium, or an aggregate or simple cyme, frequently 1- flowered. Calyx 4- to 5-lobed, tubular, obconic, or 4-angulate, rarely spathiform, without an inner continuous membranaceous rim, carinate or ecarinate, glabrous or scabrous-papillose, frequently bearing a number of minute squamellae within and at the base of the tube. Corolla 4- or 5-lobed, marcescent, tubular, funnel- form, campanulate, to rotate (in S. American species), the tube with or without faucal fimbriae; lobes with 5—9 parallel veins, sinistrally or quincuncially convolute in the bud, without interlobal plicae; interstaminal glands at the base of the tube (never at the base of the ovary) somewhat scutiform or rounded, swollen or appear- ing only as green patches on the surface of the tube, the glands never foveate. Stamens 4 or 5, included; filaments linear or variously taper-winged, rarely ciliate at the base; anthers 2- celled, oblong, rarely somewhat triangular, versatile, introrse in the bud but becoming extrorse following dehiscence. Ovary sessile or stipitate; carpels 2, 1-celled, the placentae parietal; stigmas 2, sessile or with a tinct style, oblong to flabelliform or reniform, lobed or entire, revolute or plane. ruit a cylindrical or ovoid capsule, septicidally dehiscent from the apex. globose or slightly flattened, angular or caudate, reticulate, papillose or smooth, brown or light tan. KEY TO THE SUBGENERA a. — 4-merous; vs tubular, the lobes with thin —, margins, the sinuses with ute blunt processes; anthers the en mall inner m ie? SE across the base and bearing m a aa. ied 3 or 5-merous; calyx tu ro afi t V ibit eb the si : attached at about the middle; ovules borne in two rows along the margin rarely wick a third row slong: deg carpel wall between the sutures; seeds gl e or » dech flattened, smoo th, Ce Ce b. Pedicels conspic uously shorter than the subtending internode; corolla Sere opa naked, KE Dec scattered cepe: minutely papillose fimbriae, or the riae uni SCH VENTANA bb. a single scale extending across the base of each corolla pbe- Kit Ge Seen Kë Pedicels conspicuously ae than the subtending internode; o fimbri at the base of each corolla lobe, the ho s shortly digitate, —— Ki e —— [Vor. 44 210 ANNALS OF THE MISSOURI BOTANICAL GARDEN SuBcENus I. EUBLEPHIS (Raf.) J. M. Gillett, comb. nov. Gentiana L. subgenus Eublephis Raf. Med. Fl. 1:208. 1828. (T.: G. crinita Froel.) oh Gentianella Borckh. in Roem. Archiv f. Bot. 1:29. 1796, non Moench. _ (Ta G. ciliata (L.) Borckh.) entiana L. **** Crossopetalae Froel. Gent. Diss. 109. 1796, sine ordine. Denckea Raf. in Med. Repos. II, 5:352. 1808, nom. nud. ins Crossopetalum Roth, Enum. Pl. Phaen. Germ. 1:516. 1827. (T.: C. gentianoides Roth entiana ciliata L. Gentiana L. Urananthe Gaud. Fl. Helv. 2:270, 1828, sine ordine. Gentiana L. sect. Crossopetalum Froel. ex Griseb. in Hook. Fl. Bor. Amer. 2:64. 1838. Gentianopsis Ma, in Acta Phytotax. 1:7. 1951. (T.: G. barbata (Froel.) Ma) Flowers usually large and showy, 4-merous, long-pedicellate. Calyx usually 4-angulate, the lobes with thin hyaline margins, the two inner triangular to ovate, the two outer linear to lanceolate, acute, and usually conduplicate and the margins united at the tips, each sinus with a thin inner membrane bearing few to many blunt processes. Corolla tube without faucal fimbriae, the lobes with reticulate venation connecting the main veins. Stamens inserted near the middle of the tube. Ovary with the parietal placenta extending over most of the inner surface. Seeds ellipsoid to angular, minutely or strongly papillose, or caudate, light brown. Tetraploids or hexaploids on base numbers of 11 and 13. Type species: Gentianella crinita (Froel.) G. Don. KEY TO THE SPECIES a. Plants in general branched, occasionally simple (but usually occurring among populations of branched plants), with a definite rosette of basal leaves or a tende form a rosette 1 D basal 1 ncy to fo by shortening of the lower internodes; flowers solitary or in axillary or terminal cymes; papillose. rminal flowers borne on slender pedicels extending beyond the upper pair of leaves; corolla tube glabrous below the insertion of the stamens; annuals or biennials with c. Base of the calyx and the calyx keels smooth or wrinkled, the keels glossy purple; upper leaves usually obtuse 1. G= cc. Base of the calyx near the junction with the pedicel and usually at least one pair of calyx keels covered by whitish or hyaline papillae (under magnification) occasionally present also on the margins of the lobes, the keels green or purple; upper leaves par crinita ow the flower and subtending it; corolla tube bearin rows of long cilia below the ' insertion of the stamens; perennials from slender rhiz . G. barbellata aa. Plants simple, never branched, without a basal rosette of leaves; the single flower borne. terminally on an elongate naked pedicel; seeds caudate, the surface reticulate.....— 4. G. simplex While G. simplex and G. barbellata are both morphologically and geograph- ically very sharp, clear-cut populations, the G. detonsa-crinita complex is much less convincing. The complex may be considered either as one large species Of divided into two rather weak species. The presence or absence of papillae on calyx keels is a constant character differentiating these species. The obtuse oF rounded medium leaf tips of the detonsa group versus the acute tips of the crinita group is a strong supporting character. Other characters either vary considerably or are influenced both by plant age or by the presence of the small form of the plant found so commonly among the larger members of a population. 1957] GILLETT—REVISION OF GENTIANELLA 211 I have chosen to recognize G. detonsa and G. crinita as two separate species for the following reasons: 1. Gentianella crinita ssp. crinita has an Appalachian distri- bution probably indicating an older history than that of G. defomsa. 2. The dis- junction between the non-papillose northern G. detonsa ssp. raupii and the central ssp. elegans can be more readily explained. 3. The presence of the narrow-leaved papillose-keeled ssp. macounii between these two disjunct populations can best explained by migration. 4. Present-day scant cytological evidence in the group indicates the presence of at least two chromosome series, the detonsa series with a base number of 11 and the crinita series with a base number of 13. 5. The occa- sional presence of sterile and depauperate forms in Manitoba may be due to occasional crosses between members of these two chromosome series. 1. GENTIANELLA DETONSA (Rottb.) G. Don, Gen. Syst. 4:179. 1838. Gentiana detonsa Rottb. Kiob. Selsk. Skr. (Acta Hafn.) 10:435. 1770. (T.: Collector unknown, herb. Rottbóll, C!) Annuals or biennials, 0.5—9.0 dm. tall, branched from the base or above, or simple, the basal branches curved or straight, shorter than (or nearly equal to) the main stem. Basal leaves elliptic, obovate-elliptic to spatulate, the apex rounded, obtuse, or acute, attenuate below to about the width of the stem, 0.5-3.5 (—6.0) em. long, 0.1—1.5 cm. wide, forming a dense rosette, or frequently reduced to one Pair; median leaves linear or linear-lanceolate to elliptic or spatulate, usually rounded, the base attenuate to about the width of the stem or slightly broader, frequently subamplexicaul or clasping, 1.5—6.5 cm. long, 0.1—0.7 cm. wide. Flow- ers solitary and terminal, long-pedicellate, the terminal pedicels 1-30 cm. long, those of the basal or lateral branches when present shorter, occasionally equal. Calyx narrowly to broadly funnelform, the tube 9-14 mm. long, 5-7 mm. wide below the lobes, the inner lobes ovate-triangular, acute, about 10 mm. long, the outer lobes longer, lanceolate and conduplicate, frequently carinate at the tip, acute, up to 14 mm. long or equal to the inner lobes; keels 1 mm. or less wide, scarcely prominent, smooth and reticulate, generally purple-tinged; sinuses acute, covered on the inside by a small straight to sulcate membrane extending to the center of the outer lobes and bearing few to many blunt processes up to 0.5 mm. long. Corolla pale to dark blue, narrowly funnelform, 2-4 (-5) cm. long, 4-10 (-15) mm. wide at the orifice, the lobes oblong or somewhat expanded above, oe 15 mm. long, 5-7 (—15) mm. wide, the tips erose to denticulate, the margins with few to several cilia 0.5-1.5 mm. long in the lower half. Stamens included, the filaments inserted near the middle of the corolla tube, the marginal wings 1.5 mm. wide at the base, tapering gradually above; anthers 2.5 mm. long, 1.0 mm. a attached above the middle; interstaminal glands oblong, crescent shaped to rounded, swollen below and tapering to the tube above. Pistil short-stipitate, the gynophore 2-4 mm. long, obscurely attenuate; ovary fusiform, about 12 mm. long, 2.0 mm. wide; stigmas sessile or short-stipitate, flabelliform, reniform, or oblong, < 5 mm. wide. Capsule as long as the corolla tube or slightly longer, dehiscing in the upper 212 [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN third, the tips of the valves recurving. Seeds ovoid to oblong, slightly flattened, the surface reticulate and with slender or rounded, inflated or somewhat collapse hyaline papillae irregularly dispersed or concentrated at the angles and ends. Circumboreal, extending from North America to Europe through Greenland and Iceland, and to Asia through Siberia; in North America from both the east and the west coasts of Greenland as far north as Disko and Cape Oswald, on the mainland westward to Coppermine and Kotzebue, southward on the east to Quebec and Newfoundland and on the west through Alaska and the Yukon through the Rocky Mountains and the Sierra Nevada to the Sierra Madre of Mexico; resolving into several subspecies and numerous minor local variants. : In a variety of habitats: tidal estuaries, river banks, mountain meadows, saline meadows, about hot springs and sloughs, in open or sparsely wooded plains, but in general preferring lime soils or shallow soil over limestone; flowering and fruiting times various depending upon latitude but generally from late June until November; at altitudes from sea-level to 2500 meters KEY TO THE SUBSPECIES a. Margins of the corolla lobes ciliate at least pedicels 2-15 cm. long; in the lower half, the tips erose to dentate; a pos not swi vollen. 8; gynophore short, never as long as the capsule; interstaminal H ned seite restricted to the uds occasionally ttenuate i ruptly constricted; corolla slender, mm E ue os — ntly constricted or scarcely pren ned a rifice, the tube aedi obconic. piens d west of the Mackenzie Mou e. sim seed or the branches arising from the base; basal rosette poorly developed or qct to a single pair of leaves. Circumboreal above the north- ern limit of trees, Kotzebue to Coppermine and the coast of Greenland AMAN Uti ripe ew enr is; G: » — ssp. defonsa cc. Tou 2—6 dm. tall, simple or the branches arising from the axils of cauline leaves rely from the base; Sen Di rosette well developed, the leaves dense, ns reduced particularly i in smaller specimens, Yukon and Alaska, chiefly along y dia Lacs on River ace; tube abru de Set dicel; corolla ery - Ew cuo lg es AO. the parts broadly obconic or poc es: Fl —4 cm. long, rarely longer, 0.5— .wi the orifice, the keels of de cal ot prominent; caulin e leaves "am to spatulate. (—2,0). dm. hi gh, profusely branched from the densely the green or Ren d branches curved-ascending; corolla somewhat short and truncate, FAS 5 w 12 the lengt » of the tube. Fast bila of James Bay, ies and Newfou . detonsa ssp. seo ee. Plants kasalik Ge Hira: high, branched E oc Ban plants, the b es an Gd requently and stron ngly purpled; corolla gie ngate iod round weis equal in witty to the corolla tube. A highly variable complex ext Ps i ‘from the rides Valley east to Hog cm. long and 1.6 cm. wide at the orifice, the Kai bag the calyx prominent, Pri at the base of the tube; cauline leaves lan vorn to mpi f. Terminal pedicels of mature plants e (-16) cm. long; flowers Leg long; plants generally br gang from the base, Central Rocky ege? ein kde 4 G. detonsa ssp. elegans ff. Terminal j pedicels 9-30 cn cm. long; ‘flowers s 4. 5-8. 0c m. long; Ee S culis branched above the base. Southern Arizona; in the Sierra Madres sou ge A EE E EE Eed dE d Ee, 1957] GILLETT—REVISION OF GENTIANELLA 213 aa. Margins of the corolla lobes entire, devoid of any fimbriae or teeth, the tips entire to erose; pedicels 2—3.5 cm. lo: g. Plants 4-9 dm Get be MER in the upper part, rarely simple; ën short, rang 5 mm. long; calyx keels rarely papillose; interstaminal gla n ntly free base and ayar outhern Sinaloa and Zac o Mic e án.. EG. etonsa a nct 88. Planes 0.3—4.5 dm. sl ER from t Mei e or x Disp synoghor frequently alm ong as the capsule, 8-11 mm. long; aha keels always smoo E e ees sland mg below and fused to de pie. tube. California to -— evada and da O aed ssp. bolopetala la. GENTIANELLA DETONSA (Rottb.) G. Don, ssp. DETONSA I ciliata Gunn. Fl. Norv. 2:88. t. 2. f. 3-5. 1772, ex ic. entiana serrata Gunn. B detonsa (Rottb.) Wahl. Fl. Zeg 71. 1812 eg NM Froel. B simplex Bunge, in Nouv. Mem. Soc. Nat. Mosc. 1:22. 1. 9. f. 1. Gentiana e SE Bunge, loc. cit. 225. £. 9. f. 3. 1829, ex 1 ie Gen Ge tiana ba réie Froel. y Richa e n: Macnab, in Edinb. New Phil. Jour. 19:62. 1835. n., MO, photo!) Anthopog on detonsa (Rottb.) Raf, Fl, Tellur. 3:25. 1837. roel.) Raf. loc. cit. 25. psi detonsa Fries B barbata (Froel.) Griseb. in Hook. Fl. Bor.-Amer. 2:64. 1838. entiana detonsa Fries "y — Gene gl Griseb. irs cit. 1838. Vict. in Contr. Lab. Bot. Univ. Montréal 20: Porsild & Porsild s. n.!) 18. Gentiana richardsoni Porsild, in Nat. Mus. Can. Bull. No. 121: :274—275. 1951. (T: Gentianopsis barbata (Froel.) Ma, in Acta Phytotax. 1:8. 1951. (Synonyms referring to other Eurasian types are not included). Subspecies detonsa has a circumboreal distribution but appears to occupy rather y and Iceland, on localized areas within its range. It occurs in Finland, Norwa both east and west coasts of Greenland, in America and Asia. Between Greenland and Coppermine, District of Mackenzie, the disjunction in distribution is con- stantly being reduced by additional collections. Westward, subspecies detonsa occurs at Atkinson Point at the mouth of the Mackenzie, at Kotzebue Sound, and after a further gap, in Mongolia and Siberia, being notably absent from Kamchatka. Found in moist ppg along sandy sea beaches; flowering from mid-July to early August; fruiting in August. I have included G. bibaka within the synonymy o à emphasize that the name barbata should not be applied to our North American material. G. barbata, if considered to be distinct at a later date, a applies to the taller Eurasian material and appears to be transitional to ssp. detonsa. In this view I follow in part Grisebach, because I have had little European material at my dis- - However, our North American material in my opinion is not identical With the taller European material nor with the various species from Asia illustrated by Yu-chuan Ma in his recent revision. f this subspecies in order to [Vor. 44 214 ANNALS OF THE MISSOURI BOTANICAL GARDEN Ld sr s G. detonsa G. crinita © ssp. detonsa ©) ssp- crinita .® ssp. yukonensis & ssp. procera y © ssp. nesophila @ ssp. victorinii @ ssp. raupii [ ssp. macounii A ssp. elegans : ssp. superba ADO ssp. lanceolata AL ssp. holopetala Ty f N Pid | e = e A, e ~ A d Kl 9| e x "KA s "X rus Lef MÉI Fig. 1A. Distribution of the subspecies of Gentianella detonsa and of G. crinita. Note: Through an error have been interchanged. in drawing, the symbols for G. detonsa ssp. nesophila and ssp. kolope o 1957] GILLETI—REVISION OF GENTIANELLA 215 Fig. 1B. Calyx and corolla of Gen- tianella detonsa ssp. detonsa (right, approx. nat. size) and of G. crinita ssp. crinita (left, X 14). UNITED STATES: ALASKA: Keewalik Spit, Miller 25c (US); Kotzebue, Anderson 4683b (S), 4766 (S), Lepage 25406 (DAO), Scamman 4063 (GH), 4663 (CAN). GREENLAND: Igdungujak, Hoff s.n. (C); Ekaluit i Igalikofjord, Lindemann s. n. (C); Itivdleq-Fjord, Qingua, Porsild s.n. (C, CAN, MO, US); Igaliko (Gardar), Porsild & orsild s.n. (C, CAN, GH, JBM, US), Rosenvinge s.n. (C, CAN), Vahl s.n. (C); Julianehaab, Rosenvinge s. n. (C), Vahl s. n. (C), Porsild 8085 (CAN). CANADA: MACKENZIE: Coppermine, Findlay 241 (DAO); Atkinson Point, 70° N, 131° 20' W, Porsild &$ Porsild 2653 (CAN). lb. GENTIANELLA DETONSA (Rottb.) G. Don, ssp. yukonensis J. M. Gillett, subspec. nov. Herbae annuae. Caules erecti, 2-6 dm. alti, simplices sive ramosi, ramis ex axillis foliorum caulinarium interdum radicalium emergentibus. Folia radicalia pluralia in rosulam congesta elliptica vel spathulata apice obtusa, 6-25 mm. longa, 5-10 mm. lata; folia media linearia vel lineari-lanceolata, 20-60 mm. longa, 1-6 mm. lata. Flores et semina subspeciei typica similes. ather high and the distribution is available for distribution. The range of subspecies detonsa corresponds rather closely to that region north of the Nordenskiold line which in turn roughly corresponds to the northern limit of trees in North America. Subspecies yukonensis occurs south of this line and apparently is restricted to the valley of the Yukon River and its tributaries. [Vor. 44 216 ANNALS OF THE MISSOURI BOTANICAL GARDEN gei STA KA: dierfe Anderson s. n. (S) ; Mackintosh, Rae & Brown 10065 (S); bermea Summit and ME Porsild & Porsild 440 (CAN, GH, US); near Delta Junc- miles south of D ëwech 60° 00’ N, 145° 45' W, June 28, 1951, W. J. Cody 8 e M. Eo ser? (DAO, HOLOTYPE). CAN YUKON: * Pely River, Mayo District, Bostock 226 (CAN); Ranch Valley, Gorman 1081 (CAN, NY, US); €: nasum Noel 38 (UBC), 14184 (CAN); Lake Kluane to Don Jek River, Müller s. . (PH); above pen Selkirk, Tarleton 115a (US), 115b (NY); Cormacks, Yukon River, ahed 567 (CAN lc. GENTIANELLA DETONSA (Rottb.) G. Don, ssp. nesophila (Th. Holm) J. M. Gillett, comb. & stat. nov Gentiana nesopbila Th. Holm, in Ottawa Nat. 15:11. 1901. (T.: Macoun s. n.!) Northwestern Newfoundland, the Mingan Islands and Anticosti, and on the east coast of James Bay, on sandy or gravelly calcareous soil, at elevations from sea level to about 20 meters; flowering from late July until early September, ruiting from August until October. CANADA: KEEWATIN: Boat O » Manitounuk Hon je & Lepage 12944 (DAO), I mu (GAS South ton Idande, Jobansen 273 WFOUNDLAND: Cape St. George, Mackenzie & kreie, III21 (GH) ; Cape Norman, Wiegand & Long 28927 (GH); Cook Point, Fernald, Gilbert & Hotchkiss ag tt (GH); Brig Bay g e a 26969 Ds Cow Head north of St. Paul's Bay, Fernald "o y oed 3898 (GH); Port au Port de Gravels"), Fernald Si y and 99 (G A Dutilly 2 us d 15797 (CAN EC: Anticosti Island, nr zor (CAN s. E . (CAN); mouth of Jupiter River, Wyatt, Sbaler P Verrill, s. n. en). Rousseau 524! I (JBM), Victorin > Rolland 20 Fd JBM ); ux zeen , ut ill omg der ĝi wi i D D et le 4 x 9 , 3 5 ( Es Kegashka Bey. jo [ME x (CA N). , » & 33 The population of ssp. nesopbila is relatively distinct but there are occasionally specimens from Newfoundland that resemble the typical subspecies rather closely. James Bay material frequently approaches ssp. raupii (for example, Lepage 12877 1957] GILLETT—REVISION OF GENTIANELLA 217 ld. GENTIANELLA DETONSA (Rottb.) G. Don, ssp. raupii (Porsild) J. M. Gillett, comb. & stat. nov. Gentiana raupii Porsild, in Sargentia 4:60. 1943. (T.: Porsild 6571!) Sandy river shores, clay banks and salt plains along the Mackenzie River basin at low elevations, with scattered stations along Hudson Bay; flowering from late June to early August; fruiting in August. CANADA: ONTARIO: Raft River, James Bay [?], Spreadborougb s. n. (CAN); Weenusk, Dutilly @ Lepage 16925 (DAO); Swan River, Lepage 31666 (DAO). MACKENZIE DISTRICT: Wood Buffalo Park, Moose Lake, Raup 3026 (CAN); Fort Norman, Kindle s. n. (CAN); Norman Wells, Wynne-Edwards 8564 (CAN); Ramparts, i t Bear La This subspecies must for the present remain a pot-pourri for that heterogenous assemblage of material occupying the Mackenzie Valley and extending to Hudson Bay in the east. The southern boundary of the subspecies is ill defined because the northern prairie provinces and southern districts of Mackenzie and Keewatin are still poorly collected. In the Hudson Bay and James Bay regions some specimens approach ssp. nesophila. As the detonsa complex invaded the regions formerly occupied by the ice the movement was apparently accompanied by a segregation of numerous minor Ee Variants. Variation was also produced by local environmental factors US DUM in a bewildering array of forms any of which may crop up with little regard for geographical continuity. le. GENTIANELLA DETONSA (Rottb.) G. Don, ssp. elegans (A. Nels.) J. M. Gillett, comb. & stat. nov. Gentiana thermalis O. Ktze. Rev. Gen. 2:427. 1891, ex char. (T.: Kuntze s. m.) Gentiana elegans A. Nels. in Bull. Torr. Bot. Club 25:276. 1898. Tr D ond I 539!) Gentiana elegans A. Nels. var. unicaulis A. Nels. loc. cit. 277. 189 (T.: Nelson 4173 ) ee elegans A. Nels. var. brevicalycina Wettst. ex ung 901, : 1, ex char. . 1906. 917. I have followed the general practice of employing the oldest specific SC for my subspecies in accordance with Recommendation 71A, Part 2 of the R (1950 ed. Stockholm). In some cases, however, where doubt exists mër the location or identity of the type of the earliest specific name, I have MS next epithet for which a widely distributed or well-preserved type exists. e epithet elegans has been selected here, rather than thermalis, because 1t 1s more [Vor. 44 218 ANNALS OF THE MISSOURI BOTANICAL GARDEN widely known and because Nelson's specimen is of fine quality and is distributed among at least three herbaria and probably more. Kuntze's specimen has not been found. Rydberg, in making the combination Anthopogon thermalis, remarked "CG. thermalis Kuntze, based on depauperate specimens." Common throughout the Rockies from southern Montana to Wyoming, Utah and Idaho, west to eastern Nevada, south to New Mexico; in mountain meadows and along streams and roadsides; frequent about hot springs; flowering throughout July and August; fruiting in late S ex and September. ete STATES apo: Arc geg Co.: Chromo, Wooton 2910 (US). Boulder Co.: Ward, SE 2741 ae Chaffee Co.: Buena Vista, Maguire & Piranian 12856 (GH, MO). HUNE Valley, Horner s. n. (GH). Conejos Co.: Cumbres Pass, Pages oy H). Custer Co.: Sangre de Cristo Range, Ewan 15390 (GH). Eagle ae? Leadville Pia Eggleston II Mo (US). Garfield Co.: Trapper's Lake, Hermann sat Ca j 2 (GH, MO, US). Middle Park, Grand Lake, Engelmann s.n. (MO) Md Dark: Clokey & Cakes es (MO, US, WS). Grant Co.: Rabbit Ears Pass, Baker 7 (US). Gunnison Co.: Kebler P: Baker 785 (GH, MO). Jackson Co.: Mt. Richtophen, Baker 6 (US). Lake Co.: in Lakes, Wolf 789 (US). La Plata Co oc 3 mann s.n. (US). Larimer Co. ameron Pass, Baker s.n. ( ); Estes Park, Allen 113 (MO). Mineral Co.: Pagosa Peak, Baker 522 (GH, MO, US). Park o.: South Park, were Letterman s. n. (MO). Routt Co.: ese Tweedy 4098 (US). San Juan Co.: Silverton, Popenoe s. n. (MO, PH). San : Trout Lake, Mig ND & Payson P (GH, MO). Summit Co.: Breckenridge, poste hel s. n. (MO, IDAHO: Bannock Co.: Blackfoot River, mg 10002 (US E Bingham Co.: 10 mi w. Gray's Lake, Gulet, Senn & Frankton 6039 (DAO). Fremont Co.: Ei ms Fork, Snake Griet Cantie er s. n. (US); Henry's Lake, mI & rient pus 5 (GH US). NTANA: Carbon Co.: Rock Creek, Elliott 76 (DAO, GH). Gallat : Gallatin basin, Blankinsbip b 351 Lage US); 8 mi. w. Eldridge, Hitcbeock g Mik 1 I 15132 (MO, WS). dison Co.: Beaver Creek, Hitchoork & Mublick 15120 (MO, WS); Madison Forest, ned n. (ID). Park Co.: 2 mi. s.e. Cooke City, Hitchcock & Mublick 13617 Mo c US, WS). Stillwater Co.: Absaroka Nat. Forest, Hitchcock & Mublick 13406 NEVADA: Elko Co.: Ruby Valley, Heller 9531 (GH, MO, NY, PH, US). Eureka : MO). : ax a: UE Park, Standley 14469 (GH, US). Mora Co.: Santa Fe Forest, rank pe 19031 (NY, US). Santa Fe Co.: Santa Fe, Alcott s. n. (MO). Taos Co.: ae ta Canyon, Sts it 2150 (F, WS). Sandoval Co.: Jemez Mts., Goodwin in UTAH: : Wet ee? Vi » Bacigalupi GH). Grand Co.: La La Sal Mts., Payson & P. "Zeg " 4075 (GH, MO). " wee DÀ E diet RM 15 mi. e. Cedar City, Hitchcock, Retbke, & Van R aadsbooven 262 (0 (US, Ws). eid Co.: 10. mi. W- er Dale, Maguire & Richards 15931 (GH, ey i Sanpet te Co.: of re bay Pi Tide- SE 1887 (US). ed Co.: Fish Lake, ose Su (US). Sum à ME Peak, Jones s. n. (MO, US), Payson & PA 08 (GH, vor KE NE Mountains Clos 1 LS . Wayne Co.: Rabbit Valley, Ward 607 (MO, Die. : ee Medicine Bow MO, ) pen ulc gem 007 (GH US); Cenc Nelson 8 H, MO, US); Telephone Mines, Nelson 7888 (GH, MO, ya De peus Co.: ro la Mts., eise dre Own- bey 325a (WS). Lincoln Co.: Iiic MEO 5175 (DAO). Park Co: Shoshone (GH, MO). Sublette Co.: Wind River Mts., Forwood s. n. (US): Survey CUL ch (GH, MO). Teton Co.: zeg Lake $ Merrill & Wilcox 1082 (GH, US). Yellowstone National Park: Lower Geyser B sin, Thompson 14159 (GH, MO, US, WS); Upper r Gey- ser Basin, Mearns 473 (US); Eege Lake, Nelson & Nelson 6764 (GH, MO, US) 1957] GILLETT— REVISION OF GENTIANELLA 219 1f. GENTIANELLA DETONSA (Rottb.) G. Don, ssp. superba (Greene) J. M. Gillett, comb. & stat. nov. Gentiana macrocalix Lex. in La Llave & Lex. Nov. Veg. Descr. 1:19. 1824, ex char. Gentianella macrantba D. Don, ex G. Don, Gen. Syst. 4:179. 1838, ex char. Gentiana ciliata Moc. & Sessé, ex G. Don. loc. cit. 119. 1838, nom nud. in syn Gentiana macrantba (D. Don, ex G. Don) Griseb. in DC. Prod. 9:102. 1845. enliana serrata Gunn. var. grandis A. Gray, Syn. Fl. N. Amer. 2:116. 1886. C 1 8!) Wright 1 Gentiana superba Greene, Pittonia 1:155. 1888. (T.: Forrer 44!) Gentiana grandis (A. Gray) Th. Holm, in Ottawa Nat. 15:110. 1901. As explained under ssp. elegans, I have tried to select as subspecific names the earliest ones available, or those most widely known (although not obliged to do so). I believe this course more reasonable than to propose entirely new names. In ssp. superba the earliest name was not selected because of difficulty in the identity or location of types. The type of Gentiana macrocalix Lex. is given as "Habitat prope Vallisoletum et Irapaeum. Floret Novembri—Lex." The type has not been seen. "I[rapaeum" is probably Uruapan (del Progreso) in Michoacán. The description is not suffi- ciently clear to enable positive identity with the north Mexican population. The actual type locality may be within the range of ssp. lanceolata. The next available name, macrantha, also exhibits difficulties, for the type is a Sessé & Mociño collection. The type has not been seen but a photograph at Chicago and a Sessé & Mociño sheet at the Gray Herbarium yield no locality data. Accord- ing to Hemsley (Jour. Bot. n. s. 8:275. 1879), Drs. Parry and Palmer later col- lected in the State of San Luis Potosí in essentially the same region as did Mociño & Sessé, The exact locality is still uncertain but still may be somewhat outside the range of this subspecies. The name grandis is avoided because of the possibility of confusion with Genti- ana grandis H. Sm. (Anz. Akad. Wiss. Wien, math.-nat. Cl. 63:100. 1926). Greene's name for which the type Forrer 44 (UC) plus two isotypes (F, GH) have been seen and for which the locality is known is thus emp oyed. G. macro- calix and G. macrantha are placed in synonymy provisionally. Arizona in the Huachuca and Santa Rita mountains, in the Sierra Madre of Chihuahua south to Durango and Zacatecas at altitudes of 2000—2600 meters; in moist areas and roadsides, steep open slopes, and in open pine forest; flowering from September until late November. Se techn Co.: Huachuca Mts., Lemmon 2823 (PH, US), 2883 (GH); Fort Huachuca, Wilcox 489 (US). Santa Cruz Co.: Santa Rita Mts., Peebles 6 Harrison 2055 PLY core s. n. (US). , x D CHIHUAHUA: e near Huajochic, Lumboltz s.n. (GH); Sierra Madre, Moun Valley, e S) oi near Guachochi, Goldman 173 (GH, US); San Julian, Nelson 4942. ( Gud reegen ichi, Rio Mayo, Gentry 2720 (F, GH, MO, UC); 10 mi. s.e. Colonia Garcia, & Barber 320 (F, NY); south Chihuahua, Zingg 5. ?- (F). [Vor. 44 220 ANNALS OF THE MISSOURI BOTANICAL GARDEN DURANGO: Sierra Madre, west of Durango, Forrer 44 (F, GH, UC); Rosilla, Collins & Kempton s. n. (US); El Salto, Martinez s. n. (F). SONO Wrigbt 1658 (GH, MO, NY, PH, US); Batocomori and Santa Cruz, Thurber 926 (GH, NY). ZACATECAS: Sierra de Valparaiso, Goldman 6 (GH, US). LOCALITY UNKNOWN: Sessé & Mociño 685, 1371 (F), s. n. (GH, photo F). 1g. GENTIANELLA DETONSA (Rottb.) G. Don, ssp. lanceolata (Benth.) J. M. Gillett, comb. & stat. nov. Leianthus lanceolatus Benth. Pl. Hartw. 24. 1839. (T.: Hartweg 209!) Gentiana lanceolata (Benth.) Griseb. in DC. Prod. 9:102. 1845. Mexico from perhaps southwestern Chihuahua in the Sierra Madre to Michoa- cán, and in Vera Cruz about Cofre de Perote; altitude unknown; flowering from early September to late November; fruiting during November and December. EXICO: JALISCO: Bolanos, Hartweg 209 (GH, photo DAO, MO). GUANAJUATO: Duges 2 e iesu A Bag Kach 3941 (F, GH, MO, NY, US); Morelia, Arsène s.n. SINALOA: Cueva del Diablo, Dehesa 1562 (US). ZACATECAS: Sierra de los Morones, Plateado, Rose 2716 (F, GH, NY, US). Subspecies lanceolata is distinguished chiefly by the entire corolla lobes, rela- tively slender flowers, and linear leaves. In Zacatecas, where the range overlaps that of ssp. superba, the flowers are larger and leaves broader. Since the material at and is scant and extremely variable, it is difficult to elaborate further on this apparently clinal effect. However, such observations as have been possible indicate a low genetic barrier separating these two subspecies. The north-south trend of variation and the overlapping range seem to justify the recognition of this popula- tion merely as a subspecies of G. detonsa. Mention should be made of a specimen collected by E. K. Balls (5438, US!) from Cofre de Perote, Vera Cruz. This specimen has shorter, more linear leaves than the other specimens of ssp. lanceolata and the corolla is more narrowly funnel- form and attenuate to the pedicel. One interesting feature is the habitat, which Balls describes as: "bog, among rushes, etc.” Since the slight morphological dif- ference may be due only to environmental conditions, and because the range of variability is high in this group, no taxonomic recognition is given here to i collection. th. GENTIANELLA DETONSA Gillett, comb. & stat. nov pre diens Gunn. var. holopetala A. Gray, Bot. Calif. 1:481. 1876. (T.: Bolander Gentiana bolopetala (A. Gray) Th. Holm, in Ottawa Nat. 15:110. 1901. Sierra Nevada Mountains in open wet meadows, hillsides, gravelly places and about springs in the Canadian and Hudsonian life zones, at altitudes of 2600-3300 meters; flowering from early July until about the end of August; fruiting in late August until the middle of September. (Rottb.) G. Don, ssp. holopetala (A. Gray) J. M. 1957] GILLETT—REVISION OF GENTIANELLA 221 Bolander 6359 (GH, MO, US). Mono Co.: Bloody Cañon, Chesnut & Drew s. n. (US); mora Pass, Wiggins 9522 (CAN, GH, WS). Mono-Madera Cos.: Yosemite Nat. Park dary, Evans s. n. (F). Placer Co.: Lake Tahoe region, Smith s. n. (F). Plumas Co.: HS (GH). Sierra Co: mmon s.n. (GH, PH). Tulare Co: Nat. Bridge Meadow, Culbertson 4260 (GH, F, MO); South Fork, Kern River, Rotbrock 313 (MO, PH, US). Tuolumne Co.: Tuolumne Valley, Brewer 2847 (MO, US). NEVADA: Boat Corral, Coville & Funston 85 (US). Gentianella detonsa ssp. bolopetala is easily recognized by the long gynophore, entire corolla lobes, and the usually acute cauline leaves. Since these characters are variable, it seems advisable to consider the Californian population at a rank lower than species. Furthermore, occasional specimens superficially are very similar to specimens of subspecies detonsa from Alaska. The one Nevada collection seen resembles ssp. elegans, so that these subspecies may also intergrade. The Nevada collection lacks adequate data to enable it to be plotted. An interesting character of ssp. holopetala is the almost hyaline wedges in the corolla tube below the sinuses, which in many herbarium specimens appear to be somewhat slack, rather suggesting the plicae between the lobes in Gentiana. For the type. 2. GENTIANELLA CRINITA (Froel.) G. Don, Gen. Syst. 4:179. 1838. Gentiana crinita Froel. Gent. Diss. 112. 1796, ex char. Annuals (biennials or winter annuals?), 1-6 dm. tall, branched from the base or above, or simple. Basal leaves lingulate to spatulate, the apex rounded, obtuse or acute, the base attenuate to the stem, 0.8-1.6 (-3.3) cm. long, 0.1-0.6 (-1.0) cm. wide, soon withering or deciduous; median leaves linear, linear-oblong, lanceolate- Ovate to ovate, elliptic in small forms, usually acute, the base attenuate to the width of the stem, or rounded or subcordate and clasping, 1.0-8.0 cm. long, 0.1— 2.0 cm. wide. Flowers few to very numerous, or solitary, —— 54 Jepson, Fl. Calif. 31:88. 1939. the terminal pedicels j [Vor. 44 222 ANNALS OF THE MISSOURI BOTANICAL GARDEN 2—22 cm. long, those of the branches usually shorter. Calyx narrowly to broadly funnelform, the tube 8-15 mm. long, 5-15 mm. wide below the sinus, the inner lobes ovate-triangular, subulate to acute, the outer lobes longer, lanceolate and conduplicate, the margins frequently carinate near the tip, acute, longer than or equal to the inner lobes; keels 0.7—1.0 mm. wide, scarcely to strongly prominent, weakly to strongly papillose (under magnification), dark green to purple-tinged; sinal membrane slightly to strongly curved and bearing few to many blunt pro- cesses up to 0.3 mm. long. Corolla pale to deep blue, rarely white, narrowly to broadly funnelform, 25-60 mm. long, 8-15 mm. wide at the orifice, the lobes oblong to obovate-oblong, the tips denticulate to short (1 mm.) -ciliate, the margins with fimbriae to 5 mm. long in the upper half. Stamens included, the filaments inserted in the lower third of the corolla tube, the marginal wings 1.3-3.0 mm. wide at the base, tapering above, strongly incurved; anthers 2-5 mm. long, about 1.5 mm. wide, attached in the upper third; interstaminal glands oval to rounded, prominent and frequently deep green. Pistil sessile to short-stipitate, the gyno- phore 2—4 mm. long, becoming obscure with age, ovary fusiform, 12-25 mm. long, 2-5 mm. wide; stigmas rounded to flabelliform or reniform, 1-3 mm. wide. Cap- sule as long as the corolla, dehiscing in the upper third, the tips of the valves slightly recurving. Seeds oblong and angular, light brown, the surface covered with rounded to elongate inflated brown papillae, particularly concentrated at the angles and ends. North American, extending from the Appalachian Mountains and the Gulf of St. Lawrence in the east to the Great Lakes region and westward across the plains to the Rocky Mountains in Alberta, resolving into several intergrading subspecies apparently of a clinal nature and with numerous minor local variants. Habitats various according to the subspecies, but in general preferring alkaline areas of about pH 6.7; flowering and fruiting roughly from June to late September; at usually low altitudes from sea-level to a few hundred meters. Froelich provided an adequate description of Gentiana crinita but cited no specimens. In synonymy Gentiana ciliata L. (Syst. Pl. 1:645. n. 27) “a var. ameri- cana" is given. By “var. americana" Froelich apparently referred to the American element from which Linnaeus drew his description. At this time the problem of selecting a type is deferred because of the difficulty of obtaining certain literature and photographs. Because Fernald? interpreted the American portion of the Gen- tiana ciliata L. (Sp. Pl. ed. 2, 334. 1762) to be his G. victorinii, the difficulties are increased. 55 Fernald in Rhodora 25:85-89, 1923. 1957] GILLETT—REVISION OF GENTIANELLA 225 KEY TO THE SUBSPECIES a. Flowers 25-60 mm. long; corolla lobes with lateral fringes 2-5 mm. long in the upper half of the lobe, the tips frequently short-ciliate also; calyx keels 0.1-0.8 mm. broad at the base. b. Median leaves ovate to lanceolate-ovate, rounded to subcordate at the base, linear to elliptic only in the small phase. North Carolina to Massachusetts and Maine west to Minnesota and Manitoba 2a. G. crinita ssp. crinita bb. Median leaves linear to linear-lanceolate. Upper New York through the Great region to Minnesota and Manitoba . G. crinita ssp. procera aa. Flowers 10—40 mm. long; corolla lobes with short lateral fringes, the tips dentate or erose, rarely ciliate; calyx keels absent or not prominent. c. Flowers 30-35 mm. long; corolla lobes orbicular; style 1.0—1.5 mm. long; stigmas flabelliform. Intercotidal zone of the St. Lawrence River. Quebec.....2c. G. crinita ssp. victorinii cc. Flowers 25-40 mm. long; corolla lobes usually oblong, somewhat truncate; s igmas Steg sessile. Gaspé, Quebec, and James Bay, southward to North Dakota, westward o erta 2d. G. crinita ssp. macounii The subspecies of G. crinita overlap in distribution a great deal so that many individual collections at the margins of ranges are difficult to assign. A general clinal condition exists between Alberta and the Appalachians. The eastward ex- treme has ovate to ovate-lanceolate leaves, long lateral corolla lobe fringes, long corollas, long-papillose seeds, coarsely papillose calyx keels, relatively distinct gyno- Phores, and stems branching in the upper part. The western population consists of slender plants with linear leaves, shorter lateral fringes, small corollas, short seed papillae, less papillose keels, sessile capsules, and the stems branch chiefly from the base. Although no sharp division exists there is a change along certain lines that is useful for separation into subspecies. On the east side of the cline, ssp. brocera differs from ssp. crinita principally in its linear leaves, although the flower characters seem to remain much the same. Other characters change gradually to merge with the plains subspecies. The western ssp. macounii is reserved for the short-flowered members, although occasionally long-flowered individuals appear. The key, of course, should be regarded as a guide only. 2a. GENTIANELLA cRINITA (Froel.) G. Don, ssp. CRINITA Gentiana ciliata L. Syst. 1:645. 1756, in part, as to American element. Gentiana fimbriata Andr. Bot. Rep. 509. 1808, ex ic. Denckea crinita (Froel.) Raf. in Med. Repos. II. 5:352. 1808. Anthopogon incarnatum Raf. New Fl. N. Am. 4:90, 1836, ex char. In the Appalachians from western Nort northwest to eastern Ontario and western Quebec, forming a cline with G. crinita ssp. procera and with ssp. ™4 as far west as Minnesota and Manitoba. In a variety of habitats: roadsides, clearings in open woods, swampy ground, gravel, sand, railway embankments, and shallow soils over limestone, in general Preferring calcareous habitats. In the western part of the range found in open oak Woods, on damp prairies, and boggy places; flowering throughout August to early October; fruiting from about mid-September until late October. west to the Great Lakes region, counii, and occurring [Vor. 44 224 ANNALS OF THE MISSOURI BOTANICAL GARDEN CANADA: MANITOBA: Ochre River, Groh s.n. (DAO); Stony Mountain, Groh s.n. (DAO, MO), Macoun s. n. (CAN, F); Beausejours, Schaeffer s. n. (PH); Burnside, McMorine s. n. (DAO); Clear Lake, Riding Mt. Nat. Park, Jackson 29-18 (DAO); Winnipeg, Fowler 1 (DAO); Dauphin, génge I0503 (CAN); Grand | Rapids, Lake Winnipeg, Scoggan 4822 (CAN); Vivian, Lüve 52 65 Sa CAN, JB ONTARIO: Bruce sett d Mulligan 194 (DAO). Carleton Co.: Britannia, Macoun s.n. (CAN JT , Dow's Swamp, Gillett 6055 (DAO), Kellett s.m. (CAN, DAO, MO); oA Gillett 6056 (DAO); Lanark Co. res der: = Li GE? 47-1018 (DAO); Harwood Plains, Dore & Rhodes 12868 (DAO gi t. Thomas, Fisher s.n. (MO). Essex Co.: Turkey Point, Lake Erie, Soper yo (DAO), Grenville Co.: perenta Terrill 53 (DAO). Middlesex Co.: London, Burgess s. 3. (CAN, DAO). Peterborough Co.: Clarina, Dummer pet Dore & Heind 14062 (DAO). Stormont Co.: dee Point, Dore & Van Rens 15667 (DAO). Waterloo Co.: Galt, Herriot s. n. (CAN, DAO); German Mills, Montgomery 298.39 (DAO). Welland Niagara Falls, ach 314 x m Davids, Scott 123 AG Wentworth Co. Sire ae Cody 105 (DAO). York Co.: Toronto, Pennell 13124 (PH). BEC: Huntingdon Co.: ne Tao & Kucyniak. f ^ (DAO, JBM). eeng STA ee EE 5 field Co.: Stratford, Eames s. n. (GH). Hartford Co.: Southing- s Bissell 146 (M Litchfield Co.: South Canes. Greenman 1446 (GH, MO, u > Us). Co.: Sout: Kennedy s.n. (GH). New Haven Co.: New Haven, Safford 25 ILL Co.: Babcock s.n. (US); Chicago, Blankford s.n. (F). Lake Co.: rakean. ae n. i Stephenson Co.: dcin obnson s. n ? INDIANA: : 22064 (GH). e Co.: ore Deam 15057 (GH ake Co.: Clarke, Dune 3970 (F, GH). Porter Co.: Dune Park, Chase 2115 (F, US), Gillett 6058 (DAO). St. Joseph Co.: Churchill s. n. (MO); Chain GE Nieuwland 11650 (MO, US). Steu un Co.: Graveyard Lake, Deam. s.n. (MO, kee Co.: Postville?, Schutz 182 (GH). Johnson Co.: Fitzpatrick © hig s. m. .. (F, GH e Cumberland Co.: Falmouth, Blake s.n. (PH); Cape Elizabeth, Gayle 904 ` : ; West Harpswell, Read 3492 (MO). Kenn sc E Co.: MARYLAND: Baltimore Co.: Gresham s. n. (US), Morris s. n. (PH). EC GE Co.: Chahi Churchill s.n. (GH). E ai Worthin (US). Bristol Co.: Nonquit, Sturtevant s. n. (MO). Essex Co.: Bowford, Mar rdoc. us 5264 4 (F). MORS Co.: Buckland, Forbes s. n. (GH). I : West Granville, Seymour 362 (GH, MO). i ES . , sos T (US). Middlesex Co.: re E E ). Plymouth Co.: Kingston, Ridler s. ^. (F). e UST : Boston, Faxon & Faroe s.n. (GH). Toce Co.: Bolton, Churchill s." MICHIGAN: Berrien County: Millspaugb 38 Pi. G Co.: Alma, Davis (US). Jackson Co.: Camp & p s.n. (F B. Dë e eges idland, Las Wap" ge KEE Co.: Geddes, Farwell 7604 (GH). Wayne Co.: Dearborn, Chandler s.n MINNESOTA: Clea Co.: head pH). Fill- more & Olmsted Cos.: Chatfield, Hie 2: EE e Be Tti Sing. Mearns s.m. (US). Houston Co.: Freiberg s.n: (MO). Morrison Co.: Sandberg 877 (U US). Geyer s.n. (MO, US). Todd Co.: Philbrook, Hotchkiss Y s. n. (US). HAMPS, knap Co.: Gilmanton, Gilbr th s. GH Cheshire Co Marlboro, Re h SA GH). Grafton Co.: Pih: Kees à (GH). pil- Pelham, Todel s.n. (GH). Rockingham Co.: Derry, Batchelder s. n. (MO): 1957] GILLETT—REVISION OF GENTIANELLA 225 NEW JERSEY: Bergen Co.: Tenafly, we mg n. (GH). Cape May Ge ean View, Meredith s.n. (GH). Essex Co.: ` Lighthie v s.n. (MO). Lake Re: Fisher s. n. (MO, US). a Co.: Watchung, Moldenke 7377 em. N). Sussex Co.: Egypt Mills, Bartram s. n. (PH). ‘Union Co.: Summit, Parker w york: Albany Co.: Albany, Nt. s.n. (NY). Broome Co.: Glenwood Ravine, Millspaugh s.n. (F). Columbia Co.: Canaan, Harrison 11 (US). Dutchess Co.: Clove, Standley & — ied pe^ Erie Co "Buffalo, Clinton s. n. (PH). Jefferson Co: Os sn. (US). Tompkins Co.: Fall Creek below Varna, Eames & MacDaniels 4827 (GH). Washington Tripoli, West Fort Ann, Burnham s.n. (GH). Westchester Co.: m Mohegan, ee 9328 (PH NORTH ta AROLINA: Macon Co. ntahala Mountains, SCH ris OHIO: Erie Co.: Milan Tow dus " . Mosele s. n. (F es Co. e, ownship y (F). PENNSYLVANIA: Berks Co.: Small s.n. (F). Bucks Co.: Bridgeton, Bright 11036 WS). Centre Co.: Scotia, Kelly & Kelly s. n. (GH). Chester Co.: Rothrock s. E ). (G n. (NY). Pontius & Bartley , US) 33914 (GH). Luzerne Co.: Wilkes-Barre, Thurston s.n. ( Summit, Small s. n. (F, PH). Montgom e Co.: Greenlane, KE 279 (GH). N hampton Co.: RAS Dowell 1 (US). Philadelphia Co.: Wissahickon, et sm. (GH). York Lon Loganville, hails 349 (MO). Wyoming Co.: Oste. Glowenke 9370 (GH e — Providenc ce Co.: East nes al dag 0656 (GH). Bennington Co.: Shaftsbury, Knowlton s.n. ast Rutland Co.: West Rutland, e s. 1. (GH); Ira, SEH s. 9. (F): indham Co.: Whittingham, t. John 421 (PH, US). Mire Geen? Greebbriér hee Lewisburg, — s.n. (GH). Ss Brown Co.: Fort Howard, Shuette s. n. (F, GH). dee Co.: Prairie NSIN: du Chica, Hale s. n. (MO Mc Co.: Spa "sis Palmer 28466 ). Racine Co.: ips Davis s. $ e Ree wie Gilbert & Gilbert s.n. (GH). Wal- . (GH). Co.: Lake ( Gela; Sud $ 9X (E). Because G. crinita ssp. crinita has been observed and collected in the field, a brief Juste: of its habitat and mode of growth is introduced here During September, 1951, a small colony near the village of bert Carleton unty, Ontario, about 20 miles west of Ottawa, was kept under observation. The Colony was restricted to an area about one-eighth of a mile long and a few hundred Yatds wide, and was further restricted to sballow soil over limestone. Plants were rather abundant, and grew in open Thuja occidentalis-Populus tremuloides woods. Halenia deflexa also was found here but was restricted to a slightly drier micro- habitat situated in clearings in Thuja woods. This microhabitat was termed a "cedar glade” in contrast to the more open and wetter habitat of G. crinita. In direct sunlight, corolla lobes of G. crinita were observed to spread at an angle of about 45° but were closed in shaded places. The color of the lobes varied from à purple-blue to slightly reddish. The only insects observed about the flowers were a few ants and flies although thrips were found within the corollas. (Wee 44 226 ANNALS OF THE MISSOURI BOTANICAL GARDEN The frequency and distribution of small, depauperate individuals within this population were of particular interest; for such plants had not only been observed on herbarium sheets, but had been deliberately selected out and described as forms and varieties by early authors. These small plants are about 5-20 cm. tall, and usually bear a single flower considerably smaller than those of the well-developed individuals. Moreover, in comparison with well-developed plants, the leaves are smaller and elliptic rather than ovate. In our colony, small plants were scattered among larger plants without any apparent preference to microhabitat. No definite explanation can be given at this time for the occurrence of these plants, but it is suggested that they are seedlings produced by early-flowering individuals during the same year. They are of interest, too, because they also occur in populations of other taxa in various parts of the continent. Large numbers of them were seen m Yellowstone Park, Wyoming, interspersed within a population of G. detonsa ssp. elegans, particularly. n an effort to determine their life span, large plants from this colony were tagged and observed again the following year. In this region they prove to be annuals. No over-wintering rosettes were observed. 2b. GENTIANELLA CRINITA (Froel.) G. Don, ssp. procera (Th. Holm) J. M. Gillett, comb. & stat. nov. Gentiana barbata B Browniana Hook. ex Macnab, in Edinb. New Phil. Jour. 19:62. 1835. T.: Macnab s. n., photo MO!). Anthopogon virgatum Raf. Fl. Tellur. 3:25. 1837, ex char. From upper New York through the Great Lakes region to Minnesota, North Dakota and Manitoba. Clinal in nature, closely resembling ssp. crinita in the eastern part of the range except for the narrower leaves, approaching ssp. macoumi in the western part, with shorter cilia, more linear leaves, lax habit and smaller seed papillae. In the western part including those individuals with coarse stems and large flowers. CANADA: MANITOBA: Aweme, Criddle s.n. (DAO, MO); Bird’s Hill, Winnipeg, Denike 1759 (DAO); Ciento, Dore 9196 (DAO); Carroll, Senn t$ Gordon 3129 (DAO); Ochre River, Scoggan 10508 (CAN) ; Stony Mountain, Macoun s. n. (CAN, F) em ONTARIO Hay Bay, Krotkov 7212 (GH). Gray Co.: Craigleith, VETE SC Huron Co.: Banks of the Maitland River, Mac (MO); SR ) Wolsley Lake, Thompson s.n. (UBC); Fishing Is. M ECH Whitefish Is., Macoun s. n. (CA : suns d Eaa e ae ; Ogden 2395 (US); Manitowaning Is., Og arbour, yon s.» ( Simc ++ Collingwood, Victorin, Rolland & Meilleur 45073 (DAO GH). Waterloo Co.: Galt, Herriot s. n. (DAO, MO). UNITED STATES: Kan ILLINOIS: Cook Co.: Babcock s.n. (MO). Du Pa 3 o8 (GH). à : i ge Co.: Moffatt 508 ( & Cook Cos.: Elgin, Sherf 1986 (MO). Kane Co.: Kankakee, Greenman 3538 (MO): Sarnia, Macoun s. n. (CAN). 1957] GILLETT—REVISION OF GENTIANELLA 227 Lake Co.: ice es) 6050 (GH, MO). McHenry Co.: Ringwood, (PH). McLea rmal, Vasey s. ». (GH). Winnebago Co: eg wi Bebb et ; H). INDIANA: Fulton Co.: Bruce Lake, Deam 46341 (GH). Kosciusko ye vro Lake, Friesner estes (DAO, WS). Lake Co.: Pine, Chase s.n. (GH). La Porte Co. Mill Creek, Friesner 14190 (NY). Marshall do Lake Maxinkuckee, wi EI Gei 1436 (US ). St. Joseph Co.: Chain Lake, Nieuwland 11642 (MO). Wayne Co.: ome — Nieuwland a n. (US). Co.: Mason s RS SUME s.n. (MO). Dickinson Co.: West Okoboji. Lake, ‘Shimek s s.n. (GH, MO, US, WS). met Co.: Cratty s.n. (PH). Palo Alto Co.: sie i Hayden I0I9I (GH, Mo. NY). Winneshiek Co.: Fremont Twp., Shimek s. s.n. (M MICHIGAN: Alpe : Alpena Point, Lake Huron, Cain, Raymond & Kucyniak 936, 943 (JBM). teen So Albion, Barr s. n. (F). Cha rlevoix Co.: Beaver Island, Michigan, Gillman s.n. (NY). Cheboygan Co.: Mackinaw City, Hermann 7311 (F, MO, S, WS). Emmet Co.: Wilderness State Park, Cain, Raymond & Kucyniak 600 (JBM). lonia Co.: Hubbardston, Wheeler s. n. (US). pere Co.: Lyons hes i ad es (F). Kent Co.: Grand R Rapids, Crozier s.n. (US). Mackinac [c ty Bay, L Huron, Ehlers 4905 (F, US). wee nd Co: Orian Twp., Chandler s. n. (US). Sc Rat craft Co.: Manistique, Cain, Raymond S Kucyniak 565 (DAO, JBM). St. Clair Co.: Port reso, Dodge s. n. (NY). pense fie Co.: Cavanaugh Lake, Grass] 7745 (MO). Wa : Detroit, Schott s. n. (F). in hernia Cass Co.: Big Thunder Lake, Richards 356 (F). Clay Co.: Muskoda, Ballard 3064 (GH). Clearwater Co: Itasca Park, Grant 3362 (GH, MO, US). Dakota Co.: Nichols, Rosendahl 4757 (F). Hennepin Co.: Fort Snelling, Mearns s. n. (US). SACO Co.: Richdale, Chandonnet s.n. (MO). Pope Co.: Ewen, Taylor "n O a, Holzinger s. n oi ORK: SE Co: Genesee River, Eaton n. (NY). Niagara Co.: Niagara Falls, Canby D NY); Goat Island, Niagara, Eheim s.n. (MO). St. Lawrence Co.: Pollys Creek, äeren & Maguire 1315 (GH). uam RTH ca KOTA: he Co.: Anselm, Stevens 227 (F). anor Urbana, McFarland 4507 (US). Clark Co.: Tremont City, Beate? aud (08). Sege Co.: Deer Creek, nr. Williamsport, poh pes unknown oul : Kingston, Bartley 6 Posies 7 . NSIN: "Zei Co.: Fort Howard, Schuette s. Dane Co.: i Heddle en dg Door Co.: Bailey's Harbour, Fassett e H P ip T rasura Co.: DAO, GH, MO). e Co.: Hasse . (NY). Waukesha Co.: Lapbam ( Delafield, dee s.n. (GH). 2c. GENTIANELLA CRINITA (Froel.) G. Don, ssp. victorinii (Fern.) J. M. Gillett, comb. & stat. nov. Gentiana victorinii Fern. in Rhodora 25:87. 1923. (T.: Victorin 16073!) Quebec, restricted to the intercotidal zone of the St. Lawrence River from Des- chambault to St. Jean Port Joli; flowering from late July through August until early Sang fruiting from mid-August until mid-September. Aer SS m g , Ro al Ei Leet 45546 (G . Levis uld, Victorin, Rollan de ns I3 E magny Co.: cb. Rousseau 21057 (GH), € 8 p ee (MO, WS), Rousseau foi (CAN). [Vor. 44 228 ANNALS OF THE MISSOURI BOTANICAL GARDEN d'Orleans, Sainte-Petronille, Victorin & Rolland 49500 (GH), Gillett 8 Van Rens 9704 (DAO); St. Laurent d'Orleans, Gillett & Van Rens 9722, St. Jean, Gillett & Ze Rem 9723, eastern end of Ile d'Orleans, Gillett t$ Van Rens 9724 (all DAO). Portneuf Co.: Deschambault, Gillett & Van Rens 9668 (DAO); Neuville, Gillett & Van Rens 9693 (DAO); Saint-Augustin, Victorin, Rolland & Jacques 33868 (GH). Quebec Co.: Cap Rouge, Victorin 16073 (GH, JBM), Gillett & Van Rens 9696 (DAO), Victorin, Seini Raymond & Kucyniak 3704 (CAN, DAO, GH, MO, WS). And many more sheets from lities. One of the most distinctive features of this subspecies is its unique habitat described in such detail by Raymond”? and Rousseau. During the summer of 1955 a visit was made by the author to this area. Both sides of the St. Lawrence were visited and a circuit made of l'Ile d'Orléans. From my observations it is suspected that the plants are annuals although Raymond cautiously writes: "Toutes ces espéces sont annuelles ou bisannuelles.” The water throughout the range is fresh but may be salt in the eastern part following persistent east winds. As a result of daily flooding by the tide, the plants usually are covered with debris or a film of mud. A cline in flowering time was observed extending downstream from Deschambault on the north shore to Berthier and Montmagny on the south shore where the plants were in fruit only. The small phase seen in other populations occurs here also but is less common. 2d. GENTIANELLA CRINITA (Froel.) G. Don, ssp. macounii (Th. Holm) J. M. Gillett, comb. & stat. nov. Gentiana macounii Th. Holm, ar, in Ottawa Nat. 15:110. 1901. (T.: Macoun s. Gentiana detonsa Rottb. 8, 9.1 n.!) tonsa Lunell, in Bull. Leeds Herb. 2:7. 1908. (T.: Lumell Antbopogon tonsum (Lunell) Rydb. in Bull. Torr. Bot. Club 40:463. 1913. Antbopogon macounii (Th. Holm) Rydb. loc. cit. 463. 1913. Anthopogon brocerum Lunell (incorrectly attributed to Th. Holm by Lunell), var. fonsum Lunell, in Am. Midl. Nat. 4:507. 1916. (T.: Lunmell s. n. T.: Anthopogon procerum var. tonsum Lunell f. uniflorum Lunell, loc. cit. 1916. (T= Lunell s. n.!) PE Gentiana gaspensis Vict. in Contr. Lab. Bor. Univ. Montréal 20:10. 1932. (T.: Victorin, Rolland & Jacques 337511) Gentiana tonsa (Lunell) Vict. loc. cit. 14. 1932. Alberta, Saskatchewan, and Manitoba, southern James Bay, and in Gaspé, Quebec; southward in the Great Plains through North Dakota and Minnesota and transitional to G. crinitg ssp. procera west of the Great Lakes. About the margins of sloughs and along river banks and meadows, occasionally in sandy swales and calcareous bogs. Flowering from late June to late August; fruiting during August and into September, Sê Raymond, M., in Nat. Canad. 78:81-87. 1951. 57 Rousseau, J., in Contrib. Lab. Bot. Univ. Montréal 23:1-7, 1932. 1957] GILLETT—REVISION OF GENTIANELLA 229 ANADA: ALBERTA: Lees Creek, Cardston, Macoun s. n. (CAN, TYPE) ; Fort Pitt, Saskatchewan River, Macoun s.n. (CAN); Ba nff, Macoun s.n. (GH, MO), McCalla 2175 Ph =o Sanson in 1899, Fletcher in 1808 (CAN); Red Deer, Grospe (?) (WS); Calgary, Turner 3759 (CAN); Shaganappi, Moodie 44 (F, NY); Jasper, Scamman 3357 (ED; Crates Creek, Macoun & Herriot s. n. (CAN, NY); Blindman River, Budd 1839 (SWC); Bow eng T, we s.n. (CAN, US); Fort Saskatchewan, Turner 2720 (CAN, DAO); Ed- nton, Moss 2412 (GH); Macleod North, Dixon 1533 (DAO); Lacombe, Lindsay & La? t 339 (DAO). TISH COLUMBIA: Field, Macoun s. n. (CAN, NY). Ear Aweme, Criddle Sem. (DAO); Bird's Hill, Vin nore Denike 1567 (DAO); Ochre River, Grob O); Little Saskatchewan River, Scoggan te (CAN); Lake Winnipegosis to Geer Lake, Scoggan 4639 (CAN); Verden, Groh s. n. E Stony Mountain, Macoun s. n. (F); Duck Mt., Halliday 90 (CAN); Broomhill, Sen : WO. 3116 (DAO); Shoal. ta Garton 3165 (CAN); Bellhampton, Bolton Ge Bay : Rupert House, Macoun s.n. (CAN). Bonaventure Co.: Bona Raymond 4008 (CAN, DAO, GH, MO, US, VS), Meri Victor et al. 33751 (CAN, GH, JBM, NY), Scoggan 1231, 981 (CAN), N). We K : McKague, Breitung s.n. (DEO); Nipawin, Breitung 59908 (DA MO), 354 (CAN); e Saskatchewan River, Boivin © Breitung igo (DAO); SC SC eg sn. (D 2 Canora, Carmichael 1 DAO); Prince Albert, Fraser G M e, g- ew: 442 (DAO); Pike Lake, Russell s. n. (DAO); Saskatoon, Fraser s. n. (SWC) ; Elbow, Campbell s. n. (SWC), —. (SWC); ; ead & Russell $1235 (DAO); Beasant, Moose Lake, Richardson s. n. (GH); Prongua, (CAN). Ri Jaw Sen ee 570 (DAO); Blackfoot Crossing, Bow River, Macoun s D STA y e River, ted 54 154 (GH, PH, US). NOR E" DAKOTA: ie? : Butte, Lunell s.n. (MIN, MO, NY, US). McHen me SC s. n. BN: ; Wrigbt s.n. (NY). Rolette Co.: Gravel his Mabbott 485 ( Celilo: E: EN has been included within ssp. macounii since the differences are such that they easily come within the range of variation of the remainder of the cL MR The Gaspé specimens have slightly smaller flowers and branch from base. However, nearly identical material occurs along the southern shore of James Bay, although I lack the habitat data necessary to ascertain whether it is confined to the same brackish meadow habitat as in the Gaspé plants. One or two collections from Manitoba also are exceedingly difficult to separate from the Quebec specimens. The occurrence of such plants seems to provide further evidence of a link with the western prairie population.5* geo k y e papillose calyx keel ve manh ked apara dl but a de dec cited, SUI fed dhat he aio kee fringed gentians and the smooth-keeled group form two natural divisions that should be mainta at the specific level at ak pss the present. D [Vor. 44 230 ANNALS OF THE MISSOURI BOTANICAL GARDEN Although the type collection of subspecies macounii comes from Cardston, AL berta, it is not typical of the majority of the population. Alberta material is more frequently branched from the lower part than Saskatchewan and Manitoba material. This characteristic must be considered a local variant within the population as a whole since other characters either are identical or come within the variation range of the subspecies. 3. GENTIANELLA barbellata (Engelm.) J. M. Gillett, comb. nov. gatus barbellata Engelm. in Trans. Acad. Sci. St. Louis 2:216. 1862. OT: Parry 208! Gentiana moseleyi A. Nels. in Bot. Gaz. 31:396. 1901. (T.: Moseley A.293!) Anthopogon barbellatus (Engelm.) Rydb. in Bull. Torr. Bot. Club 33:148. 1906. Acaulescent or caulescent perennials 5-15 cm. tall, from slender, fleshy, branched, horizontal thizomes, the rhizome branches erect, 5-10 cm. long, be- coming thickened towards the surface. Basal leaves oblanceolate, rarely linear, rounded or obtuse, gradually attenuate below, expanded and connate to clasping at the base, forming a membranaceous sheath enclosing the ends of the rhizome, 1.5-9.0 cm. long, 0.3—1.2 cm. wide; median leaves oblanceolate to linear, obtuse or rarely acute, attenuate to about the width of the stem at the base, connate or clasping, 2—6 cm. long, 0.5—1.0 cm. wide. Flowers solitary and terminal, sessile or short-pedicellate, the pedicels to 5 mm. long, or axillary at the node immediately beneath, the axillary flowers sessile or short-pedicellate, the pedicels 1-2 cm. long; terminal flower (and occasionally the axillary flowers) subtended by a pair of fre- quently colored, hyaline-margined, lanceolate to linear, bract-like leaves. Calyx unnelform, 11-25 mm. long, 5-10 mm. wide at the orifice; lobes triangular to lanceolate, equal, as long as or shorter than the tube, somewhat crispate near the within, narrow funnelform, 24-45 mm. long, 5-10 mm. ith obtuse to acute erose tips, spreading in flower, irregularly g as the tube or longer, 15-25 mm. long, 5 mm. long; ovary ellipsoid, sharply attenuate at the base, 10 mm. long, 3 mm. wide; stigmas sessile, elliptical, 2 mm. wide. Mature capsule as long as the corolla tube, dehiscing in the upper third, the tips of the valves erect. Seeds somewhat obovoid, papillose. Central Colorado; Wyomin Medicine Bow Range; Utah in ancisco Mountains; at elevations of 1,000—3,600 meters; on rocky, alpine and subalpine slopes, below timber line 1n 1957] GILLETT—REVISION OF GENTIANELLA 231 open coniferous forest, and occasionally in open wet meadows; flowering from mid-August until late September, fruiting from mid-September until snowfall. UNITED STATES: Coconino Co.: San Francisco Mountains, Knowlton 126 (US) F (NY), Jones $00 (NY), ‘Snow s.n. (F). Custer Co.: Sangre de Cristo Mts. near West- ANE Eggleston 6316 (US). El Paso Co: Gentian Dell, Clements & Clements 214 (GH, O, NY, US). Gunnison Co.: Taylor River, Coulter s. n. (US). Jefferson Co.: Smit Horn, Greene 327 (GH). Lake Co. Twin Lakes s, Wolf & Rothrock 794 (GH, US). e Lari Mount Flora, Snowy Range, Parry 2 , US); Estes Park, Clokey 3832 Mineral Co.: gon Wheel Gap, Murdock oe (F, MO). Leadville Forest, Eggleston 11975 (US); Mt. Harvard, Clemen vcd (NY). Saguache Co.: Weber 3598 (WS). San Miguel Co.: Wilson, Mason s.n. (NY); Breckenridge, Brandegee 253 (MO, NY); Pike’s Peak, Underwood s.n. (NY). GES unknown: Parry in 187. (NY), in 1873 (F), Parry 440 (F). Colfax Co.: Baldy Peak, Standley 14343 (US). Taos Co.: Costillo E 5j pi ei ei S). UTAH: Grand Co.: Stenize Stellen, La Sal Mts., — E (MO, US). Sanpete Co: Summit of Heliotrope, above Mayfield, Collector unknow WYOMING: Sat Co.: Medicine Bow Mts., Nelson a ae 1138 (MO, NY). Sub- lette Co.: Green River Lakes, Wind River Mts., Mutulse Creek, Eon s.n. (NY); Clif Creek, Curtis s.n. (NY), Ownbey 1129 (MO, WS); 25 m Piney, above Middle Piney Lake, Meyer & Meyer 2365 (NY). Teton Co.: Civil Creek Divide, Murie 1052 (MO). Gentianella barbellata is the only perennial species of the genus in North Amer- The flowers bear unique rows of cilia below the staminal insertion, somewhat comparable to the substaminal cilia in the related European perennial Gentianella ciliata (L.) Börner. The latter has caudate seeds, however, while in G. barbellata the seeds are papillose and ecaudate. = ` B barbellata engt dots). Fig. distrib £ Gentianella simplex (circles) and of G. B, calyx e e ao A actual size, and of G. dius (right), X about 75. [Vor. 44 2354 ANNALS OF THE MISSOURI BOTANICAL GARDEN 4. GENTIANELLA simplex (A. Gray) J. M. Gillett, comb. nov. Gentiana simplex A. Gray in Newberry, Bot. Rept. U.S. Pac. R.R. Surv. 63:87. 1857. (T.: ! ewberry s. n.! Antbopogon simplex (A. Gray) Rydb. Fl. Rocky Mts. 659. 1917. Annuals, 5-35 cm. tall; stems simple, smooth or finely striate. Leaves in 2-6 pairs, somewhat fleshy, the basal leaves spatulate, soon withering, 5-15 mm. long, 1-6 mm. wide, the apex rounded to obtuse, the base connate, the median and upper leaves elliptic, ovate, or lanceolate, 5-20 mm. long, 2-5 mm. wide, the apex obtuse, the base clasping to connate, slightly decurrent. Flowers solitary, terminal on pedicels 2-15 cm. long. Calyx narrowly funnelform, 10-25 mm. long, 5-10 mm. wide at the sinuses; lobes equal, as long as the frequently crispate tube, 5-12 mm. long, 2-5 mm. wide at the base, the outer foliaceous and veiny, occasionally cris- pate, the inner foliaceous or frequently strongly crispate and recurved; sinuses acute, covered within by a small entire membrane. Corolla pale blue to violet, narrowly funnelform, 20—40 mm. long, 4-10 mm. wide at the orifice; lobes broadly spatulate with rounded to obtuse, erose to dentate tips, erect to slightly spreading at anthesis, 5-15 mm. long, 3-6 mm. wide; sinuses acute, the lateral margins entire or with a few fringes 0.5-1.0 mm. long. Stamens included; filaments inserted in the middle of the corolla tube, the flat to slightly incurved marginal wings 1.5 mm. wide at the base, tapering slightly above; anthers 2 mm. long and wide, attached in the upper third; interstaminal glands elongate, swollen, attached 1 mm. from the base of the tube. Pistil stipitate, the gynophore 7 mm. long; ovary ovate, sharply attenuate at the base, 12 mm. long, 6 mm. wide; stigmas sessile, laminate and minutely lobed, 4 mm. wide. Mature capsule slightly exceeded by the per- sistent corolla lobes, dehiscing for one-third to half its length, the tips barely parted. Seeds elongate, caudate at both ends, 0.5-1.0 mm. long, 0.01 mm. wide, the body ovoid, the testa reticulate. Mountains of southern and southeaster n Oregon to eastern California, northern and western Idaho and western Nevada; in open meadows, wet places, and bogs; at elevations of 1200-3000 meters; flowering from late July until mid-October; fruiting from late August until the UNITED STATES: end of the growing season. i : : Bast Lake, Copeland s. së Placer Co.: Squaw Valley, Brewer hey (GH, MO, US). Placer-Eldorado H > eich egor 210 (US). Plumas Co.: Big Meadows, RS 4 ta gë B . Sierra Co.: Little Truckee River, Sonne 218 (MO, US). Siskiyou Co.: Mt. Shasta, on 17810 (WS); Mt. Eddy, Smith 554. (US); Scott Mt., Ehe s. n. (MO). Trin- (MO). T mr a am: 12755 (GH). Tulare Co.: Sequoia Nat. Park, Cronquist 2182 M Frankton 6045 (DAO). Custer Co.: Bregen ee ee? 1957] GILLETT—REVISION OF GENTIANELLA 233 Stanley Lake, Challis Nat. Forest, Thompson 13971 (GH, MO, US); Cape Horn, Mac- bride 6 Payson 3638 (GH, MO, US). NEVADA: Washoe Co.: Incline, Lake Tahoe, Mason 12208 (CAN, F, GH, MO, US, WS), Archer 6714 (NY). OR Baker Co.: Eagle Creek Meadows, Cusick 2336 (GH, MO, US, WS). Crook Co.: Ochoco Nat. Forest, Munson & Bray 194 (WS). Deschutes Co.: Deschutes River, Tumalo Village, W hited 548 (WS). Douglas Co.: Diamond Lake, Coville & Applegate 460 (US). Klamath Co.: Lake of the Woods, Applegate 321 (GH); Upper Klamath Lake, Newberry s. n. (MO). Umatilla Co.: Mottet & Peterson s. n. (ID) A very distinct species. The height of this handsome plant is rather variable. Apparently this variability depends on the dryness of the habitat and on the nature of the surrounding vegetation. The living plant has been observed in the field in Boise National Forest, Idaho. Here, hundreds of plants in all stages of develop- ment grew in damp sandy ground among sedges and among scattered seedling pines. The plants were checked carefully for any deviation from the unbranched habit, and none was found. No insect visitors were seen about the plants, but dissections of several corollas revealed the presence of thrips. Gentiana calyosa growing near by was visited by both bees and humming birds. Gentianella simplex may be related to G. ciliata of Europe. The caudate seeds of the two species are almost indistinguishable, and are unique among the fringed 8entians. G. ciliata, however, is reported to be a perennial. DOUBTFUL SPECIES GENTIANA VENTRICOSA Griseb. Gen. & Sp. Gent. 259. 1839, is figured in Hooker’s ‘Flora Boreali-Americana’ and may prove to be of considerable cyto- logical interest. I am assured by Dr. N. Y. Sandwith that no photograph of the type specimen (Drummond s.n., K: “ad cataractas fl. Saskatchewan terrarum Hudsons Bay”) could add much to Grisebach’s plate. Recently Dr. H. Scoggan of the National Museum of Canada collected another specimen of the same aspect at Grand Rapids, Lake Winnipeg, Manitoba (Scoggan 4407, Aug. 6, 1948, CAN!). These specimens have the ovary and stamens shorter than normal and the corolla is poorly developed. Apparently they do not form a large population but are found singly. It is suggested that the two polyploid series of different base num- bers occasionally may hybridize to produce this sterile hybrid. That both Scoggan's and Drummond’s specimens were collected in essentially the same region seems to indicate some cytological explanation for their occurrence. No similar oddities have been reported elsewhere in the group. GENTIANA crintra Froel. 8 CERVANTESII Griseb. in DC. Prod. 9. 102. 1845— The type, a Cervantes collection, has not been located and the description of this Mexican entity is too brief for proper interpretation. TVor. 44 234 ANNALS OF THE MISSOURI BOTANICAL GARDEN SUBGENUS Il. GENTIANELLA Gentiana L. subgen. Gentianella (Moench) Kusnez. in Engl. & Prantl, Nat. Pflanzenfam. d 1541895. SECTION AMARELLA (Gaudin, ex Griseb.) J. M. Gillett, comb. nov. Gentiana $ Amarella Gaudin, ex Griseb. Gen. & Sp. Gent. 238. 1839. Flowers 4- to 5-merous, small. Calyx lobes with green margins, rarely reduced to teeth; sinuses without an inner membrane. Corolla 4- to 5-, rare y 6- to 8- parted, the lobes ovate, triangular or oblong, entire, the orifice with or without vascularized fimbriae, interstaminal glands epipetalous, and solitary. Stamens in- serted near the middle of the tube or below; anthers oblong, slightly longer than wide; filaments glabrous. Ovary with the placenta confined to the sutures, rarely with an extra row of ovules along the ovary wall between the sutures. Seeds smooth, round to slightly flattened. Probably tetraploids on the base number nine. KEY TO THE SERIES AND SPECIES a. e orifice usually naked, without faucal fimbriae, or fimbriate and the calyx T ge » i es a a I TOPHILAE Calyx split to Ce base along one side to form a membranaceous spathiform shea ten the V ee E green, apical teeth. Southeastern Arizona, western Chihuahua and REEL, de. G. wislizeni bb. Calyx bur, ps lobes well developed; orifice of the corolla always naked, with no of fimbriae at the — of the lobes - Plants ege fro e base or simple; the ascending basal branches frequently ing reduced flow d. "Decii. lobes e ES sic? lobes broad and fo liaceous; ovary narrowly ov ellipsoid; flow n loose simple cymes or axillar ary and solitary, M in oe te cymes, Kn subtended by a pair of bract- like leaves. Als m. and. nada south to Montana... G. propinqua dd. Corolla lobes 5: det dee irregular but verd E E E broadly ovoid; flowers in compact terminal aggregate cymes or t head due to reduction Es the da rus internodes, subte ade d by the aper ber or pairs of bract-like | Southwestern Greenlan LAC M ee: ae generally econ above, the flowers of the branches about equal to those of in stem, not reduced. e. Deech perfect, abbat 2 cm. long; corolla lobes apiculate, half as long as ER tube; eux 4-10 mm. long. ; Min e to the southern Appalachians, westwar t à inne- a, Missouri, and Arkansas... quinquefolia ee. Mon meter sin amous, about 1 cm. long; corolla lobes oie to ‘rarely acute, Pee to half as long as the tube; calyx minute, about 2 mm. long. 9 Southeastern Arizona to Coahuila, Mexico. G. microcalyx aa. Orifice of the peche with dense or scattered fimbriae at the base of the lobes...........—— EE AMARELLAE dE M LE ries II. Ml i Gin Be shorter than or equal to the acute to "eng lobes; — E free at the ee (united in ssp. beterosejala) terminal flow. about n length to the —— "n ber = ee oundland to Maine and Vi Ze t; C ward to uthward in the ; Mountains southward to central UTEM Back Vtc Beet" gg tube longer than the rounded, frequently auriculate lobes; fimbriae yx ege $ md Bowers Target t ines the lateral. Attu iden KEEN AE 1957] GILLETT—REVISION OF GENTIANELLA 235 Series I. ARCTOPHILAE J. M. Gillett, ser. nov. Aloitis Raf. Fl. Tellur. 3:21. 18 (T.: A. quinquefolia (L.) Raf.) Gentiana L. sect. Arctophila Griseb. Gen. et Sp. Gent. 250. 1839. ^. GENTIANELLA wislizeni (Engelm.) J. M. Gillett, comb. nov. Gentiana wislizeni Engelm., in Trans. Acad. Sci. St. Louis 2:215. 1862. (T.: Wislizenus ! Amarella wislizeni (Engelm.) Greene, Leafl. Bot. Obs. & Crit. 1:53. 1904. Annuals 1—5 dm. tall; stems prominently angled, branched above the base, the branches ascending. ` Basal leaves ovate, oblanceolate, elliptic to spatulate, 5—20 mm. long, 3-6 mm. wide, soon withering and frequently deciduous, occasionally forming a loose rosette; median leaves ovate to lanceolate, the apices obtuse in leaves of the lower third of the plant, becoming acute above, the base cordate and clasping, 15-35 mm. long, 5-15 mm. wide; upper leaves similar to the median but coming progressively more lanceolate, the tips acute. Flowers axillary or termi- nal, in aggregate cymes, rarely in simple cymes or solitary, the flowers frequently secund or nodding; pedicels straight or strongly curved particularly when young, 3-20 mm. long. Calyx 4-6 mm. long, slit to the base along one side to form a membranaceous spathiform sheath partially enclosing the corolla, the lobes reduced to 4—5 minute, green teeth borne on the rim of the sheath. Corolla pale blue or White, tubular to narrowly funnelform, 6-13 mm. long, 2-4 mm. wide at the orifice, the lobes triangular-ovate, acuminate, one-fourth the length of the tube, Very widely spreading, bearing at the orifice few to many ciliate fimbriae, or naked, the interfilamental glands distinct and swollen, more or less oblong-scutiform. Stamens slightly exserted, the filaments inserted near the middle of the corolla tube, the wings about 2.5 mm. wide at the base and tapering slightly above; anthers oblong, 0.81.1 mm. long, about 0.75 mm. wide, versatile and attached near the middle. Pistil subsessile or with a gynophore not exceeding 1 mm. long, the ovary linear-lanceolate, 7-8 mm. long, 1.0—1.5 mm. wide, stigmas sessile, oblong, 0.6 mm. long, 0.4 mm. wide, erect. Capsule longer than the marcescent corolla, up to 14 mm. long, dehiscing in the upper one-sixth, the valves recurving at the tip. Seeds ovoid, 0.75 mm. long, 0.5 mm. wide, distinctly flattened, the surface smooth, Minutely wrinkled under high magnification, light brown. In the White Mountains of Apache County and in the Chiricahua Mountains of Cochise County, Arizona; throughout the Sierra Madres of Chihuahua, and ex- tending into Durango at altitudes of 2100-2500 m., on cool slopes and in stony Pine woods; flowering throughout September and fruiting from late September until mid-October. priest pibe Co.: White Mts, Rothrock 799 (GH, MO, US). Cochise Co Chiricahua National Mon., Barfoot Park, Chiricahua Mts., Blumer s. ». (GH, NY, US), 177 (US), Eggleston 10819 (GH, US) ; Rustler Park, Jones 28730 (MO). von : Si Rio Chico and Rio Caballo, Barlow s. n. (F, US); Santo P [un Mie Altos, Hewitt 78 (GH); Mesa west of Hop [Vor. 44 236 ANNALS OF THE MISSOURI BOTANICAL GARDEN Valley, Jones s. n. (US); Pringle 1329 (F, GH, MEXU, NY, US), 1662 (MO, NY); near lonia Garcia, Townsend & Barber 322 (F, GH, MO, NY, US); Collector unknown (MO); Pennell 19097, 19136 (US); Majarachie, Knobloch 5614 (F); mountains of Llanos, Wislizenus 206 (GH, MO). DURANGO: road between San Julián and Cerro Pierto, Nelson 4950 (US). The distinctive spathiform calyx assures instant recognition of G. wislizeni. The fimbriate corona at the orifice of the corolla is extremely variable, ranging from a dense row of ciliate fimbriae extending entirely across the base of each lobe to completely lacking in a few specimens. The lack of correlation between geog- raphy and the occurrence of plants with naked corollas prohibits recognition of them as a distinct taxon. Even as the criterion for a sporadic form the feature fails, since both types of flowers have been found on the same plant. Engelmann placed this species in Gentiana sect. ARCTOPHILA, perhaps because of the lack of fimbriae in most specimens that he examined, although both fimbriate and naked specimens were apparently seen by him. B C Fig. 3. A, calyx and corolla of Gentianella microcalyx, X 11⁄2. B, distribution of G. wislizeni (black dots) and of G. microcalyx (circles). C, calyx and corolla of G. wislizeni, X 1V4. 6. GENTIANELLA propinqua (Richards. ) J. M. Gillett, comb. nov. Gentiana propinqua Richards. in Frankl. Narr. 1st Jour. 734. 1823. (T.: Richardson 5. n., photo MO!) axillary or terminal, in simple cymes or solitary, occasionally in aggregate cymes, the terminal flower conspicuously larger the P- aleutica; pedicels 4-30 mm. long, the orrespondingly shorter. Calyx 5-12 mm. te to lanceolate, acute, the outer lobes wider s the tube, the margins minutely papillose- lateral flowers sessile or with pedicels c long, the tube 2-3 mm. long; lobes ova than the inner, about twice as long a 1957] GILLETT—REVISION OF GENTIANELLA 237 toothed or entire, the sinuses acute. Corolla pale blue to white, tubular to narrowly funnelform, 12-20 mm. long, 3-5 mm. wide at the orifice, the lobes ovate to ovate-lanceolate, one-third to half the length of the tube, the tips mucronate or denticulate, the sinuses acute or more commonly concealed by the imbrication of the lobes, the orifice of the tube naked; interstaminal glands distinct, scutiform, and appearing as a pale green patch at or slightly above the base. Stamens included or slightly exserted, the filaments inserted about the middle of the corolla tube, the wings 0.25—0.50 mm. wide at the base and tapering slightly above; anthers oblong, 0.7-1.5 mm. long, 0.6-1.2 mm. wide, versatile and attached near the middle. Pistil very shortly stipitate, the gynophore 0.5 mm. long; ovary cylindrical, 7-10 mm. long, 1-2 mm. wide; stigmas sessile, elliptic, 1.25 mm. long, 0.4 mm. wide, erect. Capsule longer than the marcescent corolla, up to 25 mm. long, dehiscing in the upper fourth. Seeds ovoid, slightly flattened, 0.5-0.75 mm. long, 0.5-0.6 mm. wide, smooth, minutely wrinkled under magnification, light tan to dark brown. Throughout Alaska and the Yukon on mountains and in valleys, extending through the Rockies to west-central Alberta and east-central British Columbia; from the Arctic Ocean occurring sporadically to the shores of Hudson Bay, with isolated stations in Beaverhead County, Montana, Gaspé County, Quebec, and the Straits of Belle Isle in Newfoundland; in a wide variety of habitats, such as alpine meadows, rocky shady hillsides, open sandy clay, gravel, or limestone areas, pine or mixed forests and in clearings, along streams and borders of marshes and in muskegs, on glacial moraines and on sea beaches, at altitudes from sea-level to about 3000 meters; flowering from mid-June until early September, fruiting from August until late September. KEY TO THE SUBSPECIES a. Terminal flowers 15-20 mm. jong. E buon than the lateral; corolla | apiculate, blue. de as Ger nan ontana......... 6a. G. propinqua = ann a. Terminal lower bios «gus ng,a CA "s lateral flowers or slightly la eg lobes ob iud wild Meatia ^ “violas to white. ër Islan d southern date 6b. G. propinqua i "idis 6a. GENTIANELLA PROPINQUA (Richards.) ssp. PROPINQUA Greng rurikiana Cham. & Schlecht. in Linnaea 1:176. 1826, ex char. (T.: Chamisso Gen d gracilis.Cham. & Schlecht. loc. cit. 1826, nom. Gentiona a setiflora Bunge, in Mem. Soc. Imp. Nat. Mosc. 7 be Mem. Soc. Nat. Mosc. 1):242. t. 9. fig. 4. 1829, ex char. et icon Gentiana pro NEE s Griseb. in Hook. Fl. Bor. Am. 2:62. 1838. (T.: Drum- (0) mon pho Gentiana Seier Griseb. in Hook. loc. aT 61. 1838. (T.: Richardson s. n.!) Gentiana arctophila B densiflora Griseb. ih ook. "a cit. 1838. (T.: Drummond s. n.!) Zeen rella opinqua (Richa rds.) Green Lodi Boc deis & Cnt. 1:53. 1904. marella ri (Griseb. ) Gree x m 1913, ex char. (T.: Fremont Amarella ventorum Rydb. in Bull. Torr. Bot. ‘Club e :463. s. n.?) Newfoundland to Alaska and southward in the Rocky Mountains to western [Vor. 44 238 ANNALS OF THE MISSOURI BOTANICAL GARDEN UNITED STATES: ALASKA: Donnelly Dort, Mile 250, Richardson Highway, Së 6294 (DAO); Jarvis (D MO); (ss in yee art and Tanana, eg 17 ie AN); Kotzebue Lund, Eschscholtz s.n. (MO); Nadie; iie 9165 (MO); Eagle Summit, Lepage 25381 (DAO); Savage River, Mexia 2100 (DAO, MO); Mt. McKinley Nat. Park, Mile 66, Nelson 5 Nelson 3978 (GH, MO, NY); Alaska Range, Nenana Valley, Porsild A Porsild 367 (CAN, GH); Castner Glacier, Porsild s Porsild 503 (CAN, GH); between Summit and McCarty, Porsild & Porsild 439 (CAN, GH); Kokrines Hills abov e Yukon 1 River, Porsild & Porsild 681 (GH); Pastolik, Norton bS Porsild & Porsild 1015 (CAN, GH); near Bluff, Seward Peninsula, Porsild & Porsild 1269 (CAN, GH); Mile 189, Richardia Hwy., Webster 121 (DAO); Fort Yukon, Kennicott s. n. (F). MONTANA: Beaverhead Co.: Sheep Mountain, Hitchcock & Mublick 12887 (MO, CANADA: ALBERTA: Lake Louise, Anderson s.n. (CAN, WS), — 45105 (CAN); Pabatac Creek, Brown 1342 (MO); Vermillion Slope, Fletcher s.n. (DAO); Blair Mt., Lamber s.n. (CAN); Banff National Park, Bow Peak, Hitch GE 8 Martin 7741 (planas : Mor GH, WS); near Banff, Macoun s. n. Sepe D Sunwapta Pass, Ledingbam 49-605 (D - AQ); Wicked River, near Peace River, Ra aup & Abbe 4331 (CAN); Smoky Lä Riles 35 (US); Jasper National Park, Mt. Edith Cav ell, Ledingham 49-627 (DAO); M e Hot Springs, Turner 5134 (DAO); Athabasca Falls, T'urner 6048 (CAN, DAO); feto Glacier 2440 (WS). yon, Groh s.n. (DAO); Kicking Horse Lake, Macoun s. n. (CAN, DAO, GH, US); Chilliwack Valley, Macoun s. n. i ts. MACKENZIE DTIG: Kiiler x (DAO); Dismal Lake, Hall s.n”. (UBC); Bernard Harbour, Johansen Di (CAN, GH); Mackenrie River, Onion, Kenni- cott & Hardisty s. n. (PH); Bear Lake, Richardson s.n. TYPE, GH); Th Hornby’s Bend, Tener 266 (CAN); Thelo: ( SE 8505 (CAN); Liard River betw hann dien ege ild qr» River Valley, Porsild o "Porsi 2 (CA ); Campbell Lake, Point Se ion, P Richardson Mts., Porsild & Porsild 6847 (CAN); i lteckicn uie n dein M CANIS 1957] GILLETT—REVISION OF GENTIANELLA 239 FRANKLIN DISTRICT: Banks Island, n. of Cape Lambton, Porsild 17593 (CAN); Vic- toria Island, vic. Holman Is. Trading Post, Porsild 1 7328 (CAN); Banks Island, Nelson MANITOBA: Churchill, Gillett 2340 (DAO, MO), Macoun s. n. (CAN), Polunin 40 (CAN), Porsild 5494 (CAN); York Factory, Bell s. n. ( CAN), Scoggan 6052 (CAN). NEWFOUNDL Henriette, Dutilly & Lepage 31244 (DAO); Lake River, Smith 8 QUEBEC: G Co.: Anse Pleureuse, Victorin & Germain 49435 (DAO); James Bay: Cap Jones, Dutilly, Lepage Si Duman 32527 (DAO). „ YUKON TERRITORY: Dawson, Calder & Billard 3380 (DAO), Eastwood 425 (F); 60- mile road, West Dawson to Alaskan border, Calder & Billard 4418 (DAO); Mayo Lake, Green 24 (DAO); Whitehorse, Gillett 3463, 3470, 3660, 3813, 3834 (DAO); Keno Hill, Gillett, Calder et al. 4338 (DAO); North Fork, Klondike River, Cockfield 51 (CAN); Cassiar Mts. 20-30 miles e. Teslin Lake, Poole s. n. ); Lake j j chell 3950 (DAO); Big Salmon, Lewes River, Gillett & Mitchell 3971 (DAO); Stewart Landing, Gillett & Mitchell 4084 (DAO); Independence Creek, Stewart River, Gillett & Mitchell 41 34 (DAO); Mt. Caribou, nr. Carcross, Gillett & Mitchell 4568 (DAO); Ram- Mackintosh, Alaska Hwy., mile 1022, Anderson & Brown 10053 (CAN); Moosehide Mt., Campbell 51 (JBM); Frith River, McEwan 194 (CAN); Canol Road, Mt. Sheldon, Porsild S Breitung 11 753 (CAN); Mile 132, Lower Lapie River Crossing, Porsild & Breitung 9991 (CAN); Mile 95, Upper Rose River valley, Porsild & Breitung 10403 (CAN); Mile 36-42, Misutlin River, Porsild & Breitung 10849 (CAN); Herschell Island, Johansen 550 (CAN); Franklin Expedition, Richardson s. n. e OCATION UNKNOWN: Arctic Sea Coast, Richardson s. n. (CAN, GH); Rocky Mts., Drummond s. n. (K, GH). Gentianella propinqua ssp. propinqua is recognized easily by the larger terminal flowers, with apiculate corolla lobes. These features, coupled with the ascending basal branches, provide good field characters for distinguishing G. propinqua from G. amarella ssp. acuta where these species occur together. This species has been examined in the field, particularly in the Yukon. The height of the plant varies considerably in the Whitehorse area, simple and branched Plants &rowing side by side in habitats ranging from forests of Pinus contorta var. latifolia to open bare sandy places. Numerous dead plants from previous years Were found in sheltered wet places, their condition clearly indicating their annual character, At Carcross, Yukon Territory, I had an opportunity to climb Mt. tibou and to observe that the plants decrease in size with increase in altitude without any apparent break in the population continuity, except perhaps on rock slides where growth of plants in general was curtailed. A journey by steamer along the Lewes and Stewart rivers revealed an almost Continuous distribution between Whitehorse and Mayo Landing. At Churchill, Manitoba, the species was not common and was restricted to 8tavelly roadsides and to the edge of the airport runways. A few specimens from southern British Columbia and the solitary specimen from Montana have corolla lobes that are broadened above the sinus and bear no apicula at the tip (arctophila type). Except that they occur in this general region, there appears to be no definite boundary to the population. One sheet, Calder & Savile 1 2038, collected at an altitude of 7,700 feet on Mount Stephens near Field, Vor. 44 240 ANNALS OF THE MISSOURI BOTANICAL GARDEN B. C., bears both apiculate and exapiculate plants. The sixteen plants on the DAO sheet are all quite small and were found on an open heath near talus slopes. The smallest plant is 3 cm. high and bears a flower fully 2 cm. long with obtuse ex- apiculate lobes; the largest plant is 6 cm. tall and has normal apiculate flowers indistinguishable from many other collections. During the summer of 1954, J. A. Calder and D. B. O. Savile encountered a patch of pure white-flowered G. propinqua located not more than a few yards away from the blue-flowered form. Although white-flowered forms occur among most colored-flowered plants, these plants show a complete lack of the purple coloration normally present in G. propinqua. Probably this is an expression of a recessive character, but because a considerable colony had apparently become established within the blue-flowered population, I feel justified in providing it with a name. GENTIANELLA PROPINQUA (Richards.) J. M. Gillett forma acyanea J. M. Gillett, form. nov. Plantae viridulae non purpureo-tinctae; floribus in vivo albis in sicco ochro- leucis. CANADA: H COLUMBIA: Steep slope at base of cliffs at 5,000 ft. near head of valley, West- ern ium Mine about 9 miles ne. of e ossing, Aug. 30, 1954, Calder, Sav. ile 8 in moss by spring Aug. 2, 1951, Cody 6294 (DAO). The paucity of data available to me prohibits any correlation of the characters of G. arctophila, as given by Grisebach, with either geographic area or altitude. Until such data are available, it seems best to relegate this name to synonymy of G. propinqua. However, the following comments regarding the types of G. arcto- bbila and G. arc. tophila B densiflora may be of interest, should the necessity arise to give the populations taxonomic recognition. A sheet from the Gray Herbarium has four different collections, among which one is labeled “Gentiana arctopbila m. B.N.A. nr. 17."; below this specimen, in another hand, appear the words “R. Mżs. Hook.” Above the label are three plants, one of which is separated from the others by a line and labelled "Arctic Sea-coast". In the New York Botanical Garden Herbarium, a Meisner Herbarium sheet is labeled “Gentiana arctophila Gr. America arctica” in the same handwriting as the name on the Gray sheet. On this same label in a different hand is the com- ment “Com. cl. Grisebach 28/ 4/65.” These specimens compare very closely with the photographs of the type from Kew. This evidence seems to indicate that the Gray sheet, at least, consists of iso- types of both G. arctophila and G. 1957] GILLETT—REVISION OF GENTIANELLA 241 6b. GENTIANELLA PROPINQUA (Richards.) J. M. Gillett, ssp. aleutica (Cham. & Schlecht.) J. M. Gillett, comb. & stat. nov. Gentiana ein Ven & Schlecht. in Linnaea 1:175. 1826, ex char. (T.: Chamisso & Eschscholtz Gentiana ke ée Cham. & Schlecht. loc. cit. 1826, nom. nud. Throughout the Aleutian Islands, extending to the mainland as far as Juneau; mountains at elevations of 100—500 meters, in stony and mossy places, dry gullies and in grassy iR in gravel; flowering and fruiting throughout August. UNITED Stat ALASKA: "Rob Kodiak group, Ge 5236 (DAO); Semisopochnoi Island, Coats 75 (US); Unimak Island, False Pass, Eyerdam E See US); Jim Fishe Ridge, Attu land, Hardy 297 (MO); Attu Island, vicin ity Massacre Bay, Lookout Hill, Van Scbaack 919 (MO), 935 (MO, US); Kodiak Island, Olga Bay, Loof d Loof 543 (GH), 530 2 valley near Cannery Station, Loof & Loo of 1689 (MO, NY); Volcano Mt., Looff & Loof 1598 (CAN); Cold Bay, Schofield 238 (DAO s e m E a Le E D so LI e D v E) B d of G. sepe NE A, calyx and corolla of Gentianella propinqua (above), X 74; an x" y distribution of G. propinqua ssp. propinqua (black dots), of G. propinqua ss; thal black dots), and of G. aurea (circles). [Vor. 44 242 ANNALS OF THE MISSOURI BOTANICAL GARDEN 7. GENTIANELLA AUREA (L.) H. Sm. ex Hylander, in Uppsala Univ. Arssk. 259. 1945. Gentiana aurea L. Syst. ed. 10. 951, 1759. (T.: Konig s.n., photo A!) Gentiana quinquefolia Oeder, Fl. Dan. t. . 1766, not L., ex i n c. Gentiana involucrata Rottb. in Kiob. Skr. Selsk. 10:434. 1770 (T.: Collector unknown, Gentiona umbellata Marsch. & Bieb. Fl. Taur. Cauc. 3:188. 1819, ex char Gentian j tiana aurea a borealis a. genuina Herder, in Acta Hort. Petrop. 1:438. 1872. (based on G. au a . Gentiana Pseud-Amarella Stev. ex Herder, loc. cit. 437. 1872, nom nud. in syn. Annuals 2-30 cm. tall, simple or branched from the base, the branches curved- ascending, frequently bearing reduced flowers. Basal leaves frequently forming a rosette, elliptic to spatulate, the apex rounded to obtuse, attenuate to the base or to a slender petiole one-third the length of the blade, 3-21 mm. long, 1-10 mm. wide; median leaves lanceolate to ovate, the apex obtuse, the base rounded, 9-26 mm. long, 4-13 mm. wide; upper leaves similar and enclosing the terminal in- florescence. Flowers axillary, in simple or aggregate cymes or in dense terminal clusters, or rarely solitary, sessile or with short pedicels 1-13 mm. long, the lateral flowers sessile or with pedicels shorter than those of the terminal ones. Calyx 3-8 mm. long, generally 5-lobed, the tube 1.5—2.0 mm. long, the lobes linear to obovate, acute, two usually larger than the others, about two to three times as long as the tube, the margins ciliate to entire; sinuses rounded. Corolla blue to violet (?), tubular to narrowly funnelform, 6-11 mm. long, 3-5 mm. wide at the orifice; corolla lobes ovate, one-third the length of the tube, the tips mucronate, the sinuses acute, the orifice of the tube naked; interstaminal glands indistinct, scutiform, about 0.25 mm. above the base of the tube. Stamens included, the filaments in- serted in the lower third of the corolla tube, the wings 0.25 mm. wide at the base and tapering above; anthers oblong, about 0.75 mm. long, 0.5 mm. wide, versatile and attached near the middle. Pistil very shortly stipitate, the gynophore 0.5 mm. long, the ovary short-cylindrical to ovoid, about 6 mm. long, 1-2 mm. wide; stigmas sessile, about 0.25 mm. long and wide, oval and revolute. Capsules as long as the marcescent corollas or slightly longer. Seeds ovoid, slightly flattened, about 0.75 mm. long, 0.6 mm. wide, the surface smooth, minutely wrinkled under magnification, light brown. In North America only in Greenland, but generally circumpolar, occurring in Norway and northern R ussia, extending into the high mountains of southern Europe, the Caucasus, the Himalayan Mountains and in the province of Kansu, China; in Greenland growing in grassy fields at altitudes probably from sea-level to about 300 meters; in the Eurasian mountains at altitudes to 3200 meters; flower- ing from mid-July until mid-August; fruiting until early September (G E GREENLAND: Igdlungujak, Jessen s.n (C); Tissaluk 61° 21', Hartz s. n. (C); Kingua, Tasersuak, Hartz e n. (C); T. M Jj Me o): liko, Har tz s.n. (C); Kakasuak near Ken d ag ed Hartz s. n. (C, GH, MO); pn Fe mann s. n. (C); Frederiksdal, Lundholm .9 Tunugdliarfik Fjord, Petersen in 188 95 30° W., Porsild 8897 (CAN); 1957] GILLETT—REVISION OF GENTIANELLA 243 Igaliko (Gardar), 60° 59' N., Porsild t$ Porsild s. n. (C, GH); Igdlorssuit, Prins Chris- tans Sund, 60° 10’ N., Porsild & Porsild (C); Igaliko, Rosenvinge s.n. (C); Kingua i Gentianella aurea is easily distinguished by the dense, terminal flower clusters subtended by the ovate upper leaves. Because of the characteristic curved-ascend- ing basal branching pattern, G. aurea somewhat resembles G. propinqua. Although some depauperate specimens lack this feature, the mucronate corolla tips are a good supplementary character for recognition. Very little material has been seen from Greenland, and therefore European specimens have been employed in drawing up the description, for the latter appear to be similar to the Greenland plants in every respect. Very little habitat data are given with the Greenland plants, so that this information also is drawn chiefly from the European collections. 8. GENTIANELLA QUINQUEFOLIA (L.) Small, Fl. Southeast. U. S. 929. 1903. (G. quinquefolia (L.) D. Lóve in Hereditas 39:227. 1953, superfluous comb.). Gentiana quinquefolia L. Sp. Pl. 1:230. 1753. (T.: Kalm s.n., Sav. Cat. 328.31, photo A!) Annuals (or biennials?) 2—8 dm. tall, profusely branched above, the branches curved-ascending, occasionally with the lower slender branches bearing reduced flowers, rarely simple in smaller plants. Basal leaves elliptic-spatulate to sub- orbicular, 10—22 mm. long, 4—10 mm. wide, the apex rounded, attenuate below to a slender petiole as long as the blade, soon withering and both basal and lower leaves usually deciduous; median leaves broadly ovate, acute, the base cordate to rounded, clasping, 15—60 mm. long, 5-35 mm. wide; upper leaves similar but Progressively smaller. Flowers in compact, umbelliform, aggregate cymes, terminal on the branches, very rarely axillary and solitary; pedicels 2-17 mm. long, those o the subterminal node frequently longer than those of the terminal node. Calyx 4-10 mm. long, the tube 1.5-3.5 mm. long, the lobes about equal to or longer than the tube, triangular to elliptic-lanceolate, the margins entire, the sinuses acute to rounded. Corolla blue to white, narrowly funnelform, 15-20 mm. long, 4-7 mm. Wide at the orifice; lobes ovate, half the length of the tube, the tips apiculate, the sinuses acute, the orifice of the tube naked; interstaminal glands scutiform, at the very base of the tube. Stamens included, the filaments inserted about the middle of the corolla tube, the wings strongly incurved, about 1 mm. wide at the base, tapering above; anthers oblong, 1.75 mm. long, 0.5 mm. wide, versatile and attached near the middle. Pistil stipitate, the gynophore 2-3 mm. long; ovary cylindrical, 12 mem long, 1.5 mm. wide; stigmas sessile, or with an indistinct style tapering into the Ovary, oblong, 1 mm. long, 0.25 mm. wide, recurving. Capsule as long as the marcescent corolla, dehiscing at the tip. Seeds rounded or slightly flattened, 0.5— 0.6 mm. long, the surface smooth, light brown. gl A polymorphic species with two clearly distinct allopatric subspecies distin- Suished chiefly by flower and calyx size. TVor. 44 244 ANNALS OF THE MISSOURI BOTANICAL GARDEN ^ v et Y Y e x A H A B A, distribution of Gentianella quinquefolia ssp. quinquefolia (black dej end of ssp. occidentalis (circles). B, corolla of G. quinque- folia Cop), ion of ssp. quinquefolia (middle), and of ssp. occidentalis (bottom), X 134. KEY TO THE SUBSPECIES a. Calyx 4—5 € long, the tube ës mm. long, the lobes 2.0—2.5 mm. long, nei triangular with hyaline margins an ery prominent keels; corolla kon mm. lon Maine westward to New York and ded Ontario, southward to Georgia, oli vum erue ESTIS BEBE SDN MOM 845. G, Geisel ssp. d " Calyx 8-10 mm. lon tube 3 ie to rep uh with thick like, keels; corolla about 20 mm. lo: Kentucky. 3.0-3.5 mm. long, the lobes Ain m. long, ellipti RES aa and broadly niti frequently — ong. Ohio to Minnesota, southward to Arkansas di a O cnp 8b. G. Cannio ssp. Dt 4 8a. GENTIANELLA QUINQUEFOLIA (L.) Small, ssp. QuINQUEFOLIA Gentiana quinqueflora Lam. Encyc. 2. 643. 1768, sphalm? Hippion quinquefolium (L (L.) Schmidt, in Roem. Archiv f. Bot. 1:11. h Gentiana amarelloides Michx. Fl. Bor. AB. 175. 1805. .(T.: ul s. n., photo M Aloitis parviflora Raf. Fl. Tellur. 3. :21. 1837, ex char Aloitis ipae dapes EM omg Lam.) Raf. is cit. 22. 1837 hat. entiana quinqueflora Lam. B parv iflora Raf. ex Griseb. in DC. Prod. 9:100. Bee? Gentiana Leuten id r var. necis (Michx.) Britt. in Mem. Torr. Bot. Club 5:2 18 Aloitis divaricata Greene, Leafl. Bot. Obs. & Crit. 1:94. 1904. Amarella amarelloides (Michx X.) Greene, loc. cit. 53. 190 ! Gentiana quinquefolia L. £. lutescens Fern. in Rhodora 19: 151. 1917. (T.: Porter s.n!) Maine to New York and southern Ontario southward in the Appalachians to North Carolina and Tennessee, at altitudes up to 1800 meters in the southern Ap- steen : Elgin Co.: Fisher 5 -m. (MO). Lambto For Dodge s. n. (MO). Welland C. Co.: (aen Falls, Scott s.n. (US). r2: rk Co.: Ei e" We V s.n. (US)- 1957] GILLETT—REVISION OF GENTIANELLA 245 sab STATEs: CONNECTICUT: Litchfield Co.: South Canaan, Greenman 1 (GH, MO). GEORGIA: Rabun Co.: Blalock, Wherry & Pennell 14087 (F, O). : N . ( ). MARYLAND: Garrett Co.: Oakland, Hermann 11380 (MO). MASSACHUSETTS: Berkshire Co.: Kitchen Brook, — s. n. (GH, MO). NEW HAMPSHIRE: Grafton Co.: Squam Lake, Faxon s. n. (GH, US). W JERSEY: Sussex Co: High Point, rper enge Sie? (MO ). Warren Co.: Phillips- "one NOE ab NEW YORK: a Co.: Hall s.n. (MO). Columbia Co.: Columbiaville, McVaugh 4037 (GH). pu. Co.: Cove, Standley & iunt 12361 (US). Erie Co.: Clinton s.n. (NY). Jefferson Co.: Wat der ui Ber .m. (US). Madison Co.: Chittenango, Morong s. n. (US). Onondaga Ges Syracuse, redis s. n. (MO). Orange Co.: Brook- Wi Collector unknown (MO). Onde Co.: Phoenix, Hastings s. n. (NY). D : j i . (MO Tompkins Co.: Ithaca, Sheldon s.n. (US). Warren Co.: Lake George, House 25901 NORTH CAROLINA: Buncombe Co.: te pies Herb. 465b (GH, MO, US). Haywood Co.: Pisgah een Hon 3606 (GH). McDowell Co.: Little Roan, Merriam S). Polk Co.: Get s, Townsend s.n. (US). Yancey Co.: s Ge zm E PENNSYLVANIA: Center Co.: Coburn , Wbyl 271 did Chester Co.: Porter s. n. (GH). Lackawanna Co.: Carbondale, Topping s. n. (US). — Co.: ‘ed os Small & Heller s. n. (US). Lycoming Co.: Williamsport, Smith s. n. (PH). Naomi Pines, Small s.n. ders Northhampton Co.: Easton, Heller 630 a US). t > F). Se p Carter Co.: Res an Mountain, Britton i^ Britton s. z4 (NY). iem Co.: Weybridge, Brainerd s.n. (NY). Bennington Co.: Man- TE Dey uo 5 (US). Caledonia Co.: East gegen AE: s.n. (GH). Orange : "i R ye d; Eggleston 1503 (U er deat por Wraps E. Bedford Co.: Peaks VIRGINIA: Au : Augusta Springs, Steele s. n. (MO, of Otter, Beyrich 87 (MO, NY). Sica Co.: Apple Orchard Mountain, Freer 1902 (GH). Frederick Co.: Cedar Creek, Canby s. US). i Rawlinson 272 (US ). rhe Co.: Shenandoah Nat. Park, Graf 3 (US). Nelson Co.: nam Steele s.n. (US). Page Co.: Hawksbill Mt., eegen 5461 (GH). Roanoke Co.: ls o or (GH Ro kbridge Co.: Steele s.m. ( PA Ch brie eg : y nee Mt., Hunn sewell 7105 (GH). Monroe and Alleghen ip Ge Allegheny geg Steele 8 Steele 418 (GH, MO, sya Pendleton Co.: Lake Terra Wer Béskley 1781 (MO). Preston Co.: Br cokside, Olds s. n 8b. GENTIANELLA QUINQUEFOLIA (L.) Small, ssp. EE à Gray) J. M. Gillett, comb. & stat. nov. Gentiana quin L identalis A. Gray, Man. ed. 1, 359. 1848, based on Hook. Bot iu GE? ^1836. (c.f. Gray in Syn. Fl. N. Amer. 2: tie? ID. Gentiana PT a L. var. occidentalis (A. Gray) A. S. Hitch. in Trans. Aca E uis 5:508. ong quinquefolia (L.) Small, var. occidentalis (A. Gray) Small, Fl. Southeast. U.S. 929, Amarella Ven (A. Gray) Greene, Leafl. Bot. Obs. & Crit. 1:53. 1904. Aloitis mesochora Greene, loc. cit. 94. 1904, ex duc Aloitis occidentalis (A. Gray) Greene, E cit. en foliosa Greene, loc. cit. 1904. Se ER ey s. - SH A entiana occidentalis (A. Gray) Fitzpat. in Iowa Nat. S Ohio westward to Wisconsin, Iowa and Minnesota, south to Arkansas dag rf tucky; occurring i in much the same habitats as the typical subspecies ze? ra On prairies and along wooded oak-juniper bluffs and wooded hillsides; as in [Vor. 44 246 ANNALS OF THE MISSOURI BOTANICAL GARDEN typical subspecies, seeming to prefer limestone areas and wet stream gullies; flower- ing dates similar to the typical subspecies. Kater ée E ARKA , Newton Co.: Vendor, Iltis s. n. (U. Ark.). Stone Co.: Blanchard Springs, Iltis s. n. (U. A k.). INOIS: deen Co.: Rantoul, Gleason s. n. (GH). Cook Co.: Glencoe, Um- bacb s. n. (GH, PH). Hancock Co.: Augusta, Mead s. n. (MO). Lake Co.: Lake Forest, š ane i ttawa, Jensen 113 (MO). K Co.: Aurora, Kane s.n. (PH, US). LaSalle Co.: O 4 Huett s. n. (GH). Menard Co.: Athens, Hall s. n. (MO, PH, US). seu Co: pe McDonald s.n. (GH, NY). St. Clair Co: French Villäge, Eggert s (GH, MO). Stephenson Co.: Freeport, Johnson s.n. (US). Tazewell Co.: East bas Chase. 3724 (US). Vermilion Co.: Fithian, Gates 2175 (US). INDIANA: Allen Co.: St. Joe River, Robinson Park, Deam 606 (MO, NY, US). Benton Co.: Shipman s.n. (PH). Howard Co.: Wildcat, e unknown ). Lagrange (NY o.: Mongo, Deam 19264 (GH). Marshall Co.: Lake Maxinkuckee, Scovell & Clark 1459 (US). St. Joseph Co.: Notre Dame, Nieuwland 10406 (US). Vigo Co.: Terre Haute, Evermann s. n. ( 1owa: Allamakee Co.: Henovár Deas -, Tolstead s. n. (MO). Cerro Gordo Co.: Maso n City, Shimek s. n. (WS). Dubuque Co.: Lux xemburg, Shimek s. n. (MO). Fayette Co.: Fink 487 (GH, US). Hamilton Co.: Webster City, Hayden së? NE Hardin Co.: Iowa Falls, Peck sn ). Poweshiek Co Grinnell, Jones s. GH). Story ^s B Winneshiek Co.: Ridgeway: Hayden 10222 (US). KENTUCKY: Short s.n. (MO, P MICHIGAN: Genesee Co.: Flint, Clark 5 (US). Ingham Co: Lansing, Tracy 5 (US). Tonia Co.: Hubbardston, Wheeler s.n. (US). Kent Co.: Grand Rapids, Cole s. Ze (MO). Saint Clair Co.: Port Huron, Dodge s. n. (GH, MO). Washtenaw Co.: Dexter, Palmer s. n. (US). Wayne Co.: Woodbridge Park, Chandler s. n. (U i MINNESOTA: Houston Co.: Freiberg s.n. (MO). Nicollet Co.: St. Peters, Geyer 234 (MO). Olmstead Co.: Pleasant Grove Twp., Moore, Leedy & Thatcher 15689 (DAO)- Wabasha Co.: Lake City, Manning s. n. (GH). Winona Co.: Whit itman, Moore & Neva WS). MISSOURI: Adair Co.: Broadbead En n. Pn ate: Co.: Van Buren, Palmer I ad (MO). Clark Co.: Bu sh s.n. (M O, NY). H : 2t 05 D ^ Co.: Pacific, Eggert s. n. (MO). Iron Co.: Royal eran "Kellogg 15260 (MO). et? SS Jewett, — res (MO). poeta A Lesterville, Steyermark wm ).. Erie Leg East Fork, Vermilion River, Mose- ley s. n. (F, WC Qi: Georgeville, Werner 539 ae Hamilton Co.: Collector unknown (PH). Ee Co.: Monclova, deu ruo s.n. (F, US). Mercer Co.: oe unknown (PH). Montgomery Co.: . (PH). ONSIN: Brown Co.: un Schuette s. n. CH). Buffalo Co.: Fountain AE P n Se Dax cae adison, McMurphy s. n. (US). Lacrosse Co.: 2 ANIM . Zo an enos. ` Merk: 1988 (DAG), a Cos: Wadmond s. n. (PH). Vernon Co.: 9. GENTIANELLA microcalyx (Lemmon) J. M. Gillett, comb. nov. Gentiana msg gc x Lemmon in Pacific en ae 23:129. Géi? "^ 1882; Engelm. eX Gra Amer. Acad. 17:222. (T.: Lemm Lemmon) Greene, JS Bot. Obs. & Crit. 1: e. 1904. Annuals 15—45 cm. tall; stems prominently angled, generally branched above rarely from the base, the branches sharply ascending. Basal leaves oblanceolate, elliptic, to spatulate, th € apex rounded to obtuse, attenuate to the base, 10-25 mm. A. Gray, in Amarella keet ( 1957] GILLETT—REVISION OF GENTIANELLA 247 long, 2-8 mm. wide, soon withering and deciduous; median leaves broadly ovate to lanceolate-ovate, the apex acute, cordate or rounded and clasping at the base, 0-40 mm. long, 3-20 mm. wide; upper leaves similar to the median, becoming progressively more lanceolate. Flowers perfect, occasionally unisexual by abortion of the anthers (or, more rarely, the pistil), in compact umbelliform aggregate cymes, terminal on the branches, very rarely axillary and solitary; pedicels of the terminal flowers about equal, 3-15 mm. long. Calyx minute, 2.0-2.5 mm. long, 4- to 5-lobed, the tube 1 mm. long; lobes triangular, acute, essentially regular, about equal to the length of the tube or slightly longer, the margins entire, the sinuses rounded. Corolla white to pale lavender, tubular, narrowly funnelform or salverform when the lobes are fully expanded, about 10 mm. long, 2-3 mm. wide at the orifice; lobes oblong, one-third to half as long as the tube, the tips obtuse to rounded, the sinuses acute, the orifice of the tube naked; interstaminal glands indistinct, appearing only as dark oblong stains by transmitted light, 0.5 mm. from the base of the tube. Stamens slightly exserted; filaments inserted near the upper third of the corolla tube, scarcely winged, the wings about 0.25 mm. wide at the base, tapering above; anthers oblong, 1.3 mm. long, 0.8 mm. wide, versatile and attached near the middle. Pistil shortly stipitate, the gynophore 0.5 mm. long, ovary cylindrical, 11 mm. long, about 2 mm. wide; stigmas sessile, oblong, 0.4 mm. ong, 0.2 mm. wide, erect. Mature capsule longer than the marcescent corolla, 12 mm. long, dehiscing at the tip. Seeds ovoid, slightly flattened, 0.75 mm. long, 0.5 mm. wide, smooth, light brown. In the Huachuca Mountains of Cochise County, the Rincon and Santa Cata- ina mountains of Pima County, and in the Santa Rita Mountains of both Pima and Santa Cruz counties, Arizona; in Mexico in the Sierra Madres of southwestern and central Chihuahua, eastern Sonora, eastwards in the mountains about Cuatro Cienagas, Coahuila, southward to Durango; in cool moist pine-oak forest and on burnt-over slopes at altitudes of 600—2000 meters; flowering from late August until the end of September; fruiting in October. Huachuca Mts., Carr Peak, Goodding 858 (GH, NY); Carr Ramsey Canyon, Jones 7 (GH anyon, Lemmon 2522 (F, GH, Us), Peebles, Harrison et al. 3399 (US) ; Fort Huachuca, Wilcox 507 (US); Towner’s (NY). Pima Co: Rincon Mts, Manning Camp, Blumer s. n. (F, GH, MO); Santa Rita Mts., Griffiths 6049 (US), Griffitbs & Thornber 182 (US), Loomis 3269 (US); Santa Catalina Mts., Mt. Lemmon, Harrison 3020 (US), Heally mgr (US); Marshall Gulch, Shreve 5394 (GH, UC, US). Santa Cruz Co.: Santa Rita Mts Canyon, P ilsbry s. n. (PH), Wooton s. n. (US); location unknown, Toumey s. n. (US). Un ARIZONA: Cochise Co.: Canyon, Gould & Haskell 3373 (GH, MO), Jones s.n. (GH); CHIHUAHUA: Temosachic, Madera, Muller 3465 (GH). COAHUILA: Cuatro Cienagas, Sierra de la Madera, Muller 3235 (MO). SONORA: Rio de Bavispe, 4 mi. e. El Belito, White 4771 (GH). [Vor. 44 248 ANNALS OF THE MISSOURI BOTANICAL GARDEN Series II. AMARELLAE J.M. Gillett, ser. nov. Amarella Gilib. Fl. Lith. Inch. 1:36. 1781, nom nud.; Raf. Fl. Tellur. 3:20. 1837. (T.: Antbopogon amarella (L.) Raf.) Gentiana L. *Coelantbae Froel. Gent. Diss. 15. 1796, proparte. Gentiana L. ***Endotricbae Froel. loc. cit. 86. 1796, ut sectio Murbeck, in Acta Hort. Berg. 23:1—28. 1892. Eyrytbalia Borckh. in Roem. Archiv f. Bot. 1:28. 1796, pro parte maj. Gentiana L. ** Amarella Gaudin, Fl. Helv. 2:270. 1828; ut sectio Griseb. Gen. et Sp. Gent. 258. 1839. 10. GENTIANELLA tortuosa (M. E. Jones) J. M. Gillett, comb. nov. Gentiana tortuosa M. E. Jones, in Proc. Calif. Acad. II, 5:707. 1895. (T.: M. E. Jones Amarella tortuosa (M. E. Jones) Rydb. in Bull. Torr. Bot. Club 40:463. 1913. Gentiana helleri Briq. in Candollea 4:331. 1931. (T.: Heller 11072!) Acaulescent or caulescent, cespitose annuals 2-8 cm. tall, from very long tap- roots. Basal leaves elliptic, lingulate, to spatulate, the apex rounded or obtuse, attenuate below to about the width of the stem, clasping, 5-25 mm. long, 2-6 mm. wide; median leaves elliptic, oblong, or lanceolate, obtuse to acute, the base attenu- ate, 5-35 mm. long, 1-5 mm. wide. Flowers solitary in the axils of the stem leaves, or on short axillary branches; pedicels short to 15 mm. long. Calyx 5-7 mm. long, the tube very short, 1-2 mm. long, the lobes closely enclosing the corolla tube, linear to oblanceolate, acute, unequal, the. margin hyaline and papillose, the sinuses rounded. Corolla white (or pale blue?), broadly funnelform to somewhat campanulate, 5-8 mm. long, 3-4 mm. wide; lobes ovate, obtuse, equaling the length of the tube, bearing a corona of 2-10 hyaline, papillose-margined fimbriae one- third the length of the lobes and extending across the base or in two distinct clusters on each side, the lower flowers occasionally naked, the sinuses acute; inter- staminal glands rather indistinct, appearing as pale green scutiform patches at the very base of the corolla tube. Stamens included, the filaments inserted in the lower half or third of the tube, the wings 0.25 wide at the base; anthers short, oblong, 0.8 mm. long, 0.6 mm. wide, versatile. Pistil short-stipitate, the gyno- phore 0.3—0.5 mm. long; ovary elliptic-ovoid, 4—5 mm. long; stigmas sessile, ellip- tic, 0.5 mm. long, 0.3 mm. wide, erect. Mature capsule as long as the marcescent corolla or slightly exserted, dehiscing in the upper half. Seeds elongate-ovoid, about 1.2 mm. long, 0.7 mm. wide, very slightly flattened, minutely wri under magnification, light brown. Mountains of southwestern Utah and southern Nevada; along damp banks and on open hillsides and bare gravelly slopes among yellow pine at elevations of 2600- 3300 meters; flowering from mid-July until late August; fruiting in September. UNITED Sieg " NEVADA: Clark Co.: Ridge to Charleston Peak, Clokey 7625 (GH, MO, NY, U mil meg O, NY, US) 8061 (NY); head of Lee oe ee Mts., Heller AH: M. E. Jones 6008 : 5 : en 4516 (UC, WS). Kan Co 37 Dë A, Hitchcock, Rathke, d Von eier d Co.: Bryce Canyon, Eastwood & Howell 7230 (US); Panguitch Lake, V 1957] GILLETT— REVISION OF GENTIANELLA 249 gies? Fig. 6. Distribution of Gentianella tortuosa in Utah and Nevada. Gentianella tortuosa, a very distinct species, appears to be rather localized since the collections came from only five localities. Distinctive field characters include the cespitose habit and the very numerous, somewhat campanulate flowers borne throughout the plant. The very long slender tap-root, sometimes twice as long as the aerial part of the plant, is a feature unique in this species. The key appears to indicate a close alliance of G. tortuosa and G. amarella ssp. acuta. It would be unwise, however, to venture an opinion on relationships here, since the floral morphology, exclusive of size differences, is startlingly similar in many species of section AMARELLA, while in vegetative characters the plants are quite different. In floral morphology, G. tortuosa bears a great similarity to G. aurea of Greenland, except for the obvious expansion of the corolla itself and the Presence of fimbriae in the throat. Since in so many instances in other species the Presence or absence of fimbriae in the corolla orifice has proved to be an unreliable character, the presence of fimbriae should not be considered discordant. The ovoid Ovary is a more reliable character to suggest affinity with G. aurea, since all other Species within the section have distinctly cylindrical ovaries which are almost in- distinguishable from one another except for size. Considering purely vegetative characters, the branching pattern of G. tortuosa, consisting as it does o many Stems, suggests affinity with G. amarella ssp. acuta. The few fimbriae in G. tortu- 954 are hyaline and short, but although frequently found in two phalanges or groups 9n either side of the midvein, possess distinct vascular bundles. ll. GENTIANELLA AMARELLA (L.) Bórner, Fl deut. Volk, 543. 1912. (H. Smith in Hylander, in Uppsala Univ. Arssk. 259. 1945, superfluous comb.) Genti : . (T.: Linn. Herb. 328.32, photo, A!)59 prime e Se WU die or Ph Nd from the base or above, the basal branches occasionally bearing reduced flowers, the upper cauline branches fre- quently reduced. Basal leaves elliptic, lingulate to spatulate, the apex obtuse to — 59 Eurasian synonyms other than the basonym are not included here. [Vor. 44 250 ANNALS OF THE MISSOURI BOTANICAL GARDEN : rounded, attenuate to the base or to a slender petiole as long as or shorter than the blade, frequently forming basal rosettes or soon withering, 3-40 mm. long, 1-11 mm. wide; median leaves oblanceolate, elliptic, ovate to ovate-lanceolate, the apex acute, occasionally obtuse or rounded in western forms, the base rounded to cordate, subconnate to somewhat decurrent or clasping, 8—60 mm. long, 3-20 mm. wide; upper leaves similar to the median or becoming progressively smaller and more acute. Flowers few to very numerous, axillary or terminal, in simple or aggregate cymes, frequently in dense clusters or solitary in the axils, sessile or borne on pedicels up to 5 cm. long. Calyx 5-25 mm. long, 4- to 5-lobed, the tube 2-40 mm. long, the lobes various, regular to irregular, one or more slightly larger than the remainder or becoming foliaceous, the margins entire, minutely denticulate to ciliate, the sinuses acute to somewhat rounded. Corolla white, blue, mauve, or yellow, or with a greenish tube, tubular, narrowly funnelform to salverform when the lobes are fully expanded, 10-25 mm. long, 2-6 mm. wide at the orifice, the lobes ovate to ovate-oblong, half the length of the tube, the tips obtuse to acute, the sinuses acute, the base of each lobe bearing a few to many slender fimbriae, free or united at the base to form a fimbriate scale; interstaminal glands at the base of the tube, scutiform. Stamens included, the filaments inserted in the lower half of the corolla tube, the wings 0.5—0.75 mm. wide at the base, slightly tapering above; anthers oblong, 0.75—1.5 mm. long. Pistil sessile or shortly stipitate, the gynophore up to 0.5-1.0 mm. long; ovary cylindrical to cylindric-ovoid, about 8-15 mm. long, 1-4 mm. mide; stigmas sessile, elliptic to oblong, erect or recurved, 0.7-1.5 mm. long, 0.5—0.7 mm. wide. Capsule as long as the marcescent corolla or slightly exceeding it, dehiscing at the tip, the valves slightly spreading. Seeds variable, ovoid, slightly flattened, 0.5-1.0 mm. long, about 0.75—1.0 mm. wide, the surface smooth, Throughout the northern temperate and arctic zones of Europe, Asia, and America, segregating into a large number of populations. In North America represented by five subspecies extending from the Arctic coasts and interior of Alaska to central Mexico; northeastward to Newfoundland and Maine, but not found in the greater Mississippi River basin nor in the eastern or southeastern United States. The typical subspecies, amarella, occurs in central and western Europe and appears to be distinguished from ssp. acuta chiefly by its large flowers. Since ssp- amarella does not occur in North America, it is not described in this revision. KEY TO THE SUBSPECIES a. One or more arek ihe die foliaceous and SCC, the Sa orob Dee wt with few (3 di d d adjacent M TOR 11a. G. pend a ect 4 oliac H f ; stems wit 5-12 or more internodes; seeds 0. m di e p ona fimbriae free; stem: b. Lateral veins of the calyx lobes esse pech pa uniting towards the base of e tube. Hon variable subspecies with numerous loc al Newfoundland, Maine, and Vermont, westward to Alaska, ake ward in the : ¥ mountains to Baja California x dais ng Mexico. n lib. G di ssp. acu ON eter or rallel ^ some distance — the sinus, 1957] GILLETT—REVISION OF GENTIANELLA 251 bb. agai veins ei the calyx lobes uniting grape below the sinu Calyx 5-11 mm. long, the tube 2-3 mm. long, the lobes thinly membranaceo "E lowers in "pies Or aggregate cymes or la ry queens 8-12 Rus mauve, F with essentially erect corolla lobes. Southern and central Mexico. ee US CY G. amarella ssp. mexicana dd. Flowers chiefly in terminal umbelliform aggregate cymes, y ive m. long, white with spreading purple-speckled corolla lobes. North-central Mexico: EE 11d amarella ssp. sandiana cc. Calyx 12-15 mm. long, the tube 3-4 mm. long, the lobes somewhat coriaceous. e. Flowers 20-25 mm. long, frequently borne in compact clusters in the si = upper leaves; aS 25-70 cm. tall. Northern Mexico and "abes rey Mex e rella mg i l ee. Flowers 15-20 mm. long, in compact or loose clusters throughout SC Sue plants 5—40 cm. tall. South-central Mexico in Michoacan and Mexico. 11f. G. amarella ssp. bartwegii * s A : ; G. amarella ssp. heterosepala grades into ssp. acuta with numerous int diate forms. lla. GENTIANELLA AMARELLA (L.) Börner, ssp. heterosepala (Engelm.) J. M. Gillett, comb. & stat. nov. Ee Ee ae Engelm. in Trans. Acad. Sci. St. Louis 2:215. 1862. (T.: Engel- . s.l Bëbee dist stegia Greene, Pittonia 4:182. 1900. (T.: Baker s. n.!) Amarella beterosepala (Engelm.) Greene, Leafl. Bot. em & Crit 1: 53. 1904. Amarella scopulorum Greene, loc. cit. 55. 1904, ex c ien bolyantba A. Nels. in Bot. Gaz. 56:68. im. TE: Walker 513!) r. U.S. Nat. Herb. 25:415-416. 1925. ui common in Utah and extending into Colorado, Wyoming, New Mexico, Arizona and Idaho; in moist meadows and open aspen forests, from 2600 to 3200 meters; flowering from mid-July through August; fruiting from late August until September UNITED STATES: Mts., Kearney & Peebles 12412 ARIZONA: Apache Co.: Hannagan Meadow, White (US). Coconino Co.: Grand Canyon of Ss Colorado, geren 173a (GH, US); San Francisco Mts., Lemmon & Lemmon s. n. (U COLORADO: Boulder Co.: idee to d, baie? 5734 (DAO). Delta Co.: Le- roux Creek, Cows» 2 (GH). ores Co.: nbey (GH, MO, W Co.: Gunnison Forest, p hj os 14631 (GH, US). Krieg Co.: near Pagos poe Baker 5m. end eor 518 (G NY, US). "Montto se Co abeguache ce eg Pay. (GH, MO). San Miguel Se Si Springs Mesa, Walker 513 (GH, RM, Y AHO: Nez Perce Co.: Zaza, St. John & Muller 8634 pa . NEW MEXICO: ee : Black Bansa N Metcalf 1257 A e US). Otero Co.: Sacra- mento Mts., ii hes ee j. a Co.: Brazos Canyon, Gi & Bellman De s US). Sender Cor Sada Mo. Bl 235 (MO, NY, US). San Juan Co. ner Get Standley 7716 (US). uche "E eter m Canyon, Miss 942 (GH). Emery Co.: Muddy Creek r region, Tidestrom 499 and Co.: Sal Mts., Payson & Payson 4073 (GH, ar Breaks SE eeng Gould 2051 ( W US); DEER xdg n. c D ES Te "(MO “Pi Co.: Tate Mine, Marysvale, Jones s. : Seen San Juan Co.: La Sal Mts., Purpus 7033 (MO, US). Sevier it Co.: Uintah ta or (M h Mts., Stokes s. n d > atch cs aes Peak, Wasatc v o LM udi Likes Mountain, Fish L e Nat. Forest, Jobnson 81 A jas ; WYOMIN' Horn Co.: Worthley 59 (US). Lincoln Co.: Cottonwood Lake, e. o : Big Smoot, P nde & neis 3765 (GH, MO). [Vor. 44 252 ANNALS OF THE MISSOURI BOTANICAL GARDEN or 7f Br Ca Ld * I al » 50] 3 vw S Lu ` R zl / o, E . d x AA Leg d on .[9 oi : : A 53 ? A 4 bd SL 1 JJ d E i 1 G. amarella : @ ssp. acuta A d. |O sse heterosepola x TP e heterose : 2 Lu Fig. 7. Distribution of the subspecies of Gentianella amarella in North America. 4 1957] GILLETT—REVISION OF GENTIANELLA 253 Engelmann's Gentiana heterosepala is reduced to subspecific rank because neither the characters are sufficiently stable nor the range sufficiently isolated to warrant recognition as a species. A center of plants of somewhat pure ssp. beferosepala occurs in Utah, New Mexico, and Colorado. However, there are inumerable inter- grades with ssp. acuta at the borders of its range. Plants of ssp. acuta throughout Wyoming and as far north as Montana show traces of heterosepala characters and various forms from this population have been described as species. Most of these I am placing in the synonomy of ssp. acuta since they lack the foliaceous calyx and many other characters of ssp. heterosepala. Undoubtedly this problem is too com- Plex to solve completely by herbarium methods and it will be necessary to make further studies to define the groups properly. The singular foliaceous calyx of ssp. heterosepala is paralleled in Europe by G. campestris. Heterose pala tendencies observed in this population include: Tendency toward larger flowers. 5. Larger, fewer seeds. . Thickening of the s 7. Larger cauline eg Gemen by a trend from acute to obtuse lea Fusion of faucal pais 9; — of the upper nodes 10. Trend from triangular to oblong calyx lobes. ON Enlargement of one or more calyx lobes. S leg stem nodes. > ww N m . NT. 19. estion of flowers by a shortening zi Ges branches ea? ra NET e Rydb., — greatly affected by babisst e Shortening of upper internodes is a trend found also in G. propinqua and in G. aurea. lib. GENTIANELLA AMARELLA (L.) Bórner, ssp. acuta (Michx.) J. M. Gillett, comb. & stat. nov. Gentiana acuta Michx. Fl. Bor. Amer. 1:177. 1803. (T.: Michaux s. n. photo MO!) Gentiana axillaris Raf. Med. Fl. 1:213. 1828, ex char. Gentiana plebeja Cham. ex Bunge, in Mem. Soc. Nat. Mosc. 7 (Nouv. Mem. Soc. Nat. Mosc. 1): 250. £. 29. f. 5. 1829, ex ic. Amarella acuta (Michx.) Raf. Fl. Tellur. 3:21. 1837 Ericala acuta (Michx.) G. Don, Gen. Syst. 4:190. 183 den acuta che B stricta Griseb. Gen. & Sp. Gent. 242. 1839. (T.: Drummond » photo Gentiana tenuis Grin. loc. cit. 250. 1839. (T.: Richardson s. n. E Mou. Gentiana aggregata Bunge, ex Griseb. in DC. Prod. 9:100. 1845, nom. entiana acuta Michx. Gentia r. nana Engelm. in Trans. Acad. Sci. St. foo. 2: 214. 1862. (T.: Parry 309!) Gentiana vein L. var. acuta (Michx.) Herder, in Acta Zeg Sege I: +428. 1872. Gentiana amarella L. var. tenuis (Griseb.) A. Gray, Syn. Fl. N r. ed. 2, 21:118. 1886. Gentiana enisótjela Greene, Pittonia 3:309. 1898. ‘(Heller & SA er 3440!) Gentiana acuta Michx. ssp. acuta Wettst. in Oesterr. Bot. Zeitschr. 50: T 1900. entiana acuta Michx. j .) Wettst. loc. cit. 194. 1900. Genti 4 Michx. wp, Webas, iei 1900. (based on G. =e Engelm.) d Tiea acuta Michx. var. strictiflora Rydb. Mem. N. Y. Bot. Gard. 1:309. 1900. (base - acuta B stricta Grise Gentiana stricta (Griseb.) Howell, FI. N.W. Amer. 1:445. 1901. Gentiana strict b.) A. Nels. in Bot. Gaz. 34:26. 1902. Genti janelle ci imei Ce Bull. zi" Bot. Club 31:631. 1904. (T.: Clements € Clements 2 ; Amarella SE (Greene) Greene, Leafl. Bot. Obs. & Crit. 1:53. 1904. [Vor. 44 254 ANNALS OF THE MISSOURI BOTANICAL GARDEN Amarella conferta Greene, loc. cit. 55. 1904. (T.: Spreadborough s. n.!) Amarella copelandi Greene, loc. cit. 53. 1904. (T.: Copeland s. n.!) mare j it 904. < Lembert s. al rella macounii Greene, loc. cit. 54. 1904. (T.: Macoun s. n.!) Amarella revoluta Greene, loc. cit. 55. 1904. (T.: Wooton 5 521) Amarella strictiflora (Rydb.) Greene, loc. cit. 53. 1904 Amarella tenuis (Griseb.) Greene, loc. cit. 53. 1904. marella amarella (L.) Cockerell, in Am. Nat. 40:871. 1906. Amarella plebeia var. bolmii (Wettst.) Rydb. in Bull. Torr. Bot. Club 33:148. 1906. Amarella gurliae Lunell, in Am. Midl. Nat. 2:142. . (T.: Lunell 816!) Amarella tbeiantba Lunell, loc. cit. 143. 1911. (T.: Lunell 818!) Amarella theiantha Lunell var. livida Lunell, loc. cit. 142. 1913. (T.: Lunell 820!) Amarella theiantha Lunell var. lactea Lunell, loc. cit. 1913 (T.: Lunell 819! Gentiana amarella L. £. michauxii Fern. in Rhodora 19:151. 1917. (based on G. acuta i ensu stricto X. Gans penta (Michx.) Hiit. in Mem. Soc. Faun. Fl. Fenn. No. 25:76. 1950. Newfoundland, Maine and Vermont, westward to Alaska, southward in the western mountains to Baja California and central Mexico. Some of the large number of reported habitats with flowering dates and eleva- tions, are as follows: Alaska: Rocky shores, sand-dunes, sea-beaches, tundras, lake shores, moist grassy swales, apparently at rather low altitudes; flowering and fruiting from June to eptember. Alberta: Roadsides, open poplar woods, alkaline meadows, dry prairies, moun- tain slopes, sides of coulees; late July until September. | Quebec: Wet fields, calcareous gravel, limestone, turfy slopes, sandy headlands, above the tide mark, near sea-level; late June to early September. Wyoming: Along creeks, open woods, alpine meadows, alkali areas up to 3000 meters; late July and throughout August. California: Grassy banks, poplar woods, grassy places in yellow pine forest, to 2500 meters; early June to September dependent upon altitude. Arizona: Dry or moist meadows to 3500 meters; flowering and fruiting from mid-August until September. (D 6023 (DAO); Keg River, Grob 2934 (DAO); Mt. Temple, Loggan, Butters, Holway / , US) ; Pincher Creek, Gillett 6026 (DAO); Waterton Lakes Nat. Park, Gillett 6029 (DAO 's Nest S); Woo m Beaverlodge, Lindsay 303 (DAO); Grande Prairie Malte ; Malte . SE : 40 (CAN); Edmonton, s.n. (CAN); Caribou Mountains, Raup 3035 (CAN); Rosedale, Moodie 1209 (F); Peace River, Indian Graveyard C i 303 E . FJ > p 8 2 - = = 5 > E AO); A ; Prince George, Groh 636 (DAO); Hudson Hope, Peace River Dist., Grob 745 (DAO); Skaget, David . ei Revel- stoke, Hitchcock & Martin 7375 (GH, NY. y, elu BC); Victor Lake, near Reves 1957] GILLETT—REVISION OF GENTIANELLA 255 WS); Cassiar Dist., Telegraph Creek, Preble & Mixter 577 (US); Mt. Robson Se Benin 3322 (GH, US); es Mts., Shaw 180 WS); Caribou, 150 Mile House, Wilson 717 (UB Foghorn Mt. s. of Waverly, Se s. n. (DAO); Tran- ço 1- Ane a s.n. (UBC); Paradise Mine, Windermere, Calder & Savile 11 300 (DAO); Takakkaw Falls, Yoho River Valley, Calder & Savile 12016 (DAO); Chute Lake, Naramata, Calder & Savile 10202 (DAO); Casca de, Calder & Savile 9539 (DAO); Quesnel, Taylor s.n. (UBC); Fairmont Hot Springs, Eastham 10137 (UBC); Williams Lake, Dog Creek Range, Carlyle s.n. (UBC); Queen Charlotte nie Masset, Newcombe s.n. (F); i DA F "ëng 14668 (D DOR: Bleac Zoe Straits of Belle Isle, Fernald * ye es 3905 (GH); Cart- wright, Hitchcock 2 Ss (US) ; L'Anse au Clair, oerte ACKENZIE DISTRICT: Lac Ste. Croix and Bear Lake, Schi 268 (US) t Smith, Preble E) Preble 171 (US); Yellowknife, Cody & Mate Pe (DAO): iade Buffalo River, Loan 327 (DAO); Seven Mile Lake, Cody & Loan 4661 (D MANITOBA: Lyleton, aue Ei äerer? 6588 (DAO); d. Mt., Burgess s.m. (DAO); Brandon, Macoun (F); peg, Denike 1689 (DAO); Wabowden Lake, Mile 137 Hudson Bay R. x ` Gillett Déi (DAO); Dauphin, Groh s.n. (DAO); Stony Mountain, Groh s. n. "oa 0); Riding Mountain Nat. Park, Rowe 60A (DAO); Cart e Senn & Gordon 3036 (DAO); Broomhill, Senn zé Gordon 3120 (DAO); Melita, Senn & Gordon 3102 (DAO); Cláplin, Spreadborougb s. ND NEW BRUNSWICK: Gloucester Co.: Grande Plaine, seg Island, Blake 5504 (GH, US), Ganong s. n. (GH) ; Goat Island, mouth of St. John River, Hay s. n. (PH). every oe Cartie oi eg 715 (DAO); St. Barbe, Fernald, Long, et al. 26972 (GH, PH); Ingornachoix ernald & Wiegand 3902 (GH, PH); St. de GR Savile së (DAO); St. Lë AM Bard Harbour, Wiegand & Gilbert 28931 (GH, 4 RIO: Sc mee eid Lake Superior n. shore, Dore 9236 (DAO); St. Ignace Islands, Lake $ * Superior, Asa s. 1. (PH). Chis: Calder ap (DAO, US); Anticosti, Baie Sainte-Claire, Vic- Ge Pes OE US); Riviére au Saumon, Vicforin et al. 21064 (F); Bonaventure Co.: ata, Port George, Lepage 12851 (DAO); Rupert House, Spafford 160 (DAO); Pringle s: n, ; MO, PH, US) ; Douglastown, rege & Pease 6605 (F); dns Minute fin , Fernald St. John (PH, US, WS); Mingan Islands, fle de Sainte- Genevieve, Vict & Rolland 21063 (GH, , PH) ; Riviére du Loup, Canby s. n. (F) ASKATC side, Blaricom 26 (F); eton House to Cumberland House, 6501 (GH); R Cowd (DAO); Yorktown, e E (D JA AO); Sut Sarhérand; Russell s. n. (DAO); ague, Breitung 476 (DAO) ; Invermay, Bees 51794 (DAO); St. Louis, Senn, Grob & Russell 2838 0; Manitou. Ee. Frankton t$ Bibbey 413 (DAO); Bjorkdale, Ge 8 Xi ers IDAOY. Albert, Senn, Groh & Russell 2817 D ; Villovbunch, vun tt Parsing wad Eege ; B : "am K gw ZEN Gillett & Boivin 6010 (DAO); Whitewood, x Ge GE M E Ca e à la ; Churchill River, Dao 8427 (DAO); NEE Su Grob a R ER ry 1 2898, ie Rockglen, Gillett & Boivin eek ta 2. YUKON; Dawson, Mesas" ); Bear Creek near Lake Desert d'Asch, Mä s.n. (PH, US). [Vor. 44 256 ANNALS OF THE MISSOURI BOTANICAL GARDEN NITED STA ALASKA: Fore Se Kennicott s. n. (F) ; Hiulink, Harrington s. n. (US); Unalaschka, Harrington s. n. (MO); Fairbanks, Palmer 1826 (US); ries Island, K Kellogg 246 (PH US); Juneau, n 1096 (F, GH, MO, US); Karluk, Kodiak Island, Rutter 71 (GH, NY, US) ; Homer, Evans 499 (U S); Dutch Harbor, Unaloskx Island, True & Prentiss 129 (US); Lake Kenai, Kenai Peninsula, Calder 5587, 6335 (DAO); Cape Phipps, Funston 72 (F, GH, MO, US); Katmai Region, Alaska Peninsula, Hagelbarger 165 (US); Attu Island, Sarana Bay, Van Schaack 1000 (MO, US); Behring Island, Macoun s.n. (US); Alitak, Looff 5 Loof 774 (UBC); Adak Island, Dorwart 45 (F); Dauer Bay, Cooper & Andrews 313 "i 2n 3 : Apac s Co.: White Mts., Eggleston 15771 (F, GH). Coconino Co.: Kaibab Kee Grand Canyon Park, Kearney & Peebles 13727 (US). Graham Co.: irs Mts., Darrow, Phillips 3 ier 106 (US). Yavapai Co.: Lemmon 4200 (GH). a Co.: Buckskin Mts., Jones 60 (US). CALIFORNIA: Butte Co.: Butte Meadows, Heller 14689 (MO, NY, pues dra Copeland Mi Er Inyo Co.: Whitney Meadows, Coville & Funston 1631 pos T mne, Yosemite Nab Geer Heller 15464 (MO, NY, PH, US). Ré Lu Bloody Ve ag Chesnut & Drew LOS Nevada Co.: Williamson s. n. (PH). Placer Co.: Summit Station, Sonne s. n. “oO S); Truckee, Sonne 217 (F). San Bernardino Co.: Pine Lake, Abrams 2897 (GH, MO, PH, US). Plumas Co.: Austin s. n. (US) ; Bear Lake, Clokey 7000 7000 (F). Shasta Co.: Newberry s.n. (US). GE beet Lemmon 190 (NY). Siskiyou Mount Eddy, Copeland s.n. (ND). uolum .: Tuolumne River, Bolander 5045 (F, MO, NY, PH US); Yosemite SH Cer s.n. (ND). Tulare Co.: Kern Ri NY H COLORADO: Boulder Co.: Boulder, Holzinger s.n. (US). Chaffee Co.: Garfield, Eggleston 6068 (GH, US). Clear Creek Co.: Georgetown, Patterson 287 (F, GH, MO, US). Custer Co.: Wet Mountain Valley, Horner s.n. (GH). Denver Co.: Glaifelter ont Co.: H). Garfield Co.: Hermann 5580 (MO). Gilpin Co.: Black Hawk, Sheldon 286 (PH, ren Berth “ts Pass, Clokey & Clokey 4241 (CAN, MO, UC, US, WS). Gunniso son, r 600 (GH, MO US WS). La Phra Co.: Duran wiit Baker et er: wen (MO, NY Js). Routt Co.: Steamboat Springs, Crandall s. n. (US). Hinsdale Co.: Lake City, Pease s. n. (NY). Jackson Co.: Lake John, Shear & SEN 4335 ^ (US). Jefferson Co.: Golden, Greene s. n. (GH). Lake Co.: Twin Lakes, Clokey 2548 (CAN, MO, NY, US, WS). Larimer Co.: Senge Lake, Baker s. n. (MO S) ineral Co.: Murdock r 1502 (CAN, US, WS E rend, t anco 02 L Pitkin Co.: Sawatch Range, Brandegee s.n. (MO). Rio Blanco Co.: North Elk Gear Sturges s. n. (GH). Rio Grande Co. : Del Norte, Evermann s. ? (US). Saguache Co.: Marshall Pass, Baker 88 5 (GH, MO US). San Juan Co.: Head of Vallecito, rere 44 (US). i : , ves mit Co.: one Mt. Guyot, Anderson s.m. (MO). Location ? € Co.: Hyndman C eek, Thompson 13520 (PH, MO, US, WS). Boise WS Sp prings Creek, Woods Sg Tidestrom bes (U Us). E ary Co.: Hughes Fork T Trail, Warren 368 (WS). Custer Bonanza, Macbride t3 ee 3475 page MO, US). Frem remont Co.: Henry’s Fork, Coulter s. n. ^ pH. US). Kootena herd 424 . - Latah Co.: Ced Donga DOOR erue BE ea : » Hifcbcoc j ; : Heller & Heller 3440 (D M 0, MO, ND, NY). GH, NY, US, WS). Nez Perce Co.: Forest | n. (GH). ONTANA: Co.: Pioneer Rang e itchcock & Mu Pii k 13030 (GH, MO UC, WS). Carbon Co.: Red Lodge, Hitchcock s Mublick 13554 SC Flathead Co.: Big Fork, Jones s.n. (G H, MO, U US). Gallatin Co.: atin Canyon bip 349 (PH, US); Glacier Co.: 3 mi. s Babb, Daubenmire 48381 (WS); Glacier Nat. Park, 1957] GILLETT—REVISION OF GENTIANELLA LaF Bowman, Evans s. n. EN Granite Co.: Garnet, Scheuber s.n. (NY). Lake Co.: Big Fork, Zeg s.n. (GH). Lewis & Clark Co.: Flathead Nat. Forest, Greg 18743 dee ). comes en Ennis, Hitchcock 16736 (UC, WS). Meagher Co.: Martinsdale, Canby MO). Missoula Co.: V codi, Kirkwood 1475 (GH). Park Co.: Wilsall, Said Ss (GH, WS). Powell Co: Flathead Nat. Forest, Hitchcock 18 549 (ID, W. t Grass Có: MacLeod, Pope 105 (NY). Teton Co.: Glacier Nat. Park, A: Elko Co.: "Ruby M ries 0467 (US). NEW MEXICO: Catron Co.: n Peak, eege OE deen Rusby 262 (F). Colfax Co.: Castle Rock, ës 4 n. ser US). e Mts, Wooton 552 (MO, US). Sandoval Co.: Jemez Mts., Bailey peers `, B5. d Mind Co.: Pecos, Standley 50rr (GH, MO, US). Santa Fe Co.: Bertaud 24 (US). Taos Co.: Carson (US). NORTH DAKOTA: Benson Co.: Leeds, Lanell 816 (MIN, NY), Lunell 818 (MIN); Butte, Lunell 820 (MIN). Bottineau Co.: Turtle Mts., Stevens 540 (DAO, or 43974 Divide GE Fortima, Metcalf 597 (US). E E Co: "Tower, Lunell $19 (MIN). OREG Crook ue Meca Coville & Apples 541 (US). be oem Co.: Big Meadows, ities 496 (GH, MO). Jefferson Co.: Metoline River, Gorman 3887 (WS). Klamath Co.: Fort Klamath, Peck 9557 (MO, PH). Union Co.: Blue Mts., Cusick 1020 (GH, US); Wallowa Mts ;j Cus ck 2101 (F, MO, US, WS). eni DAKOTA: Canes Co.: Custer, Degener & Peiler s.n. (MO, US). Lawrence Co.: Lead, Car 144 (DAO, GH, MO). Pennington Co.: Redfern, kr 4285 (GH, US); Mystic, Visker s. n. (F). UTAH: Beaver Co.: Tashar Mts., v Are 10417 (US). Cache Co.: Tony Grove c Maguire. 16996 (WS). Duchesne Co.: Krobs Basin, Hermann 5190 (GH, MO). field Co.: Aquarius Plateau , Ward 440 (MO, US). Kane Co.: Siler s.n. (MO). Sale hake Co.: Bright ton, Maguire 17317 2 US, WS). Sanpete Co.: Watson 59 (ID). t Co.: Fish one Jones 5823 (GH, MO, NY, US, WS). Washington Co.: Spring- dale, J Jones 6078 (M VERMONT: Daer? Co.: Mt. Mansfield, Smuggler's Notch, Pringle in 1871 (MO, NY, PH, US); and many other collections in most American herbaria. WASHINGTON: Clallam Co.: Mt. Angeles, does 7357 (GH, MO, PH). Ferry Co.: Coville, Hitchcock tate (DAO, ID, US i osi n Co.: Olympic Mts., Marmot Pass, Tbompson 7088 (GH, MO, PH). King Co Seattle, Freiberg s.n. (MO). ogan Co.: Salmon Meadows, Thompson 3 (GH - Oreille Co.: St. 6462 (WS). San Juan Co.: Fri E Harbor, Zeller 8 D. 918 (GH, MO, US). Skagit -= Dewey, Mason s. n. (WS). Stevens Co.: Arden, Spiegelberg 267 (WS). Whatcom Co.: Padden Lake, Suksdorf 990 (F, GH, MO, WYOMING: Yellowstone Nat. Park, Snake River, Nelson & Nelson 6447 (GH, MO, US, WS). Albany Co.: Centennial, Nelson 8803 (GH, MO, —« Big Horn Co.: Middle Ten Sleep Creek, Qood 471 (GH, US). Carbon Co: Rawlins, Goodding 570« (MO, US). Park Co.: B rtooth Lake, Williams & Williams 3755 (GH, MO, NY). Shoshone i. from Colt City, Iltis 3996 (DAO). Sheridan Co.: Spring Creek, Williams NY). lett yo Payson 2903 (GH, MO, US). Teton Co.: Teton Mts., Porter 4741 (DA Mt. View » Maguire, Piranian t$ Richards 12643 (GH). Washakie Co.: Big Horn Mts. W Porter 4343 } (GH, MO). Weston Co.: Newcastle, Nelson 2551 (F). BA pes SE La Sanca Creek n.w. La Grulla Sierra, San Pedro Mart Wiggins $$ Demaree io (F, GH); La e nr. Sech Sierra San Pedro Márir, Keng 907 75 (GH : dix er ` Nevada end "Toluci, Moore 89a (GH), Kenoyer s. n. EON: Municipio de Galeana, peak of Cerro Potosi, Schneider 971 (F, GH), iui, 4 dig (F, GH). [Vor. 44 258 ANNALS OF THE MISSOURI BOTANICAL GARDEN For the combination Gentianella amarella (L.) Bórner and other Borner refer- ences I am indebted to Dr. Harald Smith, of Uppsala, who pointed out to me that Bórner had made the combination earlier in his Flora. This Flora, although recent, is not widely distributed on this side of the Atlantic. The Library of Congress has a copy but I know of no other in the United States. G. amarella ssp. acuta is extremely variable, particularly in the number of inter- nodes, the size and shape of the calyx lobes, the length of the corolla tube and lobes, and the density of the inflorescence. Field observations indicate that great variabil- ity in height of the plant, density of the inflorescence, color of the flowers, and other features is found in local populations as well as in groups of plants of the range as a whole. The pattern of this variability is not sufficiently well understood at this time to warrant the recognition of additional forms or subspecies. It has been necessary to establish the relationship of ssp. acuta to the Euro- pean population of Gentianella amarella (sensu lato). 'The few European specimens of G. amarella that I have seen show an astonishingly great variability. Many "species" have been based upon the variants of this European population; many of them are difficult to distinguish morphologically, and perhaps should be reduced to subspecific or lesser rank. Most of the typical specimens of the European G. amarella have somewhat larger flowers than do the American plants of ssp. acuta. Since G. amarella does not occur in Greenland, the American and the European populations are disjunct. Because of this geographic disjunction and the difference in flower size, I am recognizing the American population as subspecifically distinct from typical G. amarella of Europe. The few specimens I have seen from Siberia and other parts of Asia show close resemblance to G. amarella ssp. acuta. ‘These Asiatic plants have been described under the name Gentiana ajanensis by Murbeck. However, I shall not consider the Asiatic population as synonymous with ssp. acuta until I have seen more material from that region. llc. GENTIANELLA AMARELLA (L.) Bórner, ssp. mexicana (Griseb.) J. M. Gillett, comb. & stat. nov. Gentiana mexicana Griseb. Gen. & Sp. Gent. 243. 1839, ex char. (T.: Schiede s. n.) Amarella mexicana (Griseb.) Arthur, in Torreya 12:34. 1912. Southern and central Mexico; in wet meadows and on gravelly slopes neat timber-line and along creek banks at elevations of 2000-3000 meters. Flowering from mid-September through October; fruiting in October and November. Mexico: FEDERAL DISTRICT: south of Contreras, Russell ej Souviron 197 (US); Desierto de los Leones, Lyonnet 517 (GH, MO, NY, S); La Cima, 9800 ft. (ag & Land 383 (F) cn ALGO: Pachuca District, Min f ( MEXICO: Temascaltepec, Hinton $313 (F, GH, MO, NY, US); Crucero-Agua Blanca Temascaltepec, Hinton 8330, 211 F } ; ren ei ) eme Pringle 4277 (F, MO, NY, PH,US), ^ > GH, MO, NY, US); Sierra de las Cruces eral del Chico, Parque Nacional El Cluco, Moore 1607 9? Murbeck, in Acta Hort. Berg. 23:1-28. 1892. 1957] GILLETT—REVISION OF GENTIANELLA 259 Oaxaca: Sierra de Calvellinas, 9000 ft., Smith 667 (MO, US). PUEBLA: Ixtaccihuatl, Purpus 1757 (F, MO, US). VERACRUZ: Cofre de Perote, 9300 ft., Balls 5435 (US). LOCALITY UNKNOWN: Ehrenberg 79 (GH). lld. GENTIANELLA AMARELLA (L.) Börner, ssp. sandiana J. M. Gillett, subsp. nov. Plantae 2.0—3.5 dm. altae ramosae. Folia basalia elliptica vel spatulata 4—10 mm. longa jam marcida superiora ovata vel ovato-lanceolata, 10—25 mm. longa 5-10 mm. lata, acuta. Flores numerosi, 9-17 mm. longi, aut infundibuliformes aut hypercrateriformes cum lobis maxime dilatatis sessiles aut pedicellis ca. 8 mm. longis in cymis terminalibus axillaribusve dense aggregatis umbelliformibus vel corymbiformibus. Calix 5-17 mm. longus, tubo 2.0—2.8 (—3.0) mm. longo, lobis lineari-oblongis acutis subaequalibus quam tubo aliquantum longioribus sinibus rotundatis margine minute denticulato. Corolla alba 10—15 mm. longa lobis albis aut purpureo-maculatis basim fimbriatis. Anthera 1 mm. longa 0.5 mm. lata. Ovarium sessile ca. 12 mm. longum 1.5 mm. latum; stigmata oblonga patula vel erecta ca. 0.5 mm. longa. Semina immatura rotundata laevia maturitate ignota. Restricted to Chihuahua and Durango. Habitat probably similar to that of G. wislizeni for some material is from the same locality. At an altitude of about 2500 meters, flowering in October. EE Palmer 334 (GH, NY, US); 65 mi. east of Batopilas, Goldman 188 (US). DURANGO: Mesa de Sandia, Pringle ro111 (F, GH, MEXU, MO nororvrr, NY, US). Very similar to G. wislizeni in many respects but having a distinctly tubular calyx. The dense terminal corymbs of fimbriate flowers with the speckled corolla lobes are striking features of this subspecies. lle. GENTIANELLA AMARELLA (L.) Bórner, ssp. wrightii (A. Gray) J. M. Gillett, comb. & stat. nov. Gentiana wrightii A. Gray, Syn. Fl. N. Amer. ed. Amarella wrightii (A. Gray) Greene, Leafl. Bot. Amarella cobrensis Greene, loc. cit. 56. 1904 Gentiana townsendii Briq. in Candollea 4:329. 1931. Northern Mexico and in adjacent New Mexico; flowering from late August until the middle of November, at altitudes of 2100—2400 meters, in moist and springy places. 2, 21:118. 1886. (T.: Wright 1659!) Obs. & Crit. 1:53. 1904. (T.: Greene s. n.!) (T.: Townsend & Barber 358!) UNITED STATES: NEW MEXICO: Grant Co.: Pinos Altos Mts., Greene s. n. (F, MO, NY, PH). Mexico: e j A hihuahua, Palmer 306 CHIHUAHUA: Sierra Madre, Jones s. n. (MO, US); southwest C , - Hà » GH, MO, US) ne Colonia Garcia, Townsend & Barber 358 (F, GH, MO, NY, US); ennell 18744, 19116 (US). ‘ SINALOA: Cerro del Viejo, San Ignacio, Montes & Salazar 83 (US). SONORA: Santa Cruz, Wright 1659 (GH, MO, NY, PH, US). LOCALITY UNKNOWN: Cima, Orcutt 3800 (F). [Vor. 44 260 ANNALS OF THE MISSOURI BOTANICAL GARDEN Subspecies wrightii is extremely variable. Greene’s Amarella cobrensis is very similar to the material collected by Wright. Gentiana townsendii appears to be but a form with exceptionally dense axillary inflorescences. Since so little material of G. townsendii is available, I do not feel that much significance can be attached to this variation, because the flowers appear to be identical in every respect to other sheets of ssp. wrightii. Gentianella amarella ssp. bartwegii and G. amarella ssp. wrightii are related to the more northern G. amarella ssp. acuta, but are distinguishable quite easily by the longer calyx tube and the nature of the lobes. The calyx lobes of ssp. acufa are thin, as are those of ssp. wrightii, and the margins are little differentiated from the central part of the lobe. In ssp. acuta the midrib of each lobe is not prominent and appears as a thin nerve. In ssp. hartwegii, however, the lobes are somewhat coriaceous, the margins somewhat thickened, and the midrib of each lobe broad and papillose, forming a flattened keel which extends the length of the calyx tube and is paralleled by a similar keel proceeding from the sinus. The midribs of the calyx lobes of ssp. wrigbtii are intermediate between those of ssp. acuta and ssp. bartwegi. 11f. GENTIANELLA AMARELLA (L.) Börner, ssp. hartwegii (Benth.) J. M. Gillett, comb. nov. Gentiana hartwegi Benth. Pl. Hartw. 47. 1840. (T.: Hartweg 351, photo MO!) Gentiana citrina Pollard, in Proc. Biol. Soc. Wash. 13:130. 1900. (T.: Pringle 4196!) xicana Griseb. ssp. hartwegi (Benth.) Wettst. in Oesterr. Bot. Zeitsch. 50:291. 1 , Gentiana mexicana ssp. hartwegi f. pringlei Wettst. loc. cit. 291. 1900. (T. Pringle 42, j Amarella bartwegi (Benth.) Arthur, in Torreya 12:33. 1912. In the states of Mexico and Michoacan, Mexico; in wet mountain meadows at altitudes of 2800-3300 meters; flowering from mid-August until October; fruiting in October and November. co: MEXICO: Nevado de Toluca, Kenoyer A F); c, Hinton 1317, 6320 (F, GH, MO, US); Valley of Tolucz, Privala véi? een MO, NY, U5), Pringle 4237 (F, GH, MO, US); Desierto Vieja, Valée de Mexico, Bourgeau 1125 (GH). ACAN: gangueo, Hartweg 371 (NY, photo F, MO). LOCALITY UNKNOWN: Sessé, Mocino, Castillo t$ Maldonado 684, 1370 (F). Wettstein considered ssp, hartwegii a subspecies of G. mexicana. While all of the Mexican plants are ve | closely allied to ssp. acuta than to ssp. in uias) ons. My own field experience with o caer er in the plants of a continuous population reduction in size regularly occurs at higher altitudes, 1957] GILLETT—REVISION OF GENTIANELLA 261 12. GENTIANELLA auriculata (Pall) J. M. Gillett, comb. nov. Gentiana auriculata Pallas, Fl. Ross. 12:102, t. 92, f. T. 1788, ex char. & ic. Dasystephana auriculata (Pall.) Borckh. in Roem. Archiv f. Bot. 11:26. 1796. Hippion auriculatum (Pall.) Schmidt, in Roem. loc. cit. 11. 1796 Amarella auriculata (Pall.) Greene, Leafl. Bot. Obs. & Crit. 1:53. 1904. Caulescent annuals (?), 2-25 cm. tall, simple or branched, occasionally from the base, more frequently above. Basal leaves forming loose rosettes, obovate- spatulate to elliptic, the apex rounded or obtuse, the base attenuate, 4—30 mm. ong, 1-15 mm. wide; median and upper leaves elliptic-ovate to broadly ovate, obtuse to rounded, the base broadened and clasping, 8—40 mm. long, 3-15 mm. wide; the upper leaves very slightly smaller or equal to the median. Flowers axil- ty or terminal, in simple or aggregate cymes or occasionally solitary, the terminal flower conspicuously larger than the lateral, the terminal pedicels extremely vari- able, subsessile to 3.5 cm. long, the lateral equal to or correspondingly shorter than the terminal, Calyx 7-9 mm. long, 4- to 5-lobed (4-lobed in North America), the tube about 6 mm. long; lobes about equal, or either the inner or the outer pair broader than the others, strongly and reticulately veined, broader than long, about one-third the length of the tube, the margins entire or minutely papillose, the base auriculate. Corolla blue, narrowly funnelform to salverform when the lobes are fully expanded, 18-28 mm. long, 5-8 mm. wide at the orifice; lobes ovate, 14-34 the length of the tube, the tips obtuse to subacute, the sinuses acute and frequently concealed by the imbrication of the lobes, the tube provided with a faucal corona, the fimbriae united at the base to form a lacerate squamella extending across the base of each lobe; interstaminal glands inconspicuous and extremely small, scuti- form, at the very base of the corolla tube. Stamens slightly exserted; filaments inserted about the middle of the corolla tube, the wings about 1 mm. wide at the base, tapering gradually above; anthers oblong, 1.6-1.8 mm. long, about 1 mm. Vide, versatile and attached near the middle. Pistil sessile, the ovary cylindrical, 15 mm. long, 1.5 mm. wide, tapering slightly near the tip; stigma sessile, oblong, 1.5 mm. long, 1.0 mm. wide, erect. Capsule slightly longer than the marcescent Corolla, about 25 mm. long, dehiscing at the tip; seeds approximately ovoid, 0.75 E long, 0.6 mm. wide, smooth, minutely wrinkled under magnification, light rown Fig. 8. Calyx (X 2), corolla (X 1), and corolla lobe of Genti ella auriculata. [Vor. 44 262 ANNALS OF THE MISSOURI BOTANICAL GARDEN On Attu Island, at the western extremity of the Aleutian Islands, on the Com- mander Islands in Behring Sea, in Kamchatka, on the islands north of Japan and about the Okhotsk Sea; on gravelly slopes at altitudes from sea level to 300 meters; flowering throughout August and presumably well into September; fruiting in September. Further habitat data are lacking since all Asiatic material examined gives no other information than the approximate location of the collection. ALASKA: Aleutian Islands, Attu Island, vicinity of Massacre Bay, Lookout Hill, Van Schaack 66 (MO, US), 748 (MO), 929 (MO, US). The description has been based primarily on Kamchatkan and Japanese material since so little is available from North America. The excellent material from Attu Island collected by Van Schaack constitutes the only record of this plant from this continent. This fact has already been pointed out by Hultén in his flora. On a field label, Van Schaack notes that the specimens are 4-merous, while the material from Kamchatka and the Commander Islands is 5-merous. However, examination of more material shows that 4- and 5-merous material occurs both in Kamchatka and in the Commander Islands (Bering Island). As seen in other species the number of corolla or calyx lobes is a variable character, both types commonly occurring even on the same plant; indeed, in G. aurea and G. amarella ssp. acuta a 5-lobed calyx is frequently associated with a 4-lobed corolla. SuscENus III. COMASTOMA (Wettst.) J. M. Gillett, stat. nov. Gentiana sect. COMASTOMA Wettst. in Oesterr. Bot. Zeitschr. 45:174. 1896. Lomatogonium A. Br. sect. COMASTOMA (Wettst.) Lóve & Lóve, in Acta Hort. Goto- burg. 20*:117. 1956. Flowers 4- to 5-merous. Calyx tube short, the lobes foliaceous, sinuses lacking an inner membrane. Corolla 4- to 5-parted, the lobes ovate, entire, the orifice with two evascular fimbriate scales at the base of each lobe, the interstaminal glands epipetalous and in pairs. Stamens inserted slightly above the middle of the corolla tube, the anthers oblong, slightly longer than wide, the filaments minutely papillose. Ovary with several rows of ovules that frequently extend nearly to the center of the carpel wall. Seeds ovoid, somewhat flattened, smooth or slightly papillose at one end. 13. GENTIANELLA TENELLA (Rottb.) Börner, Fl. deut. Volk, 542. 1912. (G tenella (Rottb.) H. Sm. ex Hylander, in Uppsala Univ. Arssk. 259. 1945, superfluous comb.) Gentiana tenella Rottb. in Kiob. Selsk. Skrift. (Acta Hafn.) 10:436. 1770. (T: Col- lector unk; s Lomatogonium tenellum (Rottb.) Löve & Löve, Acta Hort. Gotoburg. 204:117. 1956. Annuals 1-26 (usually about 6) cm. t : à all, loosely to densely cespitose or branch- ing above, or simple, the branches curved-ascending. Basal leaves two or forming dense rosettes, frequently deciduous, elliptic, obovate to spatulate, 3-10 mm. long: 1-4 mm. wide, the apex rounded to obtuse, attenuate to the base; median and upper leaves ovate to ovate-elliptic, 4-9 mm. long, 1-3 mm. wide, the apex obtuse, 1957] GILLETT—REVISION OF GENTIANELLA 263 the base attenuate, somewhat clasping. Flowers solitary, terminal or axillary. Pedicels 2-10 cm. long. Calyx 4- to 5-merous, the tube nearly obsolete, 5-10 mm. long, the outer lobes foliaceous, broadly ovate to lanceolate, obtuse to acuminate, about 10 mm. long, 277 mm. wide, with or without an outer inflated saccate pro- trusion at the base, the inner lobes lanceolate, slightly shorter than the outer and about half to one-third as wide, the margins entire, the sinuses concealed by the imbrication of the lobes. Corolla blue to white, tubular to somewhat salverform, about 11 mm. long, 2—3 mm. wide at the orifice, the lobes oblong-ovate, imbricate, one-third to half as long as the tube, the tips obtuse, the orifice bearing at the base of the corolla lobe two groups of 4—6 blunt evascular fimbriae united at the base to form two fimbriate scales, or nearly free about half the length of the lobes, the interstaminal glands indistinct, appearing by transmitted light as two greenish, oblong stains about 1 mm. above the base of the tube. Stamens included, the filaments inserted near the upper third of the corolla tube, the wings about 0.3—0.4 mm. wide at the base, slightly tapering towards the anthers, the margins minutely Papillose; anthers oval, about 1 mm. long, 0.5 mm. wide, versatile and attached near the middle. Pistil sessile; ovary fusiform, about 5 mm. long, about 1.5 mm. Wide; stigmas sessile, ovate-oblong, 0.5 mm. long and wide, erect, the stigmatic surfaces facing inwards, recurving in fruit. Capsule longer than the marcescent corolla, up to 12 mm. long, dehiscing at the tip, the valves recurving about half their length. Seeds ovoid, slightly flattened, 0.75 mm. long, 0.5 mm. wide, the surface smooth, sub-papillose to papillose at one end, light brown. LS - - -~ x ES SEE e eh Za ` Re a EE WS. dal Wd H “a p SR d S, V Co. »— Gë AS 5 Py fS AED A Fig. 9. A, distribution of Gentianella tenella in North America. B, calyx and corolla of G. tenella, X 134. [Vor. 44 264 ANNALS OF THE MISSOURI BOTANICAL GARDEN Gentianella tenella occurs in northern Europe and Asia and in the mountains of the southern parts of those continents. In North America the species occurs in Greenland, northern continental Canada and Alaska; in the Rocky Mountains as far south as Arizona and New Mexico and in Mono County, California. Because the plants are small and inconspicuous they are often overlooked by collectors. This may explain the scattered distribution pattern. G. tenella grows in a variety of habitats from sandy sea beaches, moist and dry slopes, lake margins, wet mead- ows, rocky places, to alpine tundra, at altitudes from a few feet above sea level in the northern part of the continent to 3500 meters in the southern part; flowering from late July until late August or early September regardless of latitude; fruiting from late August until September. KEY TO THE SUBSPECIES a. Calyx lobes ovate to lanceolate; branches few; cauline leaves with rident es EE EE G. e? e tenella aa. One or two calyx lobes broadly ovate; branc shes numerous; stem eg peel ibilofi shortening of the lower internodes. Pribilof Islands, Alaska............... 13b. G. ag = pri o 13a. GENTIANELLA TENELLA (Rottb.) Börner, SSp. TENELLA Gentiana koeningii Gunner, Fl. Norv. 2:102. 1772, ex char 16 Gentiana glacialis Thom. e die Hist. Fl. Dauph. 2:532. 1787, ex char. (T.: Hall 652) Gentiana tetragona Roth, Tent. Fl. Germ. 2:290. 1789, ex char Hippion E Schmid in Roem. Archiv f. Bot. 1:21. 1796, ex char. entiana borealis Bunge, in Nouv. Mem. Soc ée? Mosc. 1:251. F. I0. f. 2. 1829, ex ic. edlen tenella (Rottb.) eg ex hs. Ind. Kew. 1:533. 1893 S Gentiana monantba A. Nels. in Bull. Torr. Bt- Club 31:244. ` 1904. (T.: Clements pedis 6561 Amarella monantba (A. Nels s.) Rydb. in Bull. Torr. mE E 33:148. 1906. Amarella oi (Rottb.) Cockerell, in Amer. Nat 2 Gentiana te ane Rottb. var. ooridehtelis Rouss. & Ee ih Nat. Canad. 79:77. 1952. eau 200! Gong Geer Rottb. var. monantha (A. Nels.) Rouss. & Raym. loc. cit. 77. 1952. Gentiana tenella Rottb. var. monantha £. alba Rouss. & Raym. loc. cit. 79. 1952. (T: Duran 310 Traill d, Sórensen 3274 (C, CAN), 3276 (C); Sa Lad, Sórensen pe 2 (C ` d jord, Jensen s.n. (C); Porsild 8545 (CAN). Location unknown: Ze Hartz s.n. (C), Jensen s.n. (C), Sabanse (?), Kruuse s.n. (C); Hurry Inlet, Kru TIN DISTRICT: Southampton Island, Brown 608 (DAO); Chesterfield Inlet, Savile 1 E 1391 (DAO); egens Point, Marsh 62 (UT, photo CAN). let NZ Coppermine, ts 242 (DAO); west side Bathurst Inlet, ee River, Kelsall ay McEwan 182 fid YUKON TERRITORY: St. Elias tas e de Bakewell $0, & 150 (GH, hee — 61 BEC: Poste de Povugnituk, rive ést de la baie d'Hudson, 1S 140, & 200 (JBM); Baie Kayak dans l'estuaire de la baje Payne, Rana 1496 (J ; 5! Porsild, in Nat. Mus. Can. Bull. 121:275. 1951. 1957] GILLETT—REVISION OF GENTIANELLA c 55 UNITED STATES: ALASKA: Nome, Seward Peninsula, Scamman 3910 (GH), Thornton 439 (US); Port Clarence, Teller, Porsild P Porsild 1488 (CAN, GH); Unalaklet, Porsild & Porsild s.n. (CAN, GH, US); Kotzebue Sound, Eschscholtz s.m. (US); Kotzebue, Lepage 25404 (CAN, DAO), Scamman 4064 (GH). ARIZONA: Coconino Co.: San Francisco Mts., Knowlton 131 (US). CALIFORNIA: Mono Co.: Bloody Cañon (Mose? ass), Congdon Herb. (GH); White Mts., wie? Meadow, Duran 2816 d GH, MO, NY, US). ADO: Clear Creek Co.: r Empire, Patterson iy (C, GH, MO, NY, US). Clear Be. "gb Cos.: Mount Flora Ps s. n. (GH, M O, US); Gray’s Peak, Patter- e S Lb ; 0: G Range, Brandegee s. n. (MO). San Miguel Co.: Trout Lake, Payson & Payson 4199 (G MO). Location unknown: Bear Creek, Purpus 700 (F); Mirror Lake, Clements & Clem- ents 456 (C, GH, MO, NY, RM, US). NEVADA: Esmeralda Co.: White Mts., Duran 3105 (F, GH, JBM, MO, E NEW MEXICO: Santa Fe Co.: vic. Santa Fe, Arséne & Benedict 16099 (F, US). UTAH: Duchesne Co.: Kus Emmons, Hermann 5201 ( ei Garfield Co.: Aquarius Plateau, Dixon s.m. (F), Ward 623 (GH, MO, US). Summit Co.: Lamotte € Ro 5973 (MO): dower Henry Fork Basin, Maguire et al. teen 14619 (WS). MING: Albany Co.: Medicine Bow Mts., Ownbey 176 (WS). Park Co.: Bear- Kee? Mis , Williams e Williams 3762 (GH, MO, NY). se Ae i Co.: Wind River Mts., Nuttall ER (GH); Yellowstone Park, Lone Star, Reynolds 74 (F). The degree of variation in this species is quite as high as in the other species of the genus. While I have not seen the type of Thomas’ Gentiana glacialis, I have seen a sheet from the Copenhagen collection labeled on the back in the lower left- hand corner, “Gentiana glacialis", and in the lower right-hand corner, “dedit Vil- r$" "This material is certainly authentic, and may even be an isotype. Villars, in his Flora, refers to Gmelin’s ‘Flora Sibirica’: "Elle paroit gravée dans Gmelin Tab. 51. B. vol. IV. sous le nom de gentiana pumila centaurii minoris folio flore biloso. Stelleri pag. 105." This plate compares very closely with the Copenhagen specimen with the exception of the corolla, which is pictured as having five lobes. This feature is not critical, however, as such variation in lobe number is common throughout the genus with the exception of the fringed gentians where four parts are the rule. On the other hand, the five lobes may have been an error by the artist, for in all other respects the illustration is G. Zenella. Nelson described Gentiana monantha partially because he thought that the true G. tenella Rottb. was an arctic plant with 5-merous flowers: “The true G. tenella Rottb. is an arctic plant with 5-merous flowers, the obtuse corolla lobes as long as its tube; the calyx only half as long as the corolla and with unequal sepals. The original description of G. tenella, however, states: “fauce 4fida, barbata”, and “Perianthium 4phyllum”. i A small population from St. Paul Island, Pribilof Islands, Alaska, consists o rather short plants with a dense cluster of ascending branches. Both the 4- and 5-merous condition occurs and one or two calyx lobes are more broadly ovate than usual. This population may be referred to: [Vor. 44 266 ANNALS OF THE MISSOURI BOTANICAL GARDEN 13b. GENTIANELLA TENELLA (Rottb.) Börner, ssp. pribilofii J. M. Gillett, sub- spec. nov. Plantae 3-5 (—8) cm. altae; folia caulinaria plerumque in rosula per contrac- tionem internodiorum inferiorum; ramis numerosis ascendentibus. Flores (4)- 5-meri; calycis lobis exterioribus tum 1 tum 2 latius ovatis. ALASKA: Pribilof Islands: Seal Island, Bryant s. n. (US) ; St. Paul Island, Elliott s. n. (GH,US), Kincaid s.n. a e ROND; St. Paul o Behring Sea, in 1891 . Macoun HOLOTYPE, GH, MO, Ns in Ce QD» rus in GC (CAN, US), in 1914 n. (US). ENUMERATION OF THE SPECIES Subgenus I. EUBLEPHIS 6. Gs age: (Richards.) J. M. 1 ee? package: G. Don a. ssp. d a. gs zt opinq b. ssp. nsis J. M. Gill b forma aeyanea J J. ses . Ssp. aleut am " dre (Th. Holm) J. M. Schlecht.) J. M. Gillett d. ss Ss (Porsild) J. M. 7. G. aurea (L.) H. Sm. ex Hy- Gi illet lander e. ssp. JD: (A. Nels.) J. M. 8. a quinquefolia (L.) Small Gillett ssp. quinquefolia £f SG superba (Greene) J. M E ssp. V LEM Gray) J. M. Gil Gillet E ssp. lanceolata (Benth.) J. M. 9 x he (Lemmon) J. M. Gillett Gillett h. e Reg holopetala (A. Gray) J. M Ser. IL. AMARELLAE 2. G. Kino PNE) G. Don E — (M. E. Jones) J. M. a. ssp b. sp Procera (Th. Hol M. 11. G. amarella (L.) B weer a. ssp. deir Eer )J.M. ca SCH “victorin (Fern. M. Gillet Pa b. ssp. ns (Michx.) J. M. Gil- d. ssp. macounii (Th. Holm) J. M. lett Gillett S C. ssp. eric (Griseb.) J. M. 34. G, Ge ee (Engelm.) J. M. Gillet Gillet : d. ssp. iiia J. M. Gillett 4. G. simplex (A. Gray) J. M. e. ssp. dene (A. Gray) J. M- Gillet Gillet Subgenus II. GENTIANELLA £. sap. hartwegii J. M. Giller 12. 6G. prem ze ) J. M. Gillett Subgenus III. COMASTOMA 13. G. tenella viera Bórner a. SSp. ened : ssp. pribilofii J. M. Gillett Sect. AMARELLA x I. ARCTOPHILAE E RS beuran (Engelm.) J. M. 1957] GILLETT—REVISION OF GENTIANELLA 267 INDEX TO NAMES a s recognized in this revision are in Roman type; new names and new combinations are in bold face type; synonym talics; names weve the rank of genus—section and series names—are in ALL CAPS. Aloitis — 235 var. E f. uniflorum ........ 228 divaricata SS 244 proceru 226 Moa 245 si Sea 232 mesocbora ` 245 Sieg 217 occidentalis |. 245 tonsum 228 parviflora Se 244 virgatum 226 quinqueflora 244 Chironia 202 quinquefolia 235 202 eo 248 Cicendia tenella 264 SE 253 Crawfurdia 197 NN 254 Crea” 210 amarelloides _ 244 gentianoides 210 anisosepala __ 253 ie ytum 202 arctophylla |. 237 iris AA we. — —— auriculata ___ 261 Ts = 210 cobrensis — 259 223 conferta ` — 253 Hess ores 253 cobelandi ` ` 254 Eyrytbalia 248 gurliae .— 5 254 Frasera 202 hartwegi ` 260 Gentiana 195 beterosepala 251 acuta 253 ete R 254 ssp. acuta 234 macounii ` 254 var. — 253 mexicana ___ 258 ssp. pleb 253 microcalyx _ 246 ssp. Mick fT TT ou 293 onantba __ 26 B stric 255 occidentalis 245 var. SE 253 blebeja var. bolmii 254 ` aggreg 253 b 237 ajanensis 258 Es 254 aleutica 241 s 251 $ ARELLA 234 Strictiflora ` — 254 +*+ Amarella — 248 iig 264 amarella ` — 249 ire 254 var. acuta 253 theiantha ` 254 f. michaux 254 var. lactea 254 var. strictiflora 253 var. livida 254 var. tenuis 253 tortuosa 248 amarelloides 244 m 237 ani: doni 253 care Se 235 vt ARCTOPIHLA .— em da wrigbtii 259 arctopbila -——— Anthopogon 209 densifore sA ella 248 ares ` ee 242 barbata 213 a var. borealis a. genuina .....— 242 barbellatus 230 GOD (ii e caccia mU brevifolium 223 axillaris — 253 detonsa _ 213 barbata —— 213 elegans 217 B Brown —— 226 incarnatum 223 Y ie E 213 macounii 228 B sim ume 228 barbelleta — —— — —————— brocerum var. tonsum ___— [Vor. 44 268 ANNALS OF THE MISSOURI BOTANICAL GARDEN Pru — — ou —— 264 Pseud-Amarella 242 brachypetala a L quinqueflora 244 Cun beirh. oue eni uu Sg 209 var. MACH 245 eise Gunn esas 214 B 244 CNN ES i aa ponis aou 243 ak: meroa ooa A var. emavellaides 244 ciliata Pall. cu uc S = ders escens 244 ciliata Moc. & Sens o 219 ccidentalis 245 OMM ae o 260 quinquefolia Oki 242 * COPLANTHE = 5. — 248 217 $t COMASTOMA — e sicheidioni 213 EE Ee A rurikiana 237 npa S T scopulorum 251 rvantesii wein c iir M serrata 215 LDOCUCCONMEUNAMIN . — — 2 fe B - 213 Sect. . CROSSOPETALUM — E 210 Y - 219 E Ge 5 } | var. - flop - 220 s rete E c e A E seti - 237 var. groenlandicum. EE Secher — 232 Y Spleen o UM icta M +) vat. tone oo 0 UR — 228 strictiflora 253 ENEE 251 superba — 219 ele BONG irc aru dI tenella — 262 var. brevicalycina uec xac FI var. monantha ee YAT gece E EE $ 17 a — 264 CM deed, es nom 248 i eier subg. PURBLERRIS — < ć — 210 Weg eegen - 253 fimbriata ———— eL Ae A tetragona cii AME gaspensis —————— —.—— 228 Ibermalis —— M subsp. Cenia —— ^ c 209 sa -- 228 ege vieta SE 264 tortuosa See T rc ur 71 vr (Gray) ‘Holm. _ merid EE rae ie = I E AE unalascbensis 0 AP eren E 960 ANANTHE — ae belleri —————À AS ventricosa vu PIE beterosehda — — 0 ey Victorini er E booth 1 x o. = 220 isliz mud vem. o 00 1 05 da wrigbtii iced ERE tomm S E ët 235 lanceolata — i n 220 Gentianella 208 ouni o a a c ATE — 234 macrantha — . ee 219 E AMARELLA 234 macrocalix el ue LLL ED Ser, AMARELLAR 4m WEE BEEN 249 ssp bartwegi ENORMI MEUSE. ssp. acuta ek f. pringlei vec Lc PEU ssp. hartwegii an microcahis — ae uo bit: — — 24 WonMHa . ^ sei : voce snag eens? 258 p MEUNEM — — — —— el. EE EE SE vs 230 SS MEM 1 o 1L Lew 239 Age mut 216 is wrightii See seb o ME Ar ARCTOPHILAR — $5. 8 occitent . — — 8 TUA cda `. NIIS EE EA auriculata Se Ru phele —— — SES SUN UIS barbellata 208, 230 pies —ÓM— HE campestris 208, xi m o s 2 Sg B densiflora i San emie TEEN silii: ect 19571 GILLETT— REVISION OF GENTIANELLA Subg. COMASTOMA — . . — var. occidentalis 269 ssp. quinquefolia 244 serraía _ 213 simplex . . 232 tenella ` 262 ssp. pribilofii . 266 ssp. tenella 264 tetranda 208 tortuosa 248 slizeni 248 Gentianella 210 alenia — 197 pion 261 auriculatum 261 longepedunculatum 264 pem 244 Ixanthu 197 Jaechkea ____ 197 Lih lanceolatus 220 Lomatogonium 197, 203 Sect. COMASTOMA .. . . —. _ 262 enellum 261 Megacodon ` 203 Opsantba 209 Pleurogyne 197 Pneumonanthe 202 Spi e 209 Swe 197, 203 MISSOURI BOTANICAL GARDEN ` pO T e EDGAR ANDERSON, SE Joen A. Seen SC Curator of Useful Plants ` Rommy. E. WOODSON, JR., = Volume XLIV Winther A Annals of the Missouri Botanical en | Morphological C Com idu in Heer islas Japulus. w with S Special Reference to the American Race . Edward I L. Davis. 2 1-294 Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the uri Botanical Garden and the Henry Shaw School of Botany of recs ës University in affiliation with the Missouri Botanical arden. Information The ANNALS OF THE Missourr BOTANICAL GARDEN appears four times during the calendar year: We Sch May, September, and November. Four numbers constitute a volume Beginning with Volume 45, 1958: Subscription Price... $12.00 per volume single Numbers . .— . — 3800 ‘cach of previous issues of the ANNALS OF THE MISSOURI BOTANICAL Contents GARDEN are listed in the Agricultural Index, published by the H. W. Wilson Company. Annals of the Missouri Botanical Garden Vol. 44 NOVEMBER, 1957 No. 4 MORPHOLOGICAL COMPLEXES IN HOPS (HUMULUS LUPULUS L.) WITH SPECIAL REFERENCE TO THE AMERICAN RACE'. EDWARD L. DAVIS? The genus Humulus cannot be treated taxonomically as can a plant known exclusively in the wild. Unlike many cultivated plants, hops are fairly easily established as escapes from cultivation, and the distinction between wild and cultivated hops is not easily drawn. One may be certain that the natural distri- bution of hops has been greatly modified by man. Yet hops cannot be studied only as cultivated plants if the origin and relations of the various cultivated varieties are to be understood. The record of hop varieties is so incomplete that the history of many of the most important cultivated varieties is not known with certainty. The relationships between these varieties, as well as the more funda- mental and difficult question of the origin of cultivated hops, cannot be determined without a close examination of wild-growing hops. It has therefore been necessary to attempt a simultaneous study of cultivated and wild forms, though in the following paper the emphasis will be given only to the American hops. A later publication will consider the cultivated varieties of the world in more detail. Hops have found many uses, in addition to that of flavoring beer, which have led to their cultivation over a wide geographical area. Sometimes the use has been such that hops have been cultivated in small gardens rather than in large fields. Many of the early medicinal uses were summarized by John Parkinson (1640) who noted that the young buds served “to open obstructions of liver and spleene, to clense the blood, to loosen the belly”, and in clensing the blood hops help to cure “all manners of scabes, itch, ringwormes and spreading sores." Var- ious parts of the plant were used: eds in powder taken in drink killeth the worm in the body; the juice of the leaves dropped into the ears clenseth the corrupt sores; a syrup, made of the juice and sugar cureth those that have the Henry Shaw School of Botany of raduate laboratory of the : partial fulfillment of the requirements for part of a thesis in pa ? University of Massachusetts, Amherst, Mass. (271) [Vor. 44 272 ANNALS OF THE MISSOURI BOTANICAL GARDEN yellow jaundice, easeth headache that cometh of heate, and temper the heate both of the liver and stomach." Hops have found additional medicinal use in poultices for swellings and tumors, for relieving pain of gout and rheumatism, and for calming nervous irritation (Simmonds, 1877). They were widely regarded as having a soporific effect, and pillows were stuffed wit em. This practice attracted attention when Dr. Willis, in 1787, prescribed the use of a hop pillow for George III (Johnson, 1867). Morgan (1952) suggested that the hop-screened front porches of rural homes in the northern part of the United States served to lull farm folks to sleep after a hard day’s work. Dr. Edgar Anderson recalls the practice in upper New York State and Michigan during the nineteenth century of planting hop vines outside the bedroom window. For these medicinal uses the quantity of hops required was not great, and sufficient was easily grown in the family garden. Such small scale, but widely practiced, cultivation has given man an unusually fine opportunity to alter dras- tically the natural distribution of this plant. According to De Candolle (1883), the hop is native to Europe and Asia, from England and Sweden as far south as the mountains of the Mediterranean basin, and in Asia as far as Damascus, the south of the Caspian Sea and eastern Siberia, and it is not native to India and China. Nuttall (1847) considered it certainly indigenous to western United States, but the following letter by Asa Gray? in answer to a request for American wild hops, does not give any suggestion that they are to be found in the East: There is hardly a doubt that all the hops raised in North America are originally from er i o plants introduced from Europe. If these will answer the purpose, it would be easy to have a tains only, an ng dioecious, they can only now and then be found in fruit. Ih seen the plant growing wild, but never in such a way that I could have gathered a pound of dried hops. There is morphological and cytological evidence, presented below, for the existence of an American race of hop, quite distinct, though in minor details, from the European hop. NOMENCLATURE The genus Humulus has been placed variously in the families Moraceae,* the Urticaceae, and the Cannabineae.® The question of the organization and size of families js beyond the scope of this treatment, and I am following Rendle’ (1952) in considering it a member of the Cannabinaceae, with the closely allied hemp. E = vie? Druggist 15:111. 1886. ngler, K., und Prantl, K. Nat. Pflanzenfam. IIi:96, 1888 s Bentham, G., and Hooker, J.D. Genera Plantarum 3.3 56-357. 1883 De Candolle, A. Prodr. 161:28-31. 1869. : : "Rendle, A.B. The classification of 3 : 1 5 F 756-58. 1952. 1957] DAVIS—-MORPHOLOGICAL COMPLEXES IN HOPS 273 Linnaeus, in the first edition of the ‘Species Plantarum"? listed a single species H. lupulus L. Over one hundred years later, A. De Candolle, 1869, in his treat- ment of the Cannbineae, included three species in the genus: H. lupulus L., H. cordifolius Miq., and H. japonicus Sieb. & Zucc. He did, however, question the validity of H. cordifolius Miq. H. lupulus and H. japonicus are very distinct species. The former is a peren- nial with numerous lupulin glands upon the bracts and bracteoles of the cones, and the lower surface of the leaves, and with a diploid chromosome number of 20 (Ono, 1955, Sinoto, 1929, Winge, 1929). H. japonicus is an annual, with lupulin Blands rare or absent, and a 2 N number of 17 in the male, 16 in the female (Kihara, 1929, Winge, 1929, Sinoto, 1929). H. cordifolius Miq. will be con- sidered on a following page. After 1753, the following names were used for the European hop in disregard of the prior Linnaean name: Lupulus scandens Lam. Fl. Fr. 2:217. 1778. L. communis Gaertn. Fruct. 1:358. £. 75. 1788. L. amarus Gilib. Exercit. 2:451. 1792. H. volubilis Salisb. Prod. 176. 1796. H. vulgaris Gilib. Hist. Pl. Europ. 2:343. 1798. H. aculeatus Nutt. in Jour. Acad. Philad. N. S. 1:182. 1847. Nuttall (1847) considered that the hop native to western United States was distinct from the European H. lupulus. He established H. americanus from a collection made by Dr. Gambel in New Mexico. His description and comments ollow: H. am : Leaves 3 to 5 lobed, the upper sometimes entire, inner divisions lanceolate- acuminate, ergeet along the a apex; wier of the cone ovate, acute, the lower ones acuminate. erican I have ventured, a ink, n sufficient g ds, to separate the c rom the ur hop, as it is in the uncultivated interior of the continent beyond the reach of inhabitants, our pl ust neces b igen I compared the present with the foreign plant with some attention, and I can in all cases readily distinguish th y their fol American plant, whatever be the other variations of the attenuated ints are denticulated nea he — In the European the summit of the leaf is abruptly toothed. In the native pes e iem flowers appear to be smaller, and the scales of the cone are lik e specimens, as in the European plant, the upper leaves som are simply Le and entire, bes in ail Genen the denticulations are smaller, and more This separation was not universally accepted, and Asa Gray retained but the single species H. lupulus L. in the revised edition of his ‘Manual’.® Still later, E attempts were made to establish an American segregate. Nelson and Cocke Proposed the variety H. lupulus L. var. neomexicanus. Again the type was Pa the Southwest (Beulah, New Mexico). The lateness of the date should not pass —— s Sp- Pi. SE 1028. 1, Man. Bot. North. U.S. p.400. 1857. Proc. Biol. pe Wash. 16:45. 1903. TVOoL. 44 274 ANNALS OF THE MISSOURI BOTANICAL GARDEN unnoticed. In the fifty years since Nuttall's collection there was much oppor- tunity for cultivated forms to become established as escapes. There is absolutely nothing in their descriptions which would make it possible to differentiate between the American variety and H. lupulus L. of Europe, except the geographical location. The acuminate fruiting bract used to set off the variety occurs widely among European hops, both wild and cultivated. Rydberg, in the ‘Flora of Colorado’, lists the vd Toe H. lupulus L. neo- mexicanus Cockerell for the Rocky Mountains. He says: "The native hops of the Rocky Mountain region has deeper divided leaves SS more sharply acuminate bracts than the cultivated variety. It grows along streams from Wyo. to Utah, N. M. and Ariz." This represents the first suggestion that the American type may be identified by the divisions of the leaf. By 1917, Rydberg had raised the variety to the species level)? The lobing of the leaves in the inflorescence, which he used in the key, is highly variable and not of any taxonomic significance. : At best, the descriptions and keys of the above authors permit the classification of some herbarium specimens into one or another species or variety. They do not establish a type found predominately in North America and separated from all European forms. The confusion is summarized by Fernald!?: “The native plant sometimes called H. americanus Nutt., H. lupulus var. neomexicanus Nels. * Cockerell and H. neomexicanus (Nels. & Cockerell) Rydb.; its characters evasive. Extensive comparisons of European and American hops bring out, above all, their great similarity. It is possible to differentiate between the European and American type on the basis of a complex of characters, but only when cone and leaf material is carefully collected. Consequently, I consider that for the present it is advisable to retain the name H. lupulus L. for the wild perennial hop of Europe and America. The annual H. japonicus Sieb. & Zucc.14 is a very distinct species which, be- cause of the small number of lupulin glands, has never been used in the making of beer. Native to Asia, it is now widely scattered in North America as an adventive weed. Merrill!” proposed using the name H. scandens (Lour.) Merr. for H. japonicus on the following grounds: Loureiro’s feed description ds Së scandens Lour. 1790], based pron. a nearly glabrous form for which he saw Staminate specimens, applies D the species cur- rently kr n as H. jap We & Zucc. No other known Kwan tung ici in any famil remotely conforms to the ae crea indicated by Loureiro. Loureiro’s description is rons vague, and I am following Fernald in retain- ing H. japonicus Sieb. & Zucc M FI. Colo. p. 100. 1906 12 Fl. Rocky Mts. & Adj. Mains; P. rs e Gray's Man. Bot. 56. 1950. 2 mj Ls Fam. Nat. 2:89 (Abh. x. Akad. Münch. mds -nat. Kl. 42:213. 1846. imm SK en A commentary on Loureiro’s "Flora Cochinchinensis." Trans: Amer. Phil. Soc- 1957] DAVIS—MORPHOLOGICAL COMPLEXES IN HOPS 275 Maximowicz'® makes reference to his reduction of Humulus cordifolius Miq. to a variety, noting that De Candolle had erred in placing it near H. japonicus in his key on the grounds that lupulin glands were absent on the bracts and bracteoles. On the contrary, Maximowicz emphasizes the presence of such glands by conclud- ing that it is as suitable for use in brewing as is the European hop. Miquel’s description!" is very incomplete, being based only upon female material, and Maximowicz did not clarify the situation. Hance!? provided the first description of the male plant, but the description contains nothing which could be used to separate it from H. lupulus. Furthermore, an examination of specimens from Maximowicz's collection shows it to be identical with the European hop. Con- sequently the names H. cordifolius Miq. and H. lupulus L. var. cordifolius Maxim. are placed in synonymy with H. lupulus L. MORPHOLOGY Until recently only brief accounts of hop morphology had been published despite very extensive studies conducted upon certain parts of the plant. Within the last two years, however, this situation has radically changed. Ehara (1955) has made a comparative. study of H. japonicus Sieb. & Zucc. and H. lupulus L., including anatomical observations as well as descriptive comparisons starting from the seedling stage. Hamaguchi (1955) published an even more extensive work on H. lupulus L, including developmental analysis of the vegetative and flowering shoot. -As a consequence of these publications, a detailed account of the mor- phology of the hop plant will not be presented here, though a description of the inflorescence of the female plant is required for the discussion which follows. Unless otherwise indicated, all observations apply to H. lupulus L. The small pistillate flowers, each consisting of a cup-like perianth and a single Pistil with two elongate stigmas, are borne in the axils of bracts and bracteoles upon a condensed axis, forming the cone. Multicellular glands, containing lupulin, are ound in great number upon the bracts and bracteoles. The cone itself is a greatly condensed inflorescence, readily differentiable into nodes and internodes. At each node are normally found four flowers. This node is really a complex structure best understood from fig. 1 (modified after Hamaguchi, 1955, p. 94) in which three are diagrammed. The internodes a! and a? are reduced to the point where they are barely visible, thus condensing two nodes of these secondary axes back upon each node of the primary axis. The nodal cluster has been interpreted as cymose, with the median flowers abortive (fig. 2). When all bracts and bracteoles are removed, the cone appears in side view as shown in fig. 3A. Each nodal cluster then appears to be subtended by a single bract scar. When turned through 90°, however, it is seen that two — 16 In Franchet & Savatier's Enum. Pl. Jap. 2:489. 1879. 1 Ann, Mus, Bot. Lugd. Bat. 2:133. 1865. 18 Jour. Bot. (Brit. & For.) 20:293-294. 1882. [Vor. 44 276 ANNALS OF THE MISSOURI BOTANICAL GARDEN bracteole primary axis lower 7! internodes 2nd axes A3 bracts A Fig. 1. Diagram of a hop cone showing Fig. 2. A cluster of flowers at a three nodes. Internodes of secondary axis single node of cone. Abortive flowers are much enlarged. indicated by X. scars occur at the base of each cluster of four flowers (fig. 3B). Rendle!? has interpreted these as stipules of a former bract. T ough the two scars are not usually inserted at the same level, one appearing slightly above and overlapping the other, the infrequent cone with well-developed foliage leaves confirms Rendle's interpretation. In such cases the petiole is found directly between the two scars, Which are at the same level (fig. 3C). The slight displacement found in the normal cone must be attributed to a secondary effect of growth. Moreover, these stipular bracts (which are so widely known in the hop literature simply as bracts that they will be so called henceforth) show a vein pattern similar to that of the stipules throughout the vegetative parts of the plant, while being very different from that of the bracteoles. MATERIAL Much of the analysis of hops throughout this study has been based upon cones. They contain many diagnostic characters and are readily obtainable from many parts of the world. of the cones used for analyses were small samples taken ch are submitted to brewers for their evaluation of the 19 Classif. Fl. Pl. 2:56-58. 1952. 1957] DAVIS—-MORPHOLOGICAL COMPLEXES IN HOPS 227 Bracteole scar Bract scar cone axis with bracts and bracteoles removed: A, resting on flat Fig. 5. Hop - y surface; B and C, turned on edge; C, abnormal cone bearing foliage leaf. Some knowledge of field practices is imperative for an understanding of the field samples. Ideally, every hop field of the same variety belongs to the same clone. Vegetative reproduction, by cuttings, is the only method used commer- cially. Each spring, or for the sake of economy every several years, soil is dug back from the hop hill and all but a few of the young shoots are cut off. The cut shoots are then used to replant old hills, or for new plantings. It is a standard Procedure for farmers to sell excess hop roots to neighboring farmers, or at the Present time, in the large organizations of growers, to ship Toots from state to state. Only rarely, when new fields are being established, will such additions be kept separate. More generally new plantings are intermixed with old. This, by itself, would not offer complications if each variety remained true to type. However, not only have some new varieties been attributed to bud sports, but in- numerable worthless rogues occur each year. The superintendent of one farm in California reported he was removing almost 20 per cent of the hills because they [Vor. 44 278 ANNALS OF THE MISSOURI BOTANICAL GARDEN were off type. This was after many years without roguing, but several farmers have suggested that about 5 per cent of the plants each year are rogues. To my knowledge, no research has been done upon the source of such rogues. They may represent vegetative mutations, seedlings, or carelessly introduced contaminants. The tale told by an Oregon farmer (though it may be an unusual case and somewhat exaggerated) illustrates how contaminants may be introduced during replanting fields. He had purchased roots which he believed to be of a single variety of hop. After planting them on his own farm two-thirds of them grew into willow trees! The sequence of events which led to this can easily be followed. In many areas, the spring root cutting is not done under the surveillance of the superintendent, but the field is let out to a cutter who receives as pay the root cuttings he makes. The temptation to add to the yield by collecting roots of escaped hops from the roadsides is occasionally too great. To the contamination from such weed hops had been added the frequently present willow whose shoots are readily confused with those of hops. Besides the over-all sampling problem which results from the field heterogeneity is the heterogeneity of the individual hop hill. Almost without exception more than one root is planted per hill; most frequently the number is three. In making the collections of leaves and cones I attempted always to take both from the same vine. Usually the effort was successful. However, since four (or more) vines are trained to each wire, and these vines may originate from different root stocks, it was necessary to trace the cone-bearing side branches back to the main vine at the level (5 or 6 feet above ground) at which leaves had been selected. This always introduced opportunity for error. VARIETAL IDENTIFICATION Before turning to the morphological characters which have been used here, some previous efforts to describe varieties on structural characteristics will be considered. In the description of new varieties, it has been a standard practice to make use 0 cone shape, size, and coloration. Such descriptions are of obvious use to the grower, at least in so far as they are not too generalized. Unfortunately, exactness is not easily attained. Moreover, cone shape and size are greatly influenced by pollina- tion. For instance, a variety such as LATE cLusTERs, which produces long cones with prominent dark green streaks (the bracts) alternating with light green streaks (the bracteoles) when unpollinated, regularly has nearly spherical cones, without streaks, after pollination. Although this type of varietal description is useful to growers familiar with these changes, it is not suitable for describing pressed mate- rial, nor is it of great value to the botanist or geneticist. It is only when more definitive characters are used that it will be possible to trace the inheritance patterns in Beyond studying the varietal descriptions, many efforts have been made to study other morphological features of hops. Only a few need be mentioned here Wormald (1915) undertook an extensive examination of male hops. Leaf color, 1957] : DAVIS—MORPHOLOGICAL COMPLEXES IN HOPS 279 texture, lobing, dimension of lobes, gland number and distribution were scored. Salmon and Wormald (1921) studied time of flowering, color of stem, relative number of glands on the leaves, and on the cone; size, shape, aroma, and color of the cone axis. Schmidt (1918) investigated the possibility of characterizing clones by the number of marginal teeth on the leaves. This investigation, carried out over a number of years, provided valuable information on the change in leaf shape which takes place as a plant becomes established. More recently, Meneret et al. (1954) have reported on the use of the length-width ratios of bracts as a means for varietal identification. In the present study six characters have been chosen for scoring. They repre- sent various modifications of the characters previously used. Even more important is the semi-graphic method of expressing the results (Anderson, 1949). By this means it is possible to identify a hop sample by plotting it out and comparing it with previous plottings of known varieties. The advantage is not only in its use- ess in identifying varieties, but also in that genetical relationships are suggested when suitable botanical characters have been selected for scoring. Identification of hop varieties is not the primary concern of this paper. Rather attention is being focused upon the morphological complexes within H. lupulus. Evidence suggests that these complexes, which are based entirely upon structural features, correspond to certain chemical and perhaps cytological groupings within the species. If addi- tional research confirms this, such morphological features as are used here will have an increased value in hop breeding. At the same time it will provide an interesting example of the use of chemical, morphological, and cytological evidence for the establishment of taxa below the level of species, where morphological evidence is not adequate by itself. The following six characters have been scored and the results expressed in pictorialized scatter diagrams: On the leaf: (1) dentation number, (2) pubescence type. On the cone axis: (3) condensation; (4) angularity; (5) bract scar type. On the bract: (6) number of main veins extending into the terminal third. LEAF: (1) Dentation.—The leaf shape, and consequently the dentation number, is exceedingly variable upon each hop plant. To attempt to describe plants upon the basis of a few randomly selected leaves would be impossible; merely increasing the number of leaves sampled has not proved very satisfactory. Leaves found at the lowermost nodes are distinctly different from those at higher nodes and in the inflorescence. The leaves of the secondary and tertiary axes differ from those of the primary. Despite this great variation, truly homologous leaves may - ob- tained by making the selection from a main vine (one of the vines trained sé = the hill) of a mature plant. In this way, the phenotypic variability can be re Be to measurable proportions. The leaves in all field collections have been taken ji the main vine, at a height of about five or six feet. The dentation count was ma on the median lobe of the leaf in the following manner: From the number of all [Vor. 44 280 ANNALS OF THE MISSOURI BOTANICAL GARDEN S ES G) EID, a Type ! mam SN Ko "matt Shs AEF Fa . Under surface of hop leaf drawn at margin, showing three different patterns. Lower right, showing method of measuring angularity. pubescence dentations counted, was subtracted the number which were not provided with a distinct vein visible to the unaided eye, the result giving the “major dentations”. The number of minor dentations was also recorded in each case, but was so variable within every variety that it did not contribute to their identification and has been eliminated. In addition, the number of dentations within 5 cm. from a point On the edge of the lobe to the last dentation at the lobe apex was counted. Since Schmidt’s work (1918) had shown that dentation number steadily in- 1957] DAVIS— MORPHOLOGICAL COMPLEXES IN HOPS 281 creased during the first three years that a plant is becoming established, collections were made only from plants four years or more old, unless otherwise indicated. (2) Pubescence-—The number of hairs per unit area upon the under side of the leaf undergoes tremendous variation, and is not by itself a satisfactory char- acter. Of greater importance than the number is the pattern of their distribution. Upon all hop leaves, hairs are found on large veins. Type 1 was made up of leaves with hairs on major veins only; type 2 with hairs upon these and upon secondary veins as well; type 3 was characterized by the presence of hairs upon all veins, down to those of the smallest, and between veins on the vein islands. These types are shown in fig. 4. In most cases, type 3 possessed the most hairs in an absolute sense, but when the leaves approached the glabrous condition, hairs were not confined to large veins, as in the nearly glabrous leaves of type 1. With but few exceptions known clones remain constant as to type. Cone: (3) Condensation of tbe axis.—Using a dissecting scope, the length to the hearest mm. was measured between five bract scars. These were taken at the base of the cone, or at least never including the two uppermost nodes. It is to be expected that the greatest disturbing influence upon this and the following character is pollination. Since pollination is known to produce profound changes in the cone axis, only unpollinated (seedless) cones are compared, unless otherwise indicated. Cone axis of three different varieties: SAAZ, Czechoslovakia; AM, LATE CLUSTERS; ENG, FUGGLES. Fig. 5. [Vor. 44 282 ANNALS OF THE MISSOURI BOTANICAL GARDEN (4) Angularity of tbe axis.—The angularity of the axis was measured with the use of an eye-piece micrometer and disecting scope. A line ab (fig. 4) was drawn with the micrometer parallel to the axis, and then lines ac and bc were measured; from the right-angle triangle so constructed, angle D was obtained. This angle was in turn doubled, thus giving an approximation of the angle naturally formed. (5) Bract scar.—The scar left by removal of the bract, when seen in side view, is classified as either elongated (stretching nearly between nodes of the axis, fig. 5 SAAZ) or abbreviated (confined to the area immediately below the subtended fruits, fig. 5 ENG.). The type of bract scar served to make a major division between varieties. The scar is definitely independent of the degree of condensation or angularity of the axis as is shown by the retention of form without regard to the changes produced by pollination. BRACT: (6) Vein number.—Brewmasters have long realized that different varieties have different-shaped bracts. Such differences may be of considerable value though the variation within a cone, from base to tip, is great. Meneret et al. (1954) have been determining length-width ratios over a period of years. Though their report is optimistic, they concluded that this measure was “not completely definitive and should be considered along with other characteristics in identifying the type.” Here, the same problem was approached in another fashion: the number of major GET, Fig. 6. Three Principal bract types. veins extending into the terminal third of 4 bract was recorded. This is itself correlated with shape, a low count being found on the long acuminate bract, a gh one on the obtuse bract. For the present it appears to be more satisfactory than shape, at least for establishing differences between complexes. However, 2 When scored in this manner, The range varies with the vari fall into groups 7—8, 9, 10 or the vein number for any variety is not constant. ty, but even with this variation, the populations more. Fig. 6 shows a bract of each type. 1957] DAVIS— MORPHOLOGICAL COMPLEXES IN HOPS 283 VARIETAL STABILITY Among the most discussed, but least exactingly studied, characteristics of hops, is their reputed ability to adapt to each new environment and assume the qualities attributed to that particular region. Such views are still widely found among growers. One of the popularly held examples of this "adaptation" is based on the effect of pollination upon cone shape and color as mentioned above. LATE CLUSTERS produces long cones with dark green and light green stripes when grown in Cali- fornia, Washington, and Idaho. Those in Oregon have more nearly spherical cones without striping. The essential difference between these types is not in the area in which they are grown, but in the presence or absence of seeds. The Oregon area hops, which are always seeded, have a characteristic cone, as do the uniformly seedless California forms. In the selection of the foregoing tools of analysis, a strong emphasis was placed upon the stability within a clone as a measure of reliability. That fact at once Prejudices their use in the evaluation of the uniformity or variability as it exists within each variety over time and space. A comparison of LATE CLUSTERS from California, Washington, and Idaho, for the year 1955 (figs. 7 and 8) shows how little the difference is, in these morphological characteristics at least, within that geographical range. The variation is not significantly greater than the individual Plant variation shown in fig. 9, in which the leaf and cone measurements have been arbitrarily combined. Another comparison is based upon leaf scores only. Fig. 10 shows leaf lobe number and dentation scores for LATE CLUSTERS. For the populations studied, the leaf lobe number was greater in Idaho (median 7) than in California or Wash- ington (both with median 5). The dentation number is also greater in Idaho (42) than California (38) or Washington (36). That this may be attributed to larger leaves in the well-irrigated fields of Idaho is indicated by the dentation count per 5 cm. of leaf margins, both California and Idaho having a median of 12 and Washington a median of 11. The constancy in the morphological characters measured here in different geo- &raphical areas speaks well for varietal consistency. The number of morphological characters which may undergo some change of course is not known, but they are certainly not of enough magnitude to warrant a fear of varietal instability. It must be recognized that genetically based characters may be such that they will remain virtually unchanged over the area within which the plants may grow, or they may be capable of a fluctuating expression under changing conditions. In the latter case, they are no less genetic. [VoL. 44 284 ANNALS OF THE MISSOURI BOTANICAL GARDEN EXPLANATION OF SYMBOLS UsED IN SCATTER DIAGRAMS (Fics. 7-9, 12-14) Number of main veins at tip of lower sterile bract: 7- 8 QO e.a Pubescence: Hairs on main vein only Oo Hairs on main and minor veins d Hairs on and between veins D Number of teeth on median lobe of leaf: 20-20 O 27 -33 2 Bract scar: Elongated Abbreviated O k Sk «€ ae Pa cree Origine des Plantes Cultivées. Paris Ehara, K. (1955). Comparative morphological studies on the ue Kyushu Univ. Fac. Agr. Jour. 10: Hamaguchi, T. (1955). Studies in hop. II. Growth and differentiation of the hop plant. Bull. Brew. Sci. 2:67—120. Hance, H. F. (1882). Spicilegia florae sinensis. Jour. Bot. (Brit. & For.) 20:293—294. Johnson, C. P. Mead it The V Plants = ven Britain. pp. 231—234. Lon 92 d Kihara, H mulus japonicus. Jap. Jour. "es 4:55—635. Meneret, G., M . t, and J. Giolet A954). Form of the er among cultivated hops. Ann. d'Amélior. de E 1954: 209-246. (Rev. in Am. Brewer, Feb. 1956). Miquel, F. A. (1865). Prolusio florae japoni Ann. M t. SEN Bat. 2:p. 133. Morgan, Shirley (1952). Hops fora ere Herb Grower 6:118—122. Ween ., and T. D. A. Cockerell (1903). Three new plants from New Mexico. Proc. Biol. Soc. Washington, 16:p. 48. Nuttall, T. (1847). Descriptions of plants ree by William Gambel M.D. in the Rocky Moun- tains and Upper California. Jour. Acad. Nat. Phila, n. s. 1:181—18 Ono, T. (1955). Studies in hop. I. baa x common hop and its e En Bull. Brewing Sci. 2:1—120. Parkinson, J. (1640). "Theatrum Botanicum. p. 176. London. Rendle, A. B. (1952). The classification of flowering plants Met, F. E (1956). Recent advances in the chemistry of hop ERC and their selene is ewing technology. Tech. Proc. . Amer pue Salmon, E E. e and H. Wormald (1921). A study of the variation in das of the wild hop. Jour. et. 11:241-267. e £ (1918). Can different clones be characterized by the number of marginal teeth on the ? Compt. Rend. Lab. Carlsberg 14:1-23. ies P. L. (1877). Hops, London. Neat X. X. ane Chromosome Cytologia 1:109-191. Wings, O. (1929). On the chrom Humulus. Hereditas 12:53—65. Wormald, H. (1915). males a male hop. Jour. pis Sci. 7:175—196. studies in some dioecious plants with special reference to the allo- ANDROCALYMMA, A NEW GENUS OF THE TRIBE CASSIEAE (CAESALPINIACEAE) JOHN D. DWYER* Androcalymma Dwyer, gen. nov. Arbores. Folia imparipinnata foliolis petiolulatis alternis vel oppositis tenuiter coriaceis costa supra immersa. Flores cymulis in paniculam terminalem dispositi. Sepala 5 imbricata; petala 5 imbricata unguiculata intus carinata. Stamina 4 libera aequalia partim circum pistillum disposita antheris biporosis rectis vel deflexis subversatilibus connectivo gracillimo. Ovarium subsessile complanatum ovulis paucis. Trees. Leaves imparipinnate, the leaflets alternate or opposite. Flowers in cymules in a terminal panicle. Stamens 4, equal, the anthers erect or deflexed, subversatile, the connective slender. Pistil subsessile with few ovules. Type Species: Androcalymma glabrifolium Dwyer. Androcalymma glabrifolium Dwyer, sp. nov. Arbores circ. 30 m. altae. Ramuli ultime valde lenticellati teretes glabri. Petioli 3-6 cm. longi, 0.1-0.5 cm. lati, plani glabri; stipulae non visae. Folia aequilateralia ovato-elliptica, 4.5—10 cm. longa, 2.5-5.5 cm. lata, plus minusve apice acuta basi obtusa tenuiter coriacea supra nitidula laevia glabra costa infra im- mersa prominente basi circ. 1.5 mm. lata, venis principalibus secundariis 8—12 Prominulis arcuatis venulis laxis reticulatis marginibus parum callosis; petioluli 5-7 mm. longi graciles glabri. Flores in cymulis paucifloris dispositi, rhachidibus evidenter quam superioribus foliis brevioribus gracilibus basi ad 1.5 mm. latis basi articulatis; bracteae bracteolae caducae cicatricibus prominentibus. Sepala 5 aestivatione valde imbricata inaequalia oblonga vel ovato-oblonga circ. 3 mm. longa 1.8—2 mm. lata apice obtusa extus pubescentia ciliis mediis saepe densius dispositis intus glabra carnosa marginibus sparse ciliolatis; petala 5 subaequalia aestivatione valde imbricata oblonga 4—6.5 mm. longa 1.7-2 mm. lata apice obtusa basi unguiculata, unguibus ad 0.5 mm. longis, extus pubescentia intus glabra et evidenter carinata; stamina 4 filamentis subcrassis saepe subclavatis rectis, 2-3 mm. longis circ. 0.25 mm. in medio latis glabris mox deciduis; antherae ovato- subrotundae circ. 0.6 mm. longae vix apice biporosae subrectae vel deflexae sub- versatiles, connectivo circ. 0.2 mm. longo paullum thecis obscurato, 4-locellatae at 2 loculis; ovarium vix stipitatum vel subsessile complanatum oblongum circ. 2 mm. longum leviter omnino pubescens 2—3-ovulatum stylo brevissimo. Fructus non visi. —————— "St. Louis University and Missouri Botanical Garden, St. Louis, Mo. Issued January 29, 1958. (295) [Vor. 44 296 ANNALS OF THE MISSOURI BOTANICAL GARDEN Type Collection: Krukoff 9005, Municipality São Paulo de Olivença; basin of Creek Belém, terra firma, high land, Amazonas, Brazil (HOLOTYPE MO, isotype NY The fancied resemblance of the anthers when deflexed to a drawn cowl sug- gested the name, Androcalymma.! The pinnately compound leaves and the poro- cidally dehiscent anthers readily place it in the tribe Cassieae. Several floral characters point to its relationship with Dicorynia Benth. and Martiusia Benth. of the same tribe: the imbricate sepals and petals borne on a fleshy receptacle, the lack of an obvious hypanthium, the few stamens only partially surrounding the ovary, and the very shortly stipitate ovary. The flowers disposed in numerous few- flowered cymules and the imparipinnate leaves with alternate or opposite leaflets likewise suggest this relationship. Both Marfiusia and Dicorynia are well distributed in the Amazon Basin as well as in the Guianas. Several features of the androecium as well as the very short style of the pistil serve to distinguish Androcalymma from its allies. The filaments of the new genus are much more elongate than those of Martiusia and obviously less crassate than those of Dicorynia. The latter genus has only two stamens. The anthers of Martiusia are subulate and acuminate. Dicorynia, with unequal filaments, has larger anthers which are 4- or 8-locellate. The presence of 8 locelli in one or in both of the anthers distinguishes Dicorynia from all other genera of the Caesalpiniaceae. Worthy of special note is the slender attenuate filament tip attached to the anther of Androcalymma. This usually lies well hidden between the locules which are free at the base. It is difficult to trace the filament tip to the point of inser- tion on the connective, the base of the latter being well above the base of the locules. The subversatile condition of the anthers is apparently found elsewhere in the tribe Cassieae only in occasional species of Cassia. Krukoff, the collector of the type material, notes that the tree is 100 feet high, and that the trunk measures 18 inches in diameter. I wish to acknowledge the help of several specialists who made notations on the herbarium sheets and carried on correspondence concerning the new genus. These include: A. Ducke, R. Schery, J. Monachino, N. Y. Sandwith, and R. Cowan. Special thanks are due the last three. l'apyjp (man); kappa (cowl). 19571 DWYER—ANDROCALYMMA Fig. 1. Androcalymma glabrifolium Dwyer 297 A REVISION OF THE GENUS STYLOSANTHES* ROBERT H. MOHLENBROCK** History OF THE GENUS The genus Stylosanthes was established in 1788 by O. Swartz! with two species, S. procumbens (= S. hamata (L.) Taub.) and S. viscosa. Previous to the founding of the genus, most present-day species of Stylosanthes had been referred to other genera. Linnaeus? in 1753 had described Trifolium biflorum (= S. biflora (L.) BSP.) and in 17593 Hedysarum hamatum B (= S. viscosa Sw.) ; in 1775 Aublet* had named Trifolium guinanense (= S. guyanensis (Aubl.) Sw.). Preceding the advent of the binomial system, a few polynomials had been ap- plied to modern species of Stylosanthes. Apparently the first mention of a species now known to belong to the genus was in 1696 when Sir Hans Sloane? listed Anonis non spinosa, minor, glabra, procumbens, flore lutea as occurring in Jamaica. This is S. hamata (L.) Taub. In 1737 Burmann? recorded Trifolium procumbens zeylanicum hirsutum (= S. fruticosa (Retz.) Mohlenbrock). Willdenow, in the fourth edition of ‘Species Plantarum',* listed five species of Stylosanthes. G. Don? in 1832 listed thirteen species. Of these, however, only seven have proved to be distinct, and three of these were recorded without usage of the earliest epithet applicable. The first study of Stylosanthes was a synopsis by Vogel® in 1838. He divided the genus into two sections, EUSTYLOSANTHES and STYPOSANTHES, based on the absence or presence, respectively, of an axis rudiment (an aborted secondary axis associated with each flower). An unfortunate situation has arisen, however, since Vogel designated S. procumbens Sw. (— S. hamata (L.) Taub.) of $ sTYPOSAN- THES to be the type species. According to present nomenclatorial rules, the section or subgenus including the type species must repeat the generic name, hence Vogel's $ srYPOsANTHES must be changed to § STYLOSANTHES; his $ EUSTYLOSANTHES here is renamed § ASTYPOSANTHES. In his section including species without axis rudiments, Vogel recorded eight species, three of them new, while in the other section, seven species are listed including four new ones. All seven of Vogel’s new ` Swartz, O. Prod. Veg. Ind. Occ. 108. 1788. „Linnaeus, C. Sp. Pl. 773. 1753 x Linnaeus, C. Syst. Nat. 10:1170. 1759. k Aublet, F. Pl. Guian. 776. 1775. oane, H. M: ann, ? Vogel, in Linnaea 12: 68. * An investigation carried out in the graduate laboratory of the Henry Shaw School of Botany of i University, and submitted as a thesis in partial fulfillment of the requirements for the degree of Doctor of Philosophy. ** Southern Illinois University, Carbondale. Issued January 29, 1958. (299) [Vor. 44 300 ANNALS OF THE MISSOURI BOTANICAL GARDEN species are recognized in this study as valid, with only one of the entire fifteen listed by him placed in synonymy. Between 1838 and 1890, six species were ascribed to Stylosanthes by various authors, but all apply to species previously named. A more detailed revision was provided by Taubert!? in 1890, including morphol- ogy, generic relationships, and geographical distribution in addition to keys and de- scriptions for the species. He retained Vogel's division of the genus into two sections and recognized 22 species, four of which he himself proposed. Fifteen of these are maintained in this study, although the specific concept has changed somewhat. It is apparent that Taubert was handicapped by the limited number of specimens available for his study. Little was added to the genus until 1919 when Hassler!! enumerated the species of Paraguay and added fourteen subspecies, varieties, and forms to S. guyanensis and S. montevidensis. Blake described thirteen species between 1920 and 1926, Harms!? added one in 1923, Macbride!? one in 1943, and Standley and Williams! one in 1950. In 1945, Herter!? split the genus into two genera, Stylosanthes and Astyposantbes, equivalent to Vogel's sections and in which rank I prefer to treat them. In the present study, 25 species are recognized, three described for the first time. GENERIC RELATIONSHIPS Stylosantbes belongs to the tribe Hedysareae of the papilionaceous Leguminosae characterized by monadelphous stamens and a lomentum. The alternation of sub- basifixed and versatile anthers and the absence of stipels suggest its relationship with three other genera, Zornia, Arachis, and Chapmannia. The quartet is grouped in the subtribe Stylosanthinae. : Stylosantbes shows a close affinity with Zornia because of the large bracts sub- tending each flower and the often spicate inflorescence. The leaflets of Zornia, however, are paripinnate, the ovules are numerous, and bracteoles are lacking. Arachis is closely allied to Stylosanthes since both genera have similar stipules and 2-3 ovules per carpel, but they differ in that bracteoles are lacking and the loment matures underground in Arachis. The monotypic Chapmannia from Florida bears imparipinnate leaves, but differs in that it possesses minute stipules, numerous ovules per carpel, and lacks bracteoles. MorPHOLOGY Species of Stylosanthes usually are herbaceous or suffruticose perennials, some attaining a height of 1.5 meters. Some species are prostrate and spreading (S. viscosa, some specimens of S. bamata), others upright and ascending. The roots usually are thick and mostly straight, with few secondary roots. M Standl. & L. Wms. in Ceiba 1:145. 1950 15 Hert. in Rev. Sudamer. Bot. 7:209. 1943, 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 301 Fig. 1. Loments of Stylosanthes Smal Pees exicana, gtt 9 ae 5 a, Robertson Mi scabra, Martius 1124; 3 7. s s Se Pick 6; 8 s Pap abe ata, S. vis d Wo 9. S. sympodialis, Weber- bauer 5936; 1 S. fruticosa, Mearns 1956; 11. S. subsericea, Standley I2110; 12. S. macrocarpa, D Priegis 6721 Géi E). All figures X 374. [Vor. 44 302 ANNALS OF THE MISSOURI BOTANICAL GARDEN The stems are generally diffusely branched and woody at the base, rarely simple (S. bracteata). 'The indument of the stems is diverse. The leaves of Stylosanthes are trifoliolate with the terminal leaflet slightly larger than the lateral ones. They are borne at the base of the cleft of the bi- dentate stipules. There are no stipels. The petioles range from 1.5 to 9 mm. long. The leaflets vary considerably interspecifically in shape and size. They are borne on rhachises which seldom exceed 2 mm. in length. S. capitata and S. guyanensis may have leaflets up to 4 cm. long and nearly 2 cm. broad. In most species the leaflets are elliptic to lanceolate, obtuse to acute, occasionally mucronate, and taper- ing to the base. Some species have black-punctate leaflets. The upper surface of most leaflets is glabrous, although in some species it may be appressed white-hairy. The lower surface may be glabrous, puberulent, or bristly, at least along the costa. The nerves of the leaflets usually are prominent on both surfaces. The stipules are amplexicaul, pubescent, and bidentate at the apex. They are 3- to 11-nerved, with usually one to three nerves extending into each tooth. The stipular teeth usually are shorter than the sheath. In this study, the teeth are measured from their tip to the base of the cleft while the sheath is measured from the cleft to the base of the sheath. The inflorescence is a terminal or occasionally an axillary spike or head. In some species the inflorescence may be elongate to 20 cm. and interrupted. The number of flowers varies from one to forty or more. The flowers are sessile in the axil of the bract and adnate to it. Each flower is surrounded by a series of bracts and bracteoles. The outermost bract is similar to the stipules in being comprised of a 3- to many-nerved sheath with a bidentate apex from which arise one or three reduced leaflets which are identical to the cauline leaflets except in size. In some species (S. bracteata, S. capitata, S. angustifolia), the leaflet is reduced and the sheath is often much larger than the stipular sheaths. Within the bract is a somewhat smaller outer bracteole which usually is 3-nerved and ciliate along its margin. Inside this may be found a densely ciliate axis rudiment (§ srvLosaNTHES) or this may be lacking ($ ASTYPOSANTHES). The nature of the axis rudiment has been the subject of much conjecture, but studies show it probably to be the vestige of a secondary floral axis. In some speci- mens of S. bracteata, the axis rudiment is adnate to the outer bracteole for about one-third of its length. Essentially the same situation occurs in Vicia monantbos, according to Svenson.!ó Within the axis rudiment are one ($ ASTYPOSANTHES) or two (§ STYLOSANTHES) inner bracteoles which often are deeply cleft and usually long-ciliate near the apex. The calyx is comprised of an elongated tube expanded into five lobes at the apex. The lobes usually are unequal, with one larger than the others or rarely with two larger. Usually the four smaller lobes are connate. The largest lobe usually 18 Sven. Am. Jour. Bot. 33:394—498. 1946. 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 303 e e. Fig. 2. Loments of Stylosanthes - S. nervosa, Weberbauer on (TYPE); 2. S. bamata, Ricksecker 47; 3. S. hamata, Blake 7792; 4. s. hamata, Leonard 8797; 5. S. sericeiceps, Jahn 678 ove 6. S. biflora, Fisher s.n.; e guyanensis, Martius 1123; 8. S. angus. "ni GE 9004; E ew Herter 89069; S. cayennensis, Broadway 972 ud 11. osa, Miers nd acrosoma, Hass 606; 13. S. bumilis, Orcutt pios 14. S. figuer 8 dione $97 Ger prä S. bumilis, E 25; 16. S. bibbocampoides, Herter 85562; 17. S. montevidensis, Venturi 9957. All figures X 34. [Vor. 44 304 ANNALS OF THE MISSOURI BOTANICAL GARDEN is acute, the others obtuse or subacute. They may be puberulent or glabrous. The tube generally is glabrous. The entire calyx may be 4-15 mm. long, the tube being 3-12 mm. long. The corolla is typically papilionaceous. The petals are usually yellow or yellow- orange and purple-striate. They are inserted near the apex of the calyx tube. The standard sometimes attains a length of 10 mm. and a width of 8 mm. It is invari- ably suborbiculate. The wings are somewhat shorter and narrower than the standard, usually clawed and auriculate, sometimes spurred within at the base. The keel petals are incurved and subrostrate and are slightly smaller than the wings. The stamens are monadelphous and united into a tube which becomes split, at least partially, with age. There are ten stamens, five with small versatile anthers alternating with five bearing larger sub-basifixed anthers. The staminal tube arches into and is enclosed by the keel petals. The single carpel has a minute terminal stigma, an elongate, usually curved style which is persistent in fruit, and a subsessile ovary which contains two or rarely three campylotropous ovules borne on marginal placentae. Only one or two of the ovules develop. e fruit of Stylosanthes is a bi-articulate loment terminated by the persistent style. The upper articulation is almost always fertile, the lower abortive or fertile. The lower articulation usually is densely pilose or glabrescent (S. mexicana, S. leio- carpa, S. cayennensis, S. erecta). The upper is glabrous or puberulent to sericeous or minutely tuberculate (S. guyanensis). The fruit usually is prominently nerved. The persistent style crowns the upper articulation as a beak which is minute or up to 8 mm. long, straight, uncinate, or circinate. The beak is glabrous or puberulent to densely villous. The relative lengths of the upper articulation and the beak are diagnostic. The upper articulation is measured from its base to a ventral protuber- ance formed where the upper articulation narrows into the beak. The beak, whether erect, declined, or coiled, is measured from the ventral protuberance to its maxi- mum height as a simple vertical projection of the upper articulation. This measure- ment, therefore, usually does not represent the gross length of the beak but merely the distance that it projects above the upper articulation. The seeds are black, compressed, ovoid, estrophiolate. In the latter part of the nineteenth century, a statement crept into the litera- warns attributing two kinds of flowers to Stylosanthes biflora: apetalous ones from which the seeds are set exclusively, and petalous ones which are sterile. ‘This ap- parently has been copied by numerous authors through the years and still persists . in some of our more recent manuals. No evidence of apetalous flowers occurring in S. biflora was found in this study. GEOGRAPHY OF THE GENUS The genus Stylosanthes is indigenous to savannas and didührenas in the eastern — States, Central America, the Antilles, South America to northern Argen- tina, the Galapagos Islands, central and southern Africa, Madagascar, and southern 19571 MOHLENBROCK—REVISION OF STYLOSANTHES 305 India and Ceylon. In addition, S. bumilis is adventive in Malaysia and Australia. All eleven species of $ ASTYPOSANTHES occur in the Western Hemisphere. Some are wide-ranging and show considerable variability, although most are well fined. Five of the seven widely distributed species of Stylosanthes belong in this section. S. biflora is rather abundant over the eastern United States. The very widespread S. guyanensis with its two subspecies occurs from Central America to southern Brazil and Bolivia. S. viscosa occurs from Baja California through Central America and northern South America and the Caribbean basin. S. montevidensis and its close local relatives, S. bippocam poides and S. macrosoma, center in southern Brazil, Bolivia, and Paraguay. S. angustifolia occurs rather frequently in northern Brazil and the Guianas. The other three species of this section, S. cayennensis, S. figueroae, and S. leiocarpa, are local, the first two being restricted to extreme north- ern South America, the last to Uruguay, Paraguay, and southeastern Brazil. The fourteen species of $ STYLOSANTHES, on the other hand, are exceedingly difficult to distinguish and are mostly restricted. Only S. hamata of the Antilles and S. fruticosa of southern Africa, India, and Ceylon are widespread. S. scabra is rather abundant from Colombia to Minas Geraes, Brazil. S. macrocarpa and S. subsericea are confined to Central America and S. calcicola to Florida, the Antilles, and parts of Central America. From Colombia and Venezuela to Peru and Ecua- dor are found S. mexicana (also in Mexico and Bolivia), S. nervosa, S. tuberculata (also in the Antilles), S. sericeiceps, and S. sympodialis (also in the Galapagos Islands). S. bracteata and S. capitata center in Brazil. S. erecta occurs in east- central Africa. UsEs Certain species of Stylosanthes (S. fruticosa, in particular) are used as fodder crops. In Australia, S. guyanensis ssp. guyanensis has been planted in an effort to check soil erosion. Nodules of nitrogen-fixing bacteria are present on the roots of most of the species. Data for some specimens indicate Stylosanthes to be a good cover crop for coffee. Several folk uses are attributed to certain members of the genus. Taubert noted that a few species have been used as diuretics. Hutchinson and Dalziel’? attribute many uses for S. erecta by native Africans: ambia an infusion of the plant is taken internally for colds, etc. In N. Nigeria the r pe is bl to arrow wounds as an antidote to their poison. The plant is a common ingredient in superstitious herbal practice, Es the root eus Zeit i i inj sharp weapons ausa maga 3 rm or epared asa wash isa EECH of injury by P : DT a root t in ions. or leaf used as a daily bath for an infant will enable him to walk very early. in eg smoked like tobacco, and also made into a decoction, is a charm against injury by a blow by a cudgel, and will cause the weapon to E —— 17 Appendix to Fl. W. Trop. Afr. p. 262. 1937. [Vor. 44 306 ANNALS OF THE MISSOURI BOTANICAL GARDEN ACKNOWLEDGMENTS The author is grateful to Dr. Robert E. Woodson, Jr., who supervised the study and who meticulously criticized the manuscript; to Dr. Edgar Anderson and the staff of the Missouri Botanical Garden for the courtesies extended to him; and to his many friends and colleagues who utilized the keys and descriptions and added numerous invaluable comments. ? SYSTEMATIC TREATMENT STYLOSANTHES Sw. Prod. Veg. Ind. Occ. 108. 1788. (T: S. hamata (L.) BSP.) Astyposantbes Hert. in Rev. Sudamer. Bot. 7:209. 1943. Herbs, rarely subshrubs. Leaves pinnately 3-foliolate; stipules amplexicaul, di- vided at the apex into two teeth, adnate to the base of the petioles; stipels none. Inflorescence spicate, terminal or axillary, 1- to several-flowered; flowers yellow or orange with purple stripes; calyx 5-lobed and with a long tube; corolla 5-merous, papilionaceous; stamens 10, the filaments united into a tube which splits with age, with 5 versatile anthers alternating with 5 sub-basifixed anthers; fruit a 1- to 2-articulated loment; seeds ovoid, light brown to black, smooth, lustrous. The herbaria in which the cited specimens are deposited and their abbreviations are as follows: Chicago Natural History Museum (F), Gray Herbarium (GH), Missouri Botanical Garden (MO), New York Botanical Garden (NY), and the United States National Herbarium (US). KEY TO THE SPECIES Sect. I. srvLosaNTHEs. Each flower, or at least the lower flowers, subtended by an axis rudiment; inner bracteoles 2 (1 in S. sericeiceps). A. Bracts 13-25 mm. broad, with 13-23 conspicuous and usually broad, colored nerves. B. Fertile articulations of the loment often 2, the upper glabrous, the beak less than one- the upper articulation, glabrous or with a very few short stiff hairs; nerves of the bracteal sheath about 15; stems usually with scattered bristles. Brazil, Ve nen RE a L 1. S. capitata . Fertile articulations of the loment 1, pilose, the beak only slightly shorter than the upper articulation, pilose; nerves of the bracteal sheath about 21; stems pilosulous to villou. Brazil Paraguay... 2. S. bracteata AA. Bracts at most 10 mm. broa i C. Beak of the loment straight (rarely with marginal ci ia); teeth of the uppe Florida, Centr: erica, Mexico, Bahamas, Cuba... 3. S. calcicola CC. Beak of the loment uncinate or coiled: leaflete aw 1. B Ki et s in S. e Africa and S. bamata of the West Indies) ; sheath of the upper stipules mostly longer th d Së à SE SE - Loment completely glabrous or ome pubescence on the beak only or occasion- ally on the nerves in xicana ; leaflets sparsely but conspicuously bristly-eiliate; loment : 5 leallet Picuously bristly-ciliate; lomen often green, with both articulations usually fertile; inflorescence obovoid. Mexico, a ED D'Zeegegasggeeegeggee 62996. 4 EE. Stems glabrous or puberulent above; with only one articulation usually S ^ N E 2. D =] Mr ON eles S. mexicana leaflets without bristly cilia; loment brown, Ata ee feles inflorescence often narrow. e x DD. Loment pubescent on the body and usually on che Lack F. Stem and bracts bearing tuberculate-based hairs lower surface of leaflets usually vill 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 307 of the loment me Jm the upper articulation (sometimes equaling it in S. macrocarpa and S. nervosa). G. Beak of the lom Piscis one-third the length of the upper articulation. H. Bracts shortly ET nflorescence usually nearly as broad as long; beak of t nt short- I. Leaflets Eus Sie ed obtuse; stem bearing short dark setae, "n whole aspect of the plant dingy brown; fertile articulation usually pubescent throughout. Baal. Reunlar: Venezuela, Colombia, E S. scabra II. Leaflets not punctate beneath, acute; stem setose or puberulent to md gy brown; fectila ‘articalitions often 2, pilosulous only o the ribs. Bahamas, Cuba, ombia, Venezuela, Peru................... S. tuberculata HH. Bracts villous or pe long Seege late bristles; I et du often ibe aim longer Amn broa d; eak di he loment usually rufous-pilos : T pejs Kias br Toltotese ence 2—3 s e m. broad, the beak bro? pilose. cuador, Peru, Galapag Islands S; ts os ym podialis JJ. Bracts with ub ulate bristles; inflorescence about as broad as long; loment 1.5-2.5 . broad, the beak puberulent. Africa, Genie India. 9. S. fruticosa GG. X nearly equaling to exceeding the uppe r articulatio. . Beak of the loment strongly coiled; bracteal haat co h and often the stem sericeous and with interspersed buberculate bristles. Honduras, Mexico......... . subsericea KK. Beak of the loment curved to s AY | uncinate; bracteal sheath and stem with tuberculate bristles but not serice per articulation and beak c Wists bou t 8 mm. long, the beak about equaling the u upper werd orie plants to 0. 2. m. tall. Mexico.....11. S. macrocarpa k € D beak some- tall. M. Fertile articulations mostly 2; bracteal sheath idi ge aring long tuberculate-based brotes E iplc. obtuse to a usually See Africa, Ce Ceylon India MM. Fertile articulation usually 1; bracteal sheath short-hispid t ciliate; leaflets ually oblan Sg acute to acuminate, g ges occasionally hispid beneath. Venezuela, Peru, Bolivia, Argentina. 12. S. nervosa FF. Stem and bracts pilose, villous, or nearly glabrous, lacking — rere lower surface of leaflets gna or Ke or glabrous, never w culate bristles; beak of th t equaling or exceeding the upper Gees (except S. sericeice ps). N. Beak of the loment equaling or exceeding the upper articulation; pubescence of the bracteal sheath whitish; stem mx not ico at t throughout. Florida, S. isn o densely Bahamas, Cuba, Central America, Colombia, Venezucla........................—.—— . S. bamata NN. Beak of the loment about oneal as sedie = = upper articulation; Sege - ie sadi sheath tan or rufous; stem often pubescent throughout. pubecsent throughout ps xr above, 15-25 mm. lon Wees mg tan-p V 14. S. sl EE OO. rr pubescent Ven ve, X omg res leaflets minutely pubescent above, mm. long; bracteal sheaths rufous-pilose. Peru, Ecuador, Ser ém . sympodialis med Sect. II. AsrvPosaNTHES. Flowers not subtended by an axis rudiment; inner bracteole 1. A. Beak of the 1 pos nimiis. at most about one-fifth as long as the upper articulation, the S fertile artic ulati B. Fertile Ster ——— hairy throughout or rarely glabrous, terete, 2.5—5.0 mm ong. Eas nited S: 15. s. biflora BB. d one sc asa tuberculate near the apex or rarely with a few scattered —A rs, flattened, 1.5—3.0 mm. long. Bahamas, Central Americ vipera 6. S. guyanensis AA. Beak of eg Lon: from one-fourth as long to exceeding the upper articulation, Ca fertile articulations 1—2. [Vor. 44 308 ANNALS OF THE MISSOURI BOTANICAL GARDEN C. Leaflets 0.5—2.0 mm. broad; inflorescence very narrow and elongate; beak of the loment strongly uncinate, often 2—3 times as long as the pubescent upper articulation. Brazil, Guianas 17. S angustifolia CC. Leaflets 2-6 mm. broad; inflorescence usually capituliform, globose to ovoid or obovoid; beak of the loment various but not strongly uncinate. D. Loment glabrous (occasionally pubescent only on the beak). E. Loment with 2 fertile articulations, the beak straight or uncinate; bracteal leaflets stalked. F. Beak straight or nearly so, about equaling the upper articulation, together about 5—6 mm. long; bracts, leaflets, and stems densely short-bristly. Brazil, Uruguay, Paraguay 18. S. leiocarpa FF. Beak uncinate, one-third as long as the upper articulation, together 3.0-3.5 mm. long; bracts and sometimes the leaflets and the stems with scattered long bristles, j often glabrous. French Guiana 19. S. cayennensis EE. Loment with 1 fertile articulation, the beak strongly uncinate to coiled; bracteal leaflets sessile or subsessile. G. Loment conspicuously reticulate-nerved, about as broad as long ( or $-flowered. U i cluding the beak) ; inflorescence 4- to ex ruguay, Argentina........... 20. S. hippocampoides GG. Loment obscurely nerved, about twice as long as broad (excluding the beak); inflorescence 2- to 4-flowered. Paraguay. 21. S. macrosoma b DD. Loment pubescent (if nearly glabrous, the stem viscid). H. Beak shorter than or nearly equaling the upper articulation; stem often viscid. I. Stems with viscid hairs or short-hispid; leaflets usually punctate beneath; fertile articulations 1-2. J. Fertile articulations usually 2, the upper decidedly widest above the middle, the beak coiled, a uppe » about one-third the length of the r articulation. — Bahamas, Cuba, Central America, Mexico, South America 22. S. viscosa JJ. Fertile articulation usually 1, broadest at or below the middle, the beak nearly straight or slightly uncinate, one-half to nearly equaling the upper articulation. Colombia 25. Ik figueroae II. Stems without viscid hairs; leaflets not punctate; fertile articulation 1. Bra il, : Paraguay, Uruguay, Bolivia, Argentina, Colombia 24. S. montevidensis HH. Beak nearly 2—4 times longer than the upper articulation; stems not id. "s Central America, Mexico, Bahamas, Cuba, Colombia, Venezuela, Brazil....... 25. S. bumilis Section I. SrvLosaNTHES Section STYPOSANTHES Vog. in Linnaea 12:68. 1838. (T: S. hamata (L.) Taub.) Each flower, or at least the lower, subtended by an axis rudiment; inner bracte- o 2. Styposantbes bamata (L.) Taub., as type of the genus, automatically becomes the type of $ sTYLOSANTHES. 1, STYLOSANTHES CAPITATA Vog. in Linnaea 12:70. 1838, ex char. capituliform, about two- many-flowered, on peduncles usually 5-7 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 309 cm. long, occasionally 2—3 cm. long; bracts with a single very reduced leaflet; sheath often purplish, 8-12 mm. broad, conspicuously 11- to 17-nerved, copiously soft-pubescent; outer bracteole 1, to 3 mm. long, ciliate; axis rudiment 5-7 mm. long in fruit, very long-ciliate; inner bracteoles 2, 2.0-2.5 mm. long and much narrower than the outer, densely ciliate at the apex. Calyx tube 4-6 mm. long, the mostly acute and sparsely ciliate lobes about 2.5 mm. long. Standard obovate, 5-7 mm. long; wings obovate, 4-5 mm. long, auriculate at the base; keel petals 3-4 mm. long, falcate and auriculate. Loment to 2.5 mm. broad, reticulate- nerved; both articulations usually fertile (either sometimes abortive), the upper about 3.5 mm. long and glabrous, the lower somewhat shorter and glabrous or sparsely pubescent; beak uncinate, about 1 mm. long, glabrous or with very few short stiff hairs on the inner face. Taubert in his diagnosis of S. capitata states that the legume is uni-articulate, the lower joint being abortive. However, specimens of this species may bear two fertile articulations. The original description of S. capitata which contains no mention of the fruit apparently is based on an incomplete specimen, for Vogel comments: “Vidi specimen unum incompletum”. All species of Stylosanthes except S. capitata, S. bracteata, and S. angustifolia possess 1—3 leaflets between the teeth of the bracteal sheaths. In these three, a small laminal extension of the midvein of the sheath usually is all that is present. Specimens from Venezuela and the Piauí state of Brazil are more coarse and have the general vegetative appearance of S. scabra. The nature of the inflores- cence and loment clearly distinguishes S. capitata from S. scabra. Grows in fields, forests, or waste ground and is known only from Brazil and Venezuela. It grows at altitudes of about 1000 meters. Vogel does not specifi- cally cite any specimens but merely states that S. capitata has been collected in Brazil by Sellow between Victoria and Bahia (fig. 3). BRAZIL: MARANHAO: Rio Tocantins, Pires & Black 1994 (NY). MINAS GERAES: Belo Horizonte, de Oliveira s. n. (US), Williams & Assis 5646, 5863 (GH), 5834 (GH, MO, NY); Morro das Pedras, Williams & Assis 6475 (GH); Pampulha, Williams 5644 (NY). praví: Oeiras, Gardner 2093 (GH, NY); exact locality unknown, Lützelburg II2 (NY). VENEZUELA: ANZOATEQUI: near Mapira, Lasser 775 (US). BOLÍVAR: Ciudad Bolívar, Bailey 1434 (GH, NY, US). 2. STYLOSANTHES BRACTEATA Vog. in Linnaea 12:70. 1838. (T: Sellow 4734!) Stems herbaceous, very slender, erect, simple or occasionally branched, to 2 dm. tall, pilosulous to villous throughout. Leaflets lanceolate to eliptic, acute to obtuse and mucronulate at the apex, to 25 mm. long and 4 mm. broad, usually much narrower, pilose on both surfaces, with 5—7 pairs of veins prominent beneath; petioles 1-2 mm. long, pilosulous, the rhachis about 0.5 mm. long; sheath of the stipules 5-8 mm. long, pilosulous or villous as the stem, 17- to 21-nerved, the teeth lance-subulate, 4.5—7.0 mm. long. Spikes capitate, 8- to several-flowered, to 30 mm. long; bracts unifoliolate, pilose and ciliate, 10-15 mm. wide, the sheaths [Vor. 44 310 ANNALS OF THE MISSOURI BOTANICAL GARDEN d EE @& S. CAPITATA . BRACTEATA ` + CALCICOLA i . MEXICANA ec SCALE e $00 1600 1500 2000 wies LJ 1000 2000 T" Sree rene Fig. 3 8.0-9.5 mm. long and with 15-21 prominent purplish nerves, the teeth 2.0-3.5 mm. long, the sessile leaflet to 5.5 mm. long; outer bracteole 1, 3.0—3.5 mm. long, bifid and ciliate near the apex; axis rudiment 7-9 mm. long in fruit, very densely pilose; inner bracteoles 2, 2-3 mm. long, ciliate. Calyx tube 8.0—8.5 mm. long, the lobes 2.0-2.5 mm. long and ciliate. Standard obovate, 4.5—6.0 mm. long; wings 3.5—4.5 mm. long, auriculate at the base; keel petals 3-5 mm. long, auricu- late. Loment to 2 mm. wide, reticulate-nerved; lower articulation abortive, upper articulation 3-4 mm. long, densely pilose; beak strongly uncinate, 3.0—3.5 mm. long, pilose. The presence of the copiously and softly hairy broad b kes S. bracteata readily distinguishable from other i a qus species of the genus. This species is the most slender Stylosanthes and, when in fruit, possesses the lon pmi seg dimit: it IPIE Ires the axis rudiment is adnate to the outer bracteole for about one- third of its length. 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 311 In dry regions in Paraguay and southwestern Brazil at altitudes around 800 meters; may also occur in northeastern Argentina. Burkart! comments that specimens referable to S. bracteata collected by Spegazzini (2194 and 2195) may be from Argentina, although the labels signify only “Norte de la provincia de Corrientes" (fig. 3). BRAZIL: MATO GROSSO: Capão Bonito, Archer t$ Gebrt 105 (US). MINAS GERAES: Ituiutabá, Macedo 479 (US), 1241 (MO); Uberaba, Regnell s.n. (US). P A: be- tween Lagos and Desiro, Dusén 15684 (GH); Tibagu, Reiss s. n. (F); Villa Velha, Ponta Grossa, Hoebne 23369 (GH, ). WITHOUT PRECISE LOCALITY: Rio Vardo, Riedel s. n. (NY); Sellow 4734 (GH, F). ARAGUAY: Estancia Primera, Jörgensen 4877 (US, MO); between Río Apa and Río Aquidaban-mi, Fiebrig 5208 (GH, US); Tapiracuay River, Hassler 4363 (GH, NY); exact locality unknown, Hassler 8510 (NY). 3. STYLOSANTHES CALCICOLA Small, Man. Southeast. Fl. U. S. 730. 1932. (T: Small, Mosier, & Small 6537!) Stems erect, to 0.5 m. tall, branched from near the base, minutely pubescent along one side, or sometimes glabrous, rarely pubescent throughout. Leaflets lan- ceolate to ovate, acute or acuminate at the apex, to 15 mm. long, glabrous or ciliate on margins near the base, with 3—5 pairs of conspicuous veins; petioles 2-4 mm. long, glabrous or puberulent, the rhachis 0.2-1.0 mm. long; sheath of the stipules glabrous or puberulent, with about 7 conspicuous nerves extending into the usu- ally somewhat longer teeth. Spikes narrowly oblong or ovoid, 0.5-1.5 cm. long, crowded, 2- to several-flowered; bracts unifoliolate, the sheath equaling to twice as long as the teeth, averaging 4 mm. wide, ciliate and often bristly on the back, 5- to 7-nerved; outer bracteole 1, about 3 mm. long, ciliate at the apex; axis rudi- ment 6.0—7.5 mm. long in fruit, sparsely pilose; inner bracteoles 2, 2.0-2.5 mm. long, ciliate at the apex. Calyx tube 3-4 mm. long, about 1 mm. longer than the acute lobes. Standard obovate, 4.5—6.0 mm. long; wings 4—5 mm. long, auriculate; keel petals falcate, 3.5—4.5 mm. long. Loment 1.5-2.0 mm. broad, conspicuously nerved; both articulations usually fertile, the upper 2.5—3.0 mm. long, densely and shortly white-hairy or sometimes glabrate, the lower 2.0-2.5 mm. long, puberulent or sericeous; beak straight or only slightly curved, 1.5-2.5 mm. long, from half as long to equaling the upper articulation, with short white hairs or glabrous. Stylosantbes calcicola is distinguished from all other species of the section by its straight beak and its general lack of pubescence. Small described S. calcicola in 1932 from a collection made in Dade County, Florida, in 1915. All the Florida specimens were collected in pinelands in ham- moc In addition to Florida, the species is indigenous to Mexico (Yucatan), Guatemala, and Cuba. Apparently the first collection of this species was made from Yucatan by Gaumer (908) in 1895 or 1896. Although there are minor differences between the Florida and Yucatan specimens, the characters of the loment 18 Darwiniana 3:327. 1939. [Vor. 44 312 ANNALS OF THE MISSOURI BOTANICAL GARDEN and the general lack of pubescence throughout the entire plant seem to indicate that these populations belong to the same species. The specimens from Florida are mostly low, much-branched plants (tall and unbranched in Cornman 2132) with densely pubescent loments while those from Yucatan and Cuba are considerably taller and generally possess pubescent loments. The leaflets of the Florida specimens are shorter. Stylosantbes calcicola resembles S. bamata with respect to leaflet morphology, habit, and the densely pubescent loments. It is an apparent calciphile with a tolerance for rather warm, arid climates. UNITED STATES: FLORIDA: Dade Co.: pinelands near Murden Hammock, Small, Mo- sier & Small 6438 (MO); pinelands near Ross-Costello Hammock, Small, Mosier & Small 6537 (US). Monroe Co.: Big Pine Key, Cornman 2132 (MO). EE HUEHUETENANGO: between Nentón and Las Palmas, Steyermark 51654 MEXICO: YUCATAN: Chichancanab, Gaumer 2044 (GH, US, MO, F); Izamal, Gaumer 992 (GH, US, F); San Anselmo, Gaumer 1955 (GH, NY, MO, F), Gaumer 1956 (GH, US, MO, F), Gaumer 908 (MO, US, F); Progreso, Flores 1 F). BAHAMAS: NEW PROVIDENCE ISLAND: West Bay, Degener 18984 (GH, NY). CuBA: CAMAGUEY: Ganado, Cayo Sabinal, Shafer 863 (NY, F). 4. STYLOSANTHES MEXICANA Taub. in Verh. Bot. Brand. 32:21. 1890. CT: Schaffner 579!) see bangii Taub. ex Rusby, in Mem. Torr. Bot. Club 4:206. 1895. (T: Bang 930: Stem ascending to spreading, subligneous near the base, to 3 dm. tall, usually much branched from near the base, densely covered with appressed white hairs, often with scattered bristles at least on the lower half of the stem. Leaflets broadly lanceolate to obovate, obtuse to subacute, glabrous or nearly so above, usually sparsely bristly-ciliate, occasionally with some tuberculate-based bristles on the lower surface, with 3—4 pairs of conspicuous veins; terminal leaflet to 16 mm. long and 4 mm. broad, the lateral leaflets somewhat smaller; petioles 3-6 mm. long, shortly white-hairy to nearly glabrous, the rhachis to 2 mm. long; sheath of the stipules equaling or exceeding the teeth, somewhat veiny, pubescent like the stem. Spikes dense, more or less oblongoid, to 15 mm. long, several-flowered; outer bracts trifoliolate, the inner unifoliolate, the sheath with short white hairs or sometimes with tuberculate-based bristles, 7- to 9-nerved, 3.5—5.0 mm. long, the teeth 0.5- 1.0 mm. shorter than the sheath; outer bracteole 1, oblong-lanceolate, ciliate, bifid, 3.5—4.5 mm. long; axis rudiment to 6 mm. long in fruit, long white-ciliate; inner bracteoles 2, ciliate, 2-3 mm. long. Calyx tube to 6 mm. long, the obtuse lobes about 2-3 mm. long, usually glabrous, sometimes ciliate. Standard suborbiculate, 4.5-6.0 mm. long; wings nearly equaling the standard in length, subauriculate at the base; keel petals falcate, 3.0—4.5 mm. long. Loment 6-7 mm. long, 1.5-3-5 mm. broad, reticulate- nerved; both articulations usually fertile, the upper 3—5 mm. long, glabrous or with some appressed pilosity on the nerves when immature, the 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 313 lower somewhat shorter, glabrescent, occasionally abortive; beak recurved-uncinate or sometimes with a half-coil, 2.5—3.0 mm. long, usually about equaling the upper articulation in length, glabrous or sparsely hairy. Stylosantbes mexicana shows much variation in leaflet shape and pubescence. While most specimens possess some tuberculate hairs in addition to simple pili, Leavenwortb 135 from Nuevo León, and Purpus 4925 from San Luis Potosí, Mex- ico, lack these bristles. The leaflets of Purpus 4925 are obtuse instead of acute. On valley floors dissected by arroyos and alluvial fans, in wooded gorges, and savannas at altitudes from 800 to 2500 meters (fig. 3). At first glance, the geographic distribution of S. mexicana seems to be very peculiar. This species is known from a few states in central Mexico and from Caracas, Venezuela, and the chacos of Bolivia, widely disjunct areas. This situa- tion, however, occurs in several species of flowering plants, striking examples being Larrea divaricata (Zygophyllaceae), Atamisquea emarginata (Capparidaceae), and Koeberlinia spinosa (Koeberliniaceae). This last species is known from the Mex- ican states of Tamaulípas, Nuevo León, and San Luis Potosí and again from a small area in the western chacos of Bolivia. In searching for an explanation of this seemingly unusual distribution, John- ston!? observes that both the areas in Mexico and Bolivia are extensive regions which are characterized by low atmospheric humidity and rainfall usually less than twenty inches annually, and in which the climatic and edaphic conditions are rather similar. The plants which occur in these areas belong principally to genera of a xerophitic nature. Johnston points to the presence in the deserts of North Amer- ica of floristic elements which apparently are a part of a flora now well represented in South America. He believes we are dealing "with a very old American desert flora formerly shared by both continents. In South America it is now relatively well preserved but in North America it lingers in a few recognizable remnants." It is obvious that at the present time the extensive unbroken belt of wet trop- ical forest between the Mexican and Bolivian stations presents "effective barriers to the exchange of elements." Johnston postulates that "during dry, warm epochs, an exchange might have been effected along relatively arid coastal strips similar to... those now present in western Central America, Ecuador, and Peru.” The species from Bolivia, known as S. bangii, is indistinguishable from S. mexicana and is considered synonymous. Mrxico: Rep Palm e elef SE (F). mmateo: near El Salto, Pringle 11969 (GH F). NUEVO L Canyon, near Monterrey, Mueller & Mueller 325 (E; MNA iere? e "GH "P ips de gege, Leavenwortb 135 ( OAXACA i Mitin » vos 1 (US). QUERÉTARO: uan del Río jo Painter & Rose en 28 uis PoTOsi: Mina de San ita, Purpus 4928 (GH, NY, MO, F); (US). i San Miguelito, “Seba ner ‘B00 DR s > 579 (US, F). TAMAULÍPAs: Buena Vista Haci- enda, ette frr doo o de la aulípeca, Bartlett 10608 (F), 10612 (GH, F); near Aiken get? wem s Northe raft 793 (GH, NY, MO); vicinity of La Victoria, CNS r 490 (GH, ; act locality unknown, Viereck 3oo (US). BoLivA: COCHABAMBA: Bolivian ti Bang 963 (GH, NY, MO, US, F). 19 Jour. Arn. Arb. 21:356—363. 1940. [IVor. 44 314 ANNALS OF THE MISSOURI BOTANICAL GARDEN VENEZUELA: DISTRITO FEDERAL: Caracas, Pittier 9679 (GH, US, NY); lower Catiza, near Caracas, Pittier 7319 (GH, US). 5. STYLOSANTHES ERECTA Beauv. Fl. Owar. 2:28. 1807, ex char. Stylosantbes guineensis Schum. in Schum. & Thonn. Beskr. Guin. Pl. 357. 1828, ex char. Ononis coriifolia Reichb. ex Guil., Perr. & Rich. Pl. Senegam. 204. 1830, nom. nud. in emer guineensis G. Don, Gen. Syst. 2:281. 1832, ex char. Stylosantbes erecta var. guineensis Vog. in Linnaea 12:68. 1838, ex char. Stylosantbes erecta var. acuminata Welw. ex Baker, in Oliver, Fl. Trop. Afr. 2:156. 1871, ex char. (T: Welwitsch 2127.) Stems suffrutescent, suberect, 0.3—1.5 m. tall, much branched, puberulent above, glabrous below, rarely with bristles. Leaflets oblanceolate, mucronulate, narrowed at both ends, glabrous, often punctate beneath, usually with 2—5 pairs of con- spicuous veins; terminal leaflet to 25 mm. long and 5.5 mm. broad, the lateral ones to 13 mm. long and 4.5 mm. broad; petioles 4-6 mm. long, glabrous or puberulent, the rhachis about 1.5 mm. long; sheath of the stipules 4.0—7.5 mm. long, glabrous, puberulent, ciliate, or bristly, 5- to many-nerved, the teeth 3.5—5.0 mm. long. Spikes oblongoid to narrowly elongate and sometimes interrupted, 4- to 12-flow- ered; bracts usually unifoliolate, the leaflet averaging 6 mm. long, the sheath 4-6 mm. long, 4.0—5.5 mm. broad, ciliate along the margins and often bristly on the back, 5- to 7-nerved, the teeth 2.5—4.5 mm. long; outer bracteole 1, 3-6 mm. long, ciliate at the apex; axis rudiment 3.5—6.5 mm. long, ciliate; inner bracteoles 2, 2.0-3.5 mm. long. Calyx tube 4.0—5.5 mm. long, the lobes about 2 mm. long. Standard suborbiculate, 4.0—6.5 mm. long; wings 4-5 mm. long, auriculate, clawed; keel petals falcate, 3.5—4.5 mm. long. Loment 1.5-2.0 mm. wide, reticulate and with one longitudinal nerve per face; only the upper articulation usually fertile, 3.5—4.5 mm. long, glabrous or nearly so; beak uncinate, glabrous or short-hairy on the inner face, 1.5-2.5 mm. long, less than half as long as the upper articulation. Stylosantbes erecta is distinguished by its loments which are mostly uni-articu- late and glabrous or nearly so, although the illustration which accompanies Beau- vais' original description depicts a bi-articulate loment. The shape and pubescence of the leaflets are quite variable and have given rise to two named varieties regarding which there has been considerable confusion. Taubert cites Welwitsch 2127 under var. guineensis, but this specimen is the type for var. acuminata. Var. guineensis is usually applied to those specimens which have bristly stems. Although including var. guineensis in his ‘Leguminosae of Tropical Africa’, Baker?" states that it is hardly varietally distinct from S. erecta. Since there is no line of demarcation with reference to degree of pubescence among the specimens, it does not seem wise to retain either variety. The inflorescence is usually much longer than broad, often becoming inter- rupted, although some specimens have more ovoid and compact heads. The leaflets in some specimens are punctate beneath. 2 Legum. Trop. Afr. 320. 1926. 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 315 In addition to the localities listed, the species is reported to occur in Sierra Leone, Angola, and Gaboon. The habitats include dry gravelly hills, sea sands, sandy pastures, and sandy thickets along rivers. BELGIAN Conco: Leopoldville, from Boma to Shinkakasa, Claessens s. n. (GH). RENCH WEST AFRICA: IVORY COAST: Grand Drewin, Roberty 13710 (MO). SENE- GAL: exact locality unknown, Sieber 37 (MO, NY). oast: Keta, Darko 584 (MO); “Pram and Pram”, Robertson 32 (MO). LIBERIA: exact locality unknown, Cook 137, 145 (US). NicERIA: Lagos (Apapa), Bels 43 (MO); exact locality unknown, Vogel s. n. (GH). 6. STYLOSANTHES SCABRA Vog. in Linnaea 12:69. 1838, ex char. Stylosanthes diarthra Blake, in Proc. Biol. Soc. Wash. 33:49. 1920. (T: Jabn 169!) Stylosanthes gloiodes Blake, loc. cit. 45. 1920. (T: Townsend A 57!) Stylosanthes plicata Blake, loc. cit. 46. 1920. (T: Kuntze s.n.!) Stems suffruticose, ascending to suberect, much branched, to 1.5 m. tall, densely and shortly hairy and often setosulous, sometimes viscid, rarely glabrescent. Leaf- lets elliptic to oblong-lanceolate, obtuse, mucronate, densely and shortly hairy above and below, sometimes glabrescent, with a few scattered setae, with 4—5 pairs of usually conspicuous veins; terminal leaflet to 15 mm. long, 5 mm. broad, the lateral ones to 12 mm. long and 4.5 mm. broad, often punctate beneath; petioles 2.5—6.0 mm. long, canaliculate above, scabrous with dense short hairs, the rhachis 1.0—3.5 mm. long; sheath and teeth of the stipules variable in length, the sheath usually 1.5—5.0 mm. longer than the teeth, short-hispid, 7- to 9-nerved. Spikes short, crowded, oblongoid, several-flowered; bracts unifoliolate, the leaflet about 4 mm. long, hispid, the sheath 3.5—6.5 mm. long, 3.0—4.5 mm. broad, densely hispid, about 7-nerved, the teeth 2.5—4.5 mm. long and somewhat shorter than the sheath; outer bracteole 1, lanceolate, 2.0—4.5 mm. long, bifid and ciliate at the apex; axis rudiment 4—5 mm. long, ciliate; inner bracteoles 2, 2-4 mm. long, ciliate at the apex. Calyx tube 3.0—6.5 mm. long, the more or less acute lobes 1.5—3.5 mm. long. Standard broadly obovate to suborbiculate, to 7 mm. long and 6 mm. broad; wings clawed, auriculate, spurred within at the base, 4—5 mm. long; keel petals auriculate, 3.5—4.5 mm. long. Loment about 2.5 mm. broad, reticulate- nerved; upper articulation 2-4 mm. long, shortly hairy, the lower 2-3 mm. long, evenly pilose throughout; beak uncinate, shortly hairy, 1-2 mm. long, one-half to one-third as long as the upper articulation. The species has a wide range but is apparently most abundant in the Minas Geraes area of Brazil. It is attributed to Peru by Bentham” (fig. 4). Stylosanthes scabra of $ sTYLOSANTHES is the viscid counterpart of Stylosanthes viscosa of § ASTYPOSANTHES. Boma: LA Paz: Larecaja, Mandon 698 (GH). : pata: exact locality unknown, Salzmann s.n. (MO). MATO GROSSO: exact BRAZIL: d locality unknown, Kuntze s.n. (NY, : S GERA Ituiutaba, Macedo 2226 (MO); Santa Luzia, Williams & Assis 201 (GH, MO), Barreto 5773 (F); Serro, Williams 21 In Mart. Fl. Bras. 152:1. 1859. [Vor. 44 316 ANNALS OF THE MISSOURI BOTANICAL GARDEN i is Assis 6850 Ce US). PARAÍBA: exact EC o: de Moer 887 (NY). pera, Pickel 2556 (GH, US, 2481 ati S, NY); Tiuma, Mil ong 67 a exact locality unknown, ion! KS '(GH, N TOM F) . RIO GRANDE D Jacui, Téodoro 1783 (GH). s&o PAULO: Campinas, Santoro 638 (US); sg nida "Barão oma See? Santoro 380 (US); San José do Barreiro, Hoehne & Gebr n s H). WITHOUT PRECISE LOCALITY: São Jose Bapt., Pohl 567 (NY). OMBIA: CUNDINAMARCA: Aguadita, Ricardo s.m. (US). mura: between Hobo and Leni e, Arbeláez & Cuatrecasas 8328 (US, F). EL VALLE: Cali, Bermúdez 20 (US), Killip & Cuatrecasas 38410 (US); Carretera al Mar, west of Cali, Killip & Lehmann V 39798 (US); Cisneros, Dagua Valley, Killip 11431 (US, GH, NY); north of Palestina, Garcia 6328 (US); Trés Cruces, Bermúdez & Barkley 17C869 (NY, US). Slept Fani Townsend A 57 (U ÉRIDA: Sierra de Nevada de Mérida, Jahn 108 (US); above Los Gon- Ee reed 10234 (F). TRUJILLO: Valera, Jahn 169 (US); Nagua, Warming 101 (US). 7. STYLOSANTHES TUBERCULATA Blake, in Proc. Biol. Soc. Wash. 33:48. 1920. (T: Britton 3330!) Stems suffruticose to 5 dm. tall, erect, branched, evenly and densely pilose and with dense short tuberculate-based hairs. Leaflets elliptic to oblong-lanceolate, acute and often mucronulate at the apex, the margins bristly-ciliate, both surfaces more or less appressed-pilose and often tuberculate-hispid at least below, with 4-6 pairs of prominent lateral nerves; terminal leaflet 10-20 mm. long, 3-5 mm. broad, the lateral ones somewhat smaller; petioles 2-4 mm. long, hispid-setose, the rhachis about 1 mm. long; sheath of the stipules 5-7 mm. long, pubescent and hispidulous as the stem, the subulate teeth 2-3 mm. long. Spikes narrowly oblongoid, 6-15 mm. long, 6- to 10-flowered; bracts unifoliolate, the leaflet pilosulous and densely tuberculate-hispid, the sheath 3.5—5.0 mm. long, hispidulous, the teeth slightly shorter; outer bracteole 1, 3-4 mm. long, bifid or entire, obtuse, ciliate or glabrous; axis rudiment 3—4 mm. long in fruit, pilose; inner bracteoles 2, 2-3 mm. long, narrow, Ciliate or glabrous. Calyx tube 4-5 mm. long, glabrous, the obtuse lobes 2.5-3.0 mm. long, ciliate or glabrous. Standard 4-5 mm. long, 3—4 mm. broad; wings somewhat auriculate at the base, short-spurred within, 4-5 mm. long; keel petals falcate, 3.0—4.5 mm. long. Loment reticulate-nerved, about 2.5 mm. broad; upper articulation of the loment fertile, 3-4 mm. long, usually sparsely or occa- sionally densely pilosulous on the ribs, the lower articulation fertile or abortive; beak slightly uncinate, shortly hairy, at least near the base, 1.5—2.0 mm. long. In savannas and on open banks at altitudes from 400 to 1800 meters (fig. 4)- BAHAMAS: NEW PROVIDE CUBA: cAMAGÜry: P ein «odes: pue E 2267 Eege Dec ANTIOQUIA: Medellín, Archer 1002 (GH, US). cau as ER Paila Holton 2 (NY). Huma: Quebrada de Angeles to Río ) Cabrera, Rusby & Pennell 325 (NY); Caine PU Plata 72 (US). META DE SANTANDER: Río Zulia, Nicifaro i n., 2412 ( US). = € GG Gest, A 22001 US Dagua, Kili 207 (NY, GH US); Jamundi, ERU: CREE tee a hace 3n ima, Arbeláez 2237 (US). bene 6451 (GH, US, F). ja, below Co Gerbe valley of the Mantaro, Weber- — DISTRITO FEDERAL: around C. cil Sted, Avis a de Ot o E 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 317 8. STYLOSANTHES sYMPODIALIS Taub. in Verh. Bot. Brand. 32:19. 1890. (T: Spruce 6373, photo!) Stylosanthes psammophila Harms in Fedde, Rep. Sp. Nov. 19:69. 1923. (T: Weberbauer 5936!) Stems erect, to 4.5 dm. tall, branched from near the base, shortly villous above, glabrate below, rarely with short bristles. Leaflets narrowly lanceolate to elliptic, acute at the tip, the larger ones to 40 mm. long and 5.5 mm. broad, evenly villous on both surfaces, usually with long marginal cilia, with 4—6 pairs of moderately conspicuous veins; petioles to 10 mm. long, usually much shorter, pilose, the rhachis about 2 mm. long; sheath of the stipules covered with soft short often rufous hairs, rarely with a few bristles. Spikes few-flowered; outer bracts usually tri- foliolate, the inner unifoliolate, the sheath about 2 mm. longer than the teeth, densely covered with soft often rufous hairs, rarely with tuberculate bristles, 5- to 7-nerved; outer bracteole 1, 2.5—3.0 mm. long, pilose near the apex; axis rudiment to 7 mm. long, with long rufous hairs; inner bracteoles 2, 2.0-2.5 mm. long, pilose near the apex. Calyx tube 3-5 mm. long, the acute, ciliate lobes 1-2 mm. long. Standard about 4.5 mm. long, suborbiculate; wings 3—4 mm. long, auriculate at the base; keel petals 3-4 mm. long, auriculate. Loment strongly reticulate-nerved, to 2 mm. broad; both articulations of the loment usually fertile, the upper about 3 mm. long, glabrous below, with tawny hairs above, the lower nearly as long, densely pilose; beak 2.0-2.5 mm. long, usually slightly shorter than the upper articulation, uncinate or almost with a complete coil, densely covered with short Occurs in sandy soils on the Galapagos Islands and in Peru and Ecuador. The specimens from the Galapagos Islands have leaflets 5-15 mm. long which some- times bear short erect hairs along with the rufous pilosity on the bracteal sheaths. The continental specimens have longer and more narrow leaflets (to 40 mm. long) and usually possess only appressed pubescence on the bracteal sheaths. These latter, specimens have been determined by some as S. psammophila. In his monograph of Stylosanthes, Taubert describes under S. scabra an un- named variety from the Galapagos Islands, stating "var. caulibus villoso- pubescen- tibus, non scabris.” These specimens belong to S. sympodialis. ADOR: GALAPAGOS ISLANDS: Abingdon Island, Snodgrass & Heller 836 (GH); Deeg Island, Tagus Cove, Stewart 1605 (MO, GH, US, NY), Snodgrass & Heller 171 (GH), Howell 9509 (GH); Bindloe Island, Snodgrass 8 Heller 767 (GH), Baur 90 (GH) ; Charles anes Snodgrass & Heller 450 (GH), Stewart 1697 (MO, GH, US, NY), Baur 92 (GH), Post Office Bay, Howell 8815 (GH, us D: ig Beach, Svenson 151 ( F); Con Island, ? Rond 144 (GH); Jervis Iland, B 9 (GH); North Indefatigable Island, Snodgrass & Heller 673 (GH); Santa Cruz tden; Rond 145 (GH); South Sey- Howell 9948 (US) Svenson 11301 ( ; Salinas, Svenson 11400 (US, GH); "uta d: Svenson 11258 NY); Salinas, La Puntilla, Asplund un (US); Salinas, Asplund 56 exact locality unknown, Eggers 14798 (US). WITHOUT PRECISE LOCALITY: Espinosa 517, 90 (US). [Vor. 44 318 ANNALS OF THE MISSOURI BOTANICAL GARDEN H soo 1000 1500 2000 eege cua a aem Fig. 4 PERU: prura: Parinas Valley, Haught 1 ; i Haugbt F (NY, between Piura and Ve? ; e E E e F); Ferreyra 5960 (US). a, Weberbauer 5936 (GH, US). Tunas: 9. SrvLosaNTHrs fruticosa (Retz.) Mohlenbrock, comb. nov. 1768, non L. 91, ex char. - 1841, ex chee. (T: Burke & Zeyher 404.) arv. & Sond. Fl. Cap. 2:227. 1862, ex char. Oliv. Fl. Trop. Afr. 2:156. 1871. (T: Kotschy 425!) Stem suffrutescent, about 5 dm. tall, much branched, erect or spreading, some- times Prostrate, shortly white-hairy above, occasionally with scattered short bristles, these sometimes yellow, often puberulent or glabrescent below. Leaflets acute, 1957) MOHLENBROCK—REVISION OF STYLOSANTHES 319 rarely obtuse at the tip, glabrous or puberulent to densely short-hairy, often with scattered short setae on the margins and midvein beneath, with 3—4 pairs of conspicuous nerves, or with the veins rarely obscure; terminal leaflet to 17 mm. long (rarely to 25 mm. long) and 3.0—4.5 mm. broad (rarely to 8 mm. broad), the lateral to 16 mm. long and 3.0—4.5 mm. broad; petioles 4-7 mm. long, hispid- ulous or shortly bristly, sometimes puberulent, the rhachis 1.0-2.5 mm. long; sheath of the stipules 4.5—10.0 mm. long, densely and shortly hairy and usually with scattered bristles, or sometimes only with bristles, with 5-7 pairs of con- spicuous nerves, the subulate teeth 3-6 mm. long, never longer than the sheath. Spikes ovate, dense, 4- to 10-flowered; bracts mostly unifoliolate, the leaflet (s) often plicate, the sheath tuberculate-bristly to merely long-ciliate around the margins, 3.0—6.5 mm. long, 2.5-5.5 mm. broad, about 7-nerved, the teeth averag- ing 3—5 mm. long, often nearly equaling the sheath; outer bracteole 1, 4—6 mm. long, ciliate at the apex; axis rudiment 3-5 mm. long, ciliate; inner bracteoles 2, 2.5—4.5 mm. long, ciliate at the apex. Calyx tube 4.5—6.5 mm. long, the lobes 2—4 mm. long, ciliate. Standard suborbiculate, 4.5-7.0 mm. long; wings 4-5 mm. long, auriculate at the base; keel petals 4—5 mm. long, auriculate. Loment 1.5— 2.5 mm. broad, reticulate; both articulations usually fertile, or the lower sometimes abortive; lower articulation, when fertile, 2.5—4.0 mm. long and as long as or slightly longer than the upper, pilose; beak somewhat uncinate or scarcely coiled, 1.5-3.0 mm. long, shortly hairy. The specimens of Stylosanthes from Africa and adjacent areas which belong to $ STYLOSANTHES seem to be treated best as belonging to two highly variable species: S. erecta from West Tropical Africa, and S. fruticosa from South Africa, Madagascar, India, and Ceylon. The former has loments which are glabrous or nearly so, while S. fruticosa possesses pubescent loments. Stylosanthes fruticosa is extremely diverse in its morphology, and a large num- ber of variants have been described. In this study, the only course that seemed practical was to consider all these variants as belonging to one exceedingly broad species. The basis for the deluge of species seems to be mostly the pubescence of the stems, leaves, bracts, and stipules. In all of these, however, the loment char- acters remain essentially uniform. Thus S. aprica described from Timor, S. bojeri from Zanzibar, S. flavicans from Sudan, and S. setosa from the Cape are being treated as S. fruticosa. S. sundaica of Malaysia, considered by most recent authors as a synonym for S. fruticosa, is really a synonym for a species in § ASTYPOSANTHES Specimens collected in Kordofan by Pappi (364) and by Kotschy (425) are the basis for Baker's S. flavicans. These specimens are flavescent when dried and some- what viscid and apparently represent an extreme in the S. fruticosa complex. Many of the bristly specimens of S. fruticosa are at least subviscid, however, and no dividing line may be drawn to separate the specimens. Specimens called S. bojeri are very bristly and are sordid upon drying. This again seems to represent only an extreme, and intergradations of all degrees of pubescence may be foun Apparently the first mention of this species in the literature was in 1737 when [Vor. 44 320 ANNALS OF THE MISSOURI BOTANICAL GARDEN Burmann?? recorded the polynomial ‘Trifolium procumbens zeylanicum hirsutum, In 1768 Burmann f.” noted that Hedysarum hamatum occurred in Ceylon and Jamaica. It is evident that he included under this binomial at least two species, the present S. hamata (L.) Taub. from Jamaica which Linnaeus called Hedysarum hamatum a in 1759 and S. fruticosa from Ceylon. In 1791 Retzius?* named Arachis fruticosa from Ceylon and included under it Linnaeus’ Hedysarum hama- tum a. Nine years later, Willdenow?? described S. mucronata and listed Arachis fruticosa as a synonym. Therefore it appears that the first legitimate epithet to be applied to this species is fruticosa. Kenya: trail from Nyeri to Wambugu, Mearns 1956 (NY, US); 14 miles from Mom- bassa, vicinity of Changamme, Mearns 2230 (US) OZAMBIQUE: Lourenço Marques, Howard 56 (US), Barle 82 (US), Curtis 80 (GH); Lourenço Marques, Marracuene, Quintas 6 (MO); Delagoa Bay, Kuntze s. n. (NY); Ma- puto, Sousa 3960 (MO). NicERIA: on the Niger, Baikie s. n. (GH) ; exact locality unknown, Barter 3428 (GH). SOUTHERN RHoprsia: Matapos, Plowes 1426 (MO); Nyamandhlovu, Plowes 1603 (MO); Umtali, Chase 3231 (NY). SUDAN RDOFAN: Nile Land, Kotschy 425 (GH). TANGANYIKA: Ugogo, Dodoma, Peter 44443, 44463 (MO); Nzega, 50 miles north of Tabora, Carnochan 10 (GH); near Manyoni, north of Kilimatinbe, Carnochan 338 (GH); a E TH AFRICA: TRANSVAAL: Barberton, T borncroft 1913 (NY); near Crocodile Poont, Dyer & Verdoorn 3426 (NY); Pretoria, Mogg 14785 (US). ZULULAND: Somkele, Wood 9291 (US). CEYLON: exact locality unknown, Thwaites 1451 (GH, NY). East INDiEs: exact locality unknown, Wight 814 (NY). Innia: Mysore and Carnatic, Thompson s. n. (GH); Salem, Hosur, Yeshoda 415 (NY). 10. STYLOSANTHES SUBSERICEA Blake, in Proc. Biol. Soc. Wash. 33:50. 1920. (T: Purpus 7152!) pieces scoparia Standl. & L. Wms. in Ceiba 1:145. 1950. (T: Williams & Molina II20I ems suffruticose, usually erect, slender to robust, to 1 m. tall, often much branched, long-pilose and usually sericeous throughout, occasionally with a few bristles near the inflorescence, sometimes glabrescent. Leaves crowded, the leaflets lanceolate, acuminate, the terminal to 20 mm. long and 4.5 mm. broad, the lateral to 15 mm. long and 3.5 mm. broad, shortly white-hairy on both surfaces with occasional bristles below, with 3-5 pairs of conspicuous veins often forming 4 submarginal nerve; petioles 5-7 mm. long, densely white-hairy, the rhachis 1.5-2.0 mm. long, pubescent as the petioles; sheath of the stipules 5-9 mm. long, sericeous with occasional scattered tuberculate-based bristles, 9- to many-nerved, the teeth lance-subulate, 6.0-7.5 mm. long. Spikes narrow, oblongoid, 4- to 10-flowered, 12-20 mm. long, on peduncles 0.5—4.0 mm. long; bracts unifoliolate, often short- ?? Burmann, Thes. Zeyl. 1737. 23 Burmann f., Fl. Ind. 167. 1768. 24 Retz. s. Bot. Fasc. 5:26. 1791. 25 Willd. Sp. Pl. 3:1166. 1800. 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 321 bristly on the back, 4.0—5.5 mm. wide, the sheath 4.0-5.5 mm. long and 7- to 9-nerved, the teeth 3.5—5.0 mm. long; outer bracteole 1, 4-7 mm. long, pilose nearly throughout; axis rudiment about 6 mm. long in fruit, long-ciliate; inner bracteoles 2, 3-4 mm. long, ciliate. Calyx tube 6-8 mm. long, the lobes 3-5 mm. long, the upper two lobes obtuse and long-pilose, the lateral ones similar but some- what shorter, the lowest acute, pilose. Standard suborbiculate, 5-8 mm. long; wings mostly obovate and with a short claw, auriculate, shortly spurred, 4-6 mm. long; keel petals 3.5—5.0 mm. long, spurred within at the base. Loment about 2.5 mm. wide, conspicuously reticulate; both articulations usually fertile, the lower 3-4 mm. long and densely long-pilose, the upper 3.5—4.0 mm. long, sometimes slightly exceeding, sometimes slightly shorter than the lower, covered with long white hairs; beak 2.5—5.0 (—7.0) mm. long, white-hairy, with a distinct coil. Stylosanthes subsericea is the woodiest member of the genus and sometimes at- tains a height of one meter. The circinate beak of the pubescent loment nearly equals or surpasses the upper articulation in length. The lower leaves fall early, leaving the stem bearing only the persistent subulate stipules near the base. All the Honduran specimens previously have gone under the binomial S. sco- paria. However, the great similarity of floral structure and vegetative characters of the Honduran plants with S. subsericea of Mexico indicates these two epithets to refer to the same species. The Mexican plants lack tuberculate bristles on the lower surface of the leaflets. The species is very local and is known only through a few collections from Hon- duras and Oaxaca, Mexico, where it grows in regions of dry rocky hillsides and dry ravines, usually with pines and oaks, at elevations between 600—850 meters. Honpuras: COMAYAGUA: vicinity of Comayagua, Standley 6 Chacon 5866 (F), 5400 (B. RAZÁN: Río Yeguare, near El Zamorano, Standley & Molina 4631 (F); Que- brada de Santa Clara, near Río Yeguare, Standley & Williams 1 598 (F); above El Zamo- between Tatascan and Maraita, Molina 4079 (US); Camino Viejo betwe Tegucigalpa, Standley 14213 (US, F); near Rio Yeguare, below El Zamorano, Standley 12110 (MO, F); Río Capa Rosa, El Zamorano, Rodríguez 3662 (US, F); Quebrada de Santa Clara, near Río Yeguare, Standley & Williams 1634 (US, F); along Río Yeguare, east of El Zamorano, Standley 1077 (US, F). EL Paraiso: Quebrada de I d between Las Mesas and Yuscarán, Standley 14983 (MO, F); vicinity of Güinope, Standley, Williams, Molina t$ Padilla 2077 (F); south of Güinope, Standley 14862 (F); Rio Ye- guare, Molina 1681 (F); Caleras, Williams & Molina 14133 (F). Mexico: oaxaca: Cerro de Picacho, Purpus 7152 (GH, NY, US, MO, F); near Oaxaca, Rose & Hough 4584 (US). [Vor. 44 322 ANNALS OF THE MISSOURI BOTANICAL GARDEN B s. sussericea 4 S. MACROCARPA | € S. HAMATA Ke € S. BIFLORA SÉ | : M H T Fig. 5 11. STYLOSANTHES MACROCARPA Blake, in Proc. Biol. Soc. Wash. 33:47. 1920. (T: Pringle 6721!) Stems herbaceous throughout, much branched from the base, spreading or " cending, to 20 cm. tall, densely covered with mostly appressed-ascending pili, with short tuberculate bristles beneath each node. Leaflets elliptic, subacute, mucronate, light green, glabrous above, sparsely bristly on the midvein beneath and along the margins, with 3—4 pairs of conspicuous veins; terminal leaflet to 13 mm. long and 3.5 mm. broad; petioles 3-5 mm. long, sparsely pilose and often shortly bristly, the rhachis 1.0—2.5 mm. long; sheath of the stipules about equaling the teeth, 3—4 mm. long, pubescent like the stem. Spikes ovoid, to 15 mm. long, 5- to 10-flowered; outer bracts trifoliolate, the inner unifoliolate, the sheath pilose and bristly, 6-9 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 323 mm. long, about twice as long as the teeth, 7-nerved; outer bracteole 1, oblong, about 4.5 mm. long, ciliate near the apex; axis rudiment to 8 mm. long in fruit, long-hairy; inner bracteoles 2, 2.5—3.0 mm. long. Calyx tube 5-6 mm. long, the obtuse lobes 2-3 mm. long. Standard obovate, not clawed, about 6 mm. long; wings auriculate below and shortly appendaged within; keel petals falcate. Loment 6.0-8.5 mm. long, 3.0—3.5 mm. broad, reticulate; usually only the upper articula- tion fertile, 4.0—4.5 mm. long, evenly appressed-pilosulous, the lower abortive or occasionally fertile, densely pilose; beak strongly uncinate, evenly appressed- pilosulous, 3.5—4.0 mm. long, usually equaling the superior articulation. Stylosantbes macrocarpa, known only from hills near Oaxaca, Mexico, closely resembles S. mexicana but differs in fruit morphology. The loment of S. macro- carpa is evenly appressed-pilosulous throughout while that of S. mexicana is glabrous or puberulent only along the nerves. The flowers examined show two inner bracteoles although Blake, in his original description, states the number of "bractlets" (= inner bracteoles) to be 1, a character not in keeping with the other members of § sTYLOSANTHES. One of the large sub-basifixed anthers in some of the flowers is about twice as large as the other four. This seems to be only a sporadic anomaly. Mexico: oaxaca: near Oaxaca, Pringle 5782 (GH); Monte Alban, Pringle 372 (GH). 12. STYLOSANTHES NERVOSA Macbr. Field Mus. Publ. Bot. 13:411. 1943. (T: Weberbauer 6215!) Stems suffrutescent, to 1 m. tall, erect, branched, usually villous with inter- spersed tuberculate-based hairs, at least when young. Leaflets oblong-lanceolate, acute to acuminate, glabrous or nearly so above, puberulent beneath with occasional tuberculate hairs (at least on the costa), with 4—6 pairs of conspicuous nerves, the terminal leaflet to 20 mm. long and 3 mm. broad, the lateral somewhat smaller; petioles 2-4 mm. long, pilose with occasional tuberculate hairs, rarely glabrous, the rhachis about 1 mm. long; sheath of the stipules about 7 mm. long, usually equaled by the subulate teeth, pilose, glabrate, with 5—7 nerves. Spikes narrowly oblong- lanceolate, 2- to 8-flowered; bracts unifoliolate, the leaflet ciliate, the sheath 5-7 mm. long, with numerous tuberculate hairs, densely ciliate, 5- to 7-nerved, the teeth usually 3-5 mm. long, sometimes bearing tubercles; outer bracteole 1, 4—5 (-7) mm. long, ciliate at the apex; axis rudiment 3—4 mm. long, ciliate; inner bracteoles 2, 3—4 mm. long, ciliate at the apex. Calyx tube about 7 mm. long, the lobes 3 mm. long, the upper 4 obtuse, ciliate or glabrous. Standard 5-6 mm. long; wings somewhat auriculate at the base, shortly spurred within; keel petals 3.5—5.0 mm. long, falcate. Loment about 2 mm. broad, reticulate; only the upper articulation usually fertile, about 3.5 mm. long, evenly pilose throughout; beak about 2.5 mm. long, pilose, strongly uncinate but usually not completely coiled. 26 In Proc. Biol. Soc. Wash. 33:47. 1920. [Vor. 44 324 ANNALS OF THE MISSOURI BOTANICAL GARDEN Infrequent over a broad range, known only from a few stations in Venezuela, Peru, Bolivia, and northern Argentina (fig. 4). ARGENTINA: CORRIENTES: exact locality unknown, Ibarrola 2310 (US). sArTA: La Merced, Venturi 5136 (GH, US). BOLIVIA: SANTA CRUZ: Buena Vista, Steinbach 5397 (NY, GH). PERU: CAJAMARCA: Jaen, near mouth of Chinchipe River, Weberbauer 6215 (NY, GH, US, F). VENEZUELA: Serranias de Terepaima, Saer 631 (F). 13. STYLOSANTHEs HAMATA (L.) Taub. Verh. Bot. Brand. 32:22. 1890. Hedysarum hamatum a L. Syst. Nat. 10:1170. 1759, ex char. Stylosanthes procumbens Sw. Prod. Veg. Ind. Occ. 108. 1788, ex char. Stylosanthes humilis Rich. ex Hemsl. in synon., Biol. Centr. Am. Bot. 1:272. 1882, ex char. Stylosanthes eriocarpa Blake, Contrib. U. S. Nat. Herb. 24:4. 1922. (T: Blake 7792!) Stems ascending, spreading, prostrate, or matted, to 1 m. tall, often much branched, usually with a line of fine pubescence on one side of the stem, sometimes glabrous near the base or rarely throughout, or occasionally sericeous throughout. Leaflets lanceolate to elliptic, obtuse to subacute, to 20 mm. ong and 6 mm. broad, glabrous or shortly pilose above and below, with 3-6 pairs of conspicuous veins; petioles 2-6 mm. long, puberulent or pilose, the rhachis 0.5—2.5 mm. long; sheath of the stipules shortly hairy, sericeous, or glabrous, 3- to 11-nerved, usually ex- ceeding or occasionally surpassed in length by the teeth. Spikes rather small, ovoid to oblongoid, few- to 15-flowered; outer bracts trifoliolate, inner usually unifolio- late, very variable in pubescence of the sheath, mostly only ciliate, the sheath 5- to 7-nerved; outer bracteole 1, 2.5—5.0 mm. long, ciliate at the apex; axis rudiment 2— mm. long; inner bracteoles 2, 2.5—3.5 mm. long, ciliate at the apex. Calyx tube 4.0-7.5 mm. long, at least twice as long as the more or less acute and ciliate lobes. Standard 4-5 mm. long, suborbiculate; wings 3.5—4.5 mm. long, clawed and auriculate; keel petals 3.0—4.5 mm. long, falcate. Loment about 2 mm. broad, reticulate; both articulations usually fertile, the upper 2—4 mm. long, glabrous or puberulent in localized areas, or densely and evenly sericeous, the lower, when fertile, somewhat shorter, pilose or glabrescent; beak equaling to exceeding the upper articulation, glabrous or with short pubescence, uncinate. Stylosantbes bamata is readily distinguished by a series of characters which includes a more or less puberulent bi-articulate loment with a beak slightly exceed- ing the upper articulation and a general lack of tuberculate bristles. This species appears to be related most nearly to S. fruticosa of Africa and, indeed, the two were not separated by such early authors as Burmann f. and Retzius. Stylosantbes fruticosa usually possesses tuberculate bristles and evenly pilosulous loments. Stylosanthes bamata is similar to S. sym podialis but in the latter the beak of the loment is shorter than the upper articulation. The spikes are usually 2- to 5-flowered, but some specimens from the Antilles have large dense spikes which may bear as many as 15 flowers. The habit of this species may be erect to ascending, or prostrate and matted. This difference seems 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 325 to be correlated with altitude; the specimens from elevations between 100 and 900 meters tend to be upright while those near sea level seem to be matted. These latter have been known as S. procumbens, The range of S. hamata is broad and includes Florida, eastern Mexico and Cen- tral America, the Antilles, Colombia, and Venezuela, often becoming a pest in gardens in Jamaica, Cuba, and neighboring islands. UNITED STATES: FLORIDA: Dade Co.: Moldenke 778 (MO). BAHAMAS: ANDROS ISLAND: Spaniard Creek, Small & Carter 8882 (F, NY, GH, US); Mastic Point, Brace 7097 (NY, F); road to Little Creek, Brace 5245 (NY). Caicos ISLANDS: South Caicos, Wilson 7637 (GH, NY, F); Jacksonville, Millspaugb & Mills- paugh 9100 (NY, F). CROOKED ISLAND: Marine View Hill, Brace 4635 (NY, F); Land- rail Point, Brace 4678 (NY, F). ELEUTHERA d urrent, Coker 330 (NY). GRAND BAHAMA ISLAND: Eight Mile Rocks, Britton & Millspaugh 2454 (GH, NY, US, F). GRAND TURK ISLAND: The Wells, Millspaugb & Millspaugb 9342 (NY, US, F); "Waterloo, Millspaugh & ch a el 9034 (F). GREAT RAGGED ISLAND: exact locality unknown, Wilson 7821 (GH, NY, F). NEW PROVIDENCE ISLAND: Nassau, Hitchcock s. n. (MO, F); Clifton Point, Degener 18985 (GH, NY, Mast Nassau, Northrop & Northrop 37 (GH, , F), Curtiss 9, 118 (GH, US, NY, F), Coker 541 (NY), eu fa Fincastle, Wright 86 (GH, NY, F), off Soldier’s Dee? KT 223 (GH, NY), road to Lake Cun dn Britton 93 Sua ce of Bay Street Road, Millspaugb 2172 (E * ; eid locality un n, Brace 444 (F). M CAY: Port pores Brace 3934 (NY, F). SAN SALVADOR "ND phin GE unknown Zäre 7359 (N EY: Pueb , Cayo e omano, Shafer 2499 (NY). HABANA: Habana Palmer P Riley ó21 x "Babs 460 (NY); Regla, Sbafer 104; Playa de Cojimar, Hitchcock s. n. (F); Mar ios Ser 6089 (US). ORIENTE: Santiago, Morro Hill, Mills- paugh 1077 (US, F), Howard 5788 (GH, US, NY, MO); Guantanamo Bay, Britton 1941 (US, NY); Holguin to etc Shafer 1551 (US, NY); Santiago, Hessel 105 (NY); Santiago, Morro Castle, Havard 82 (NY); ra Underwood & Earle 1682 (NY). UT PRECISE LOCALITY: Triscornia, Hitchco ck s:n: (F). DomıNıcan REPUBLIC: BARAHONA: Los Patos, Abbot I en oe bg How- ard & Howard dee 3 (GH, NY), Fuertes 211 (GH, M F), von Türckbeim 2804 (NY). sAN JUAN; near San Juan de la Maguana, Howat p UAE 4795 (GH); El Cercado, Cas Santiago, Hondo, Howard & Howard 9293 (NY). santiaco: Santiago, Jiménez 473 (US). INGO: Monción, banks of Río Mao, Valeur 802 (US); Gua ve uhin, Abbott 937a (US), 942 (US); near Puerto Liberatador, Manzanillo Bay, Howard & Howard 9659 (NY); Monción, Valeur 248 (NY, US, MO, F); Ciudad Tru- v Allard 14350 (US), 13067 (US, F), 13114 (US), diae sh 13313 13725 (U O), 14007 (US), 14281 (US); Azua, Rose, Fitch & ussell 4028 (US, NY). Geesse CISE LOCALITY: Moca, Ekman NH 11 270 (US); Zeg Faris 38 Con NY); Llanos de Rafael, Eggers 1920 (NY, US); Constanza, von Türck eim 2529 (NY). Granp Cayman: South Sound, Kings GC 69 (MO); Bodden Bay road, Millspaugh 1335 (F). E 0463 (US); Gros More, t — EE ed Leonard 2863 de aiit GONAIVES Qu eg n e te R AR e roa re Neuve, Nash & Gk (NY). GONA roit, Leonard 3401 (GH, US). MIREBALAIS: vam s locality unknown, Cook, "Scofield " Doyle 81 (US). pu nord: St. Michel de l'Atalaye, Leonard 7037 (US); Plaisance to Marmelade, Nash 678 (NY). NORD-OUEST: Jean-Rabel, eona 12925 Leonard 1 ); , US). Holdridge 928 (US). LA VALLEE: soe e d, Leonard & Leonard 11224 (US, MO), 15583 (US). WITHOUT LITY: Plaine Centrale, Pavane-Papaye, an PRECISE H6013 (US); La Cumbre, R Geer II rg (US). [Vor. 44 326 ANNALS OF THE MISSOURI BOTANICAL GARDEN Jamaica: Sherwood Forest, ie 3137 (MO); Port Morant, Hitchcock s. n. (MO); Kingston, Hitchcock s. n. (MO); r Port Antonio, Fredbolm 3054 (US, NY); Tre- lawny, Jackson Town, Hu nnewell forti (GH); Waterloo Road, Campbell 6019 (NY); n = Mandeville ben 107 (N R ANTI ADA: exact locality unknown, Britton & Teros Jor uw T. QUIA re locality unknown, eg 553 (NY d St, Jobas, Shafer GC exact locality unknown, Hutson (US). BADOS: “Water r- ichael, Dasb 281 (US, NY, F); et Borrell 281 (US); ETRA, T 62 (US). BONAIRE: exact locality unknown, Boldingh 7246 (NY). AO: San Cruz, Britton & sn 3013. (NY). GRENADA: Richm ond Hill, Broadway s. n. (US, Y). GUADELOUP ointe-à-Pitre, Hammarlund 18 (NY); exact cetur unknown, Questel 1459 (US). ‘Nacua: Salt Pond Hill, "Fairchild 2599 (US); Matthew Town, Nash & Taylor 1069 (NY, F). MONA: exact locality unkno gét Stevens 6224 (NY), 6344 . = BARTHÉLEMY: Gustavia, Questel 116 (NY), 217 (NY); exact locality un- orsström s. n. (F). sr, com: Corn Hill, KN A 216 (GH, MO, NY, US, i S ta Ferry, Thin 482 (GH, NY); Bassin, Ricksecker 47 (US, MO, F). sr. AN: Bethania to Rosenberg, Shafer 232 (US, NY). st. KiTTs: near Basseterre, Britton & Cowell 729 (US, NY). sr. THoMas: Svenson’s Ee Britton 280 (NY); Paradise Bay, Britton & Britton 208 (US, NY). TORTOLA: Sea Cow Bay, Britton & Shafer 667 (GH, NY). VIRGIN GORDA: exact locality unknown, Su 163 (G a NY). LITTLE CAYMAN: South Side, Blossom Point, Kings LC 75 (MO). Puerto Rico: Quebradillas, Stevenson 5887 (US); Guanica, lo 3756 (NY, US); Yauco, Underwood 6$ Griggs 658 (US, NY); Guayama, Ven , Goll 531 (US); Cayey, south of Caguas, Heller & Heller 328. (US, NY, Di Bay , Sintenis 1092 (US, MO, NY); Condado, Britton, Britton & Brown 5764 (US, NY ee UATEMALA: IZABAL: trail from Los Amates to Izabal, Blake 7792 (GH, US); Puerto Barrios, Standley 25126 (NY, U : COCLÉ: “Pesonong Shefer 36 (NY). OLOMB TL uerto Colómbia, de Kattah, Molina & Bark- ley Ato25 (US), Aug (05); Barranquilla, Elias 927 IUD Puerto Colombia, Dugand 4008 (US). sotrivar: La Popa, icd Zeg ith 14067 (NY, GH, US, F); Cartegena, Molina & Barkley Boo me (US). cu ARCA: east of Apulo, Killip, Dugand & Jara- SH 38164 (US). MAGDALENA: Ris Tx lity unknown, Castañeda 252 (MO). TOLIMA: in yE Arbeláez & Mgr Ae? 65 1506 (US "ven ALCÓN: Adicora, Paraguaná, Tomayo 750 (US US, F). NUEVA ESPARTA: S Ae of Margarita, Miller E o ag 68 rayo 750 ( US. MG, b. 2 14. STYLOSANTHES SERICEICEPS Blake, in Contrib. U. S. Nat. Herb. 20:524. 1924. (T: Jabn 678! tems suffruticose, erect, to 0.5 m. tall, densely covered with shortly appressed silvery hairs, glabrescent below. Leaflets lanceolate, to 25 mm. long and 5 mm. broad, acuminate, mucronate, usually glabrous above, minutely pubescent beneath and along the margins, with 4—7 pairs of prominent veins; petioles to 6 mm. long, sericeous, the rhachis to 2.5 mm. long; sheath of the stipules about as long as the teeth, densely short-hairy, with 3 veins crowded near the center. Spikes oblongoid or obovoid, about 15 mm. long, crowded, several-flowered; outer bracts often tri- foliolate, inner unifoliolate, the sheath 5-7 mm. long, about equaling the teeth, densely covered with long tawny hairs, 5-nerved; outer bracteole 1, bifid, ciliate; about 4 mm. long; axis rudiment to 4 mm. long, densely pilose; inner bracteole 1, about equal in length to the outer but narrower. Calyx tube to 7 mm. long, the more or less pilose lobes obtuse or acute, about 3 mm. long. Standard to 8 mm. long, suborbiculate; wings 4.5-6.5 mm. long, auriculate; keel petals 4.5—6.0 mm. long, falcate. Loment 1.5-2.0 mm. broad, reticulate; both articulations fertile or 19571 MOHLENBROCK—REVISION OF STYLOSANTHES 327 the lower sometimes abortive, the upper 2.5-3.0 mm. long, evenly clothed with short silvery hairs, the lower, when fertile, nearly 4 mm. long, pilose; beak 1.5-2.0 mm. long, nearly half as long as the upper articulation, strongly uncinate, densely white-hairy. Stylosanthes sericeiceps apparently marks the transition between § sTYLOSAN- THES and § ASTYPOSANTHES in that only the lowest 2-3 flowers in each spike possess an axis rudiment and that each flower, in addition, is subtended by only one inner bracteole, the latter character one which is found throughout $ AsTYPOSANTHES. The absence of the axis rudiment in upper flowers has been noted for some speci- mens in other species of $ srYLOsANTHES (S. bamata and S. subsericea), but this is the only species in which this is combined with the absence of one of the inner bracteoles. Stylosanthes sympodialis closely resembles S. sericeiceps, differing from it in the rufous villosity of the bracteal sheaths and in the unevenly pubescent loment. There are no data as to the habitat of this species, except that it grows at altitudes between 400 and 1000 meters. It is known only from Venezuela. NEZUELA RIDA: EU ce] Jabn 678 (US, GH, NY); between Estangues and Poo. Real, Che I Y (NY, US). Section II. AsrvrosawTHEs (Herter) Mohlenbrock, stat. nov. Sect. Eustylosanthes Vog. in Linnaea 12:63. Astyposantbes Herter, in Rev. Sudamer Bot. 7: 209. 1943. None of the flowers subtended by an axis rudiment; inner bracteole 1. In the absence of previous typification, Stylosanthes humilis HBK. is desig- nated as the type for $ ASTYPOSANTHES. 15. STYLOSANTHES BIFLORA (L.) BSP. Prelim. Cat. N. Y. Pl. 118. 1883. Trifolium veier L. Sp. Pl. 773. 1753, ex char. i t. Fl. Carol. 183. 1788, ex char äech es Ais rs in Svenska Vet. Akad. Handl. 11:296. 1789, ex char. Stylosantbes FEEN Michx. Fl. Bor. Am. 2:75. 1803, ex char. tylosanthes hispid Stylosanthes bispide var. nudiuscula Michx. loc. cit. 1803, ex char. Stylosanthes a erecta Pursh, Fl. Am. Sept. 2:480. deng ex char. Séylounibe bi voy! procumbens Pursh, loc. cit. nn, exc e CECR E ve B bispidissima Torr. & Gray, Fl. N. Am. d 354. 1838, ex char. (T: «b i r. rr. Bot. Club 24:565. 1897. (T: ebe 674!) Stylosanthes riparia Kearney, in Bull. To Stylosanthes biflora var. dese Poll. & Ball, in Proc. Biol. Soc. Wash. 13:134. 1900, ex char. Stylosanthes biflora var. bispidissima Mohr, in Contrib. U. S. Nat. Herb. 6:570. 1901, ex Siyloseutbes floridana Blake, in Proc. Biol. Soc. Wash. 33:51. 1920. (T: Sudworth s. n.!) Stylosanthes riparia var. setifera Fern. in Rhodora 40: 438. 1938. (T: Fernald & Long I! eati a riparia f. ochroleuca Fern. loc. cit. 1938. (T: Fernald &$ Long 8732!) [Vor,. 44 328 ANNALS OF THE MISSOURI BOTANICAL GARDEN Stems herbaceous to subligneous at the base, to 6 dm. tall, erect or spreading, much branched or rarely sparsely branched, glabrous to puberulent to densely hispid, the bristles sometimes yellowish. Leaflets ovate to elliptic to lanceolate, obtuse to acute at the tip, glabrous on both surfaces and occasionally with spinulose teeth on the margins, sometimes punctate beneath, with 3—6 pairs of nerves; ter- minal leaflet to 40 mm. long and 10 mm. broad, usually much smaller, the lateral ones slightly smaller than the terminal; petioles 1-3 mm. long, glabrous to puberu- lent to hispidulous, the rhachis 0.5—1.5 mm. long; sheath of the stipules 3-10 mm. long, glabrous to puberulent to densely hispid, with numerous nerves, the subulate teeth 3-8 mm. long, often setose. Spikes small, 1- to 8-flowered; bracts mostly unifoliolate or the outer sometimes trifoliolate, the sheath puberulent to densely hispid, rarely glabrous, 3.0—6.5 mm. long, 2.5—5.5 mm. broad, mostly 5- to 9- nerved, the teeth usually somewhat shorter; outer bracteole 1, lanceolate, 2-3 mm. long, glabrous to ciliate at the apex; axis rudiment none; inner bracteole 1, 1.0—2.5 mm. long, glabrous to ciliate at the apex. Calyx tube 2.5—5.0 mm. long, glabrous, the more or less acute lobes 2-4 mm. long, glabrous to puberulent. Standard suborbiculate, 4.5-7.0 mm. long, to 5.5 mm. broad; wings 3.5—4.5 mm. long, auriculate, spurred within at the base; keel petals denticulate to entire at the apex, auriculate, spurred within at the base. Loment 2.5—3.0 mm. broad, strongly reticulate with vertical nerves; only the upper articulation of the loment fertile, 2.5-5.0 mm. long, obovoid, plump, puberulent throughout; beak 0.5—1.0 mm long, coiled. The Stylosanthes biflora complex is defined as including populations not only of S. biflora but also S. riparia and S. floridana of other authors. The members of this complex are relatively constant with respect to loment characters but are diverse in leaflet shape, pubescence, and to some extent habit. However, some variation may occur in shape of the loments with some articulations more inflated than others, or in degree of reticulation. Loments are generally puberulent although they may become glabrous with age. Degree of maturity affects such characters of the loment so that in some specimens, the entire range of variability may be observed. Most authors refer specimens with acute leaflets and without bristles on the stem to S. biflora (L.) BSP., and those with acute leaflets and bristles on the stem and bracteal sheaths to S. biflora var. bispidissima (Michx.) Pollard & Ball. Speci- mens with obtuse leaflets and which may be either non-bristly or bristly are regarded mostly as S. riparia Kearney or its variety setifera Fern., respectively. Specimens from the extreme southeastern United States which lack bristles completely are known sometimes as S. floridana Blake. A sharp line of demarcation among these variants is, however, difficult to determine, a fact which has been noted previously by some workers (Isely27), Supposedly the best criterion to separate S. riparia from S. biflora is the position of the beak on the upper articulation of the loment. In S. riparia, the beak is said to 27 Towa State Coll. Jour. Sci. 30:113. 1955, 19571 MOHLENBROCK—REVISION OF STYLOSANTHES 329 be oriented terminally on the loment rather than at one side as in S. biflora. However, in most specimens identified as S. riparia, the beaks appear lateral. Isely attributes the differences between fruits to "artifacts of observation” since "the degree to which the beak is or is not symmetrically placed may depend on the angle from which it is viewed". Other varieties or species which have been segregated from S. biflora seem only to represent extremes. UNITED STATEs: ALABAMA: DeKalb Co: Ruth 338 (MO). Lee gee Zeg s.n. (MO). Marshall Co.: Hubricbt B 1667 Mesh Mobile Co.: Graves 961 (M ARKANSAS: Baxter Co.: E. J. doen 5961 (MO). Ca SE Demaree 17794 ( M. Craighead Co.: eg 3495 (M ê (MO). CH g O o iv br Ka 3 8 bs kt SS ka D 2. 2 96 aree 19460 (MO). Johnson Co.: Demaree 19966 (MO). Lincoln Co.: Demaree I91 56 (MO). Logan Co.: Segen 17682 (MO). Miller Co.: Heller & Heller 4126 (MO). Prairie Co.: Demaree 22303 (MO). Pulaski Co.: Demaree 17293 (MO). Saline Co.: Demaree 21195 (MO). Sharp Co.: Emig 168 (MO). Union Co.: Demaree 19421 (MO). ll Co.: Demaree 21270 (MO). DELAWARE: Newcastle Co.: Commons s. n. (MO). Sussex Co.: Churchill s. n. (MO). Fromma: Duval Co.: Curtiss 6418 — Escambia Co.: Brinker 473 (MO). Hills- borough Co.: Ferguson. s.n. (MO). Jefferson Co.: Hitchcock s.n. (MO). Lake Co.: Nash 1309 (MO). Marion Co.: Woodson e sdey 97 (MO). Volusia Co.: Hood s. n. (MO). Groncia: Bulloch Co.: Harper 946 (MO). Cobb Co.: Duncan 8646 (MO). De- Kalb Co.: Miller, Perry, Boyd & Marn 531 (MO). n Co.: Allard 129 (MO). Madison Co.: Duncan 11588 (MO). Oconee Co.: Small s. n. (MO). Stephens Co.: fades 11737 "MOT Union Co.: Duncan 7835 (MO). ‘Whitfield Co.: Harper 393 (MO). ILLINOIS: Jackson Co.: Moblenbrock 5058 (M Kansas: Cowley Co.: White 134 (MO). EE Co. Rydberg & Imler 454 (MO). KENTUCKY: a. Co.: Wherry & Pennell 13873 (MO). Whitley Co.: Wherry 8 Pernel 2 I an (M UISIANA: Bet Parish: Ball 621 (MO). risit Montgomery Co.: Painter 1084 (MO) ISSOURI Bush 326 O). Benton teyermark 24436 (MO), 24436a (MO). Butler Co.: Steyermark 11489 (MO) Callaway Co eyermark 26175 ). Camden teyermark 20655 (MO). Cole Steyermark 14981 (MO) 5 +: Crawford Co.: Steyermark 15373 (MO). Dade Co.: Steyermark 5653 (M Co.: Kellogg 995 (MO). Douglas Co.: Steyermark 15373 (MO). Dunklin Co.: i w Blankinsbip s. n. (MO) n Co.: Pammel s. n. ( as Palmer 461 (MO) Jefferson Steyermark 1366 (MO). Lincoln teyermark 8130 (MO). Mc- Donald Co.: Palmer 6a (MO). Mari Steyermark 15303 (MO). Marion Co.: Davis 7866 (MO). Oregon Co. lmer & Steyermark 417. MO) Co.: Steyer- Palme mark 25893 (MO). Polk Co.: Steyermark 23 MO). Reynolds Co.: Steyermark 14201 (MO). St. Louis Co.: Drushel 4311 (MO). Shannon Co.: Bush s.n. (M ds Steyermark 20763 (M Ei Stone A Steyermark m. Mo). Taney Co.: m 5888 (MO). Vernon Co.: Drouet I Sëch 11217 (MO) o (MO Way Webster Co.: Steyersiutk 23697 (MO). Wright Co.: ddr s. n. (MO [Vor. 44 330 ANNALS OF THE MISSOURI BOTANICAL GARDEN New Jersey: Burlington Co.: Meredith s.n. (MO). Camden Co.: Martindale s. n. (MO). Gloucester Co.: Parker 12430 (MO). Hunterdon Co.: Fisher s.m. (MO). Middlesex Co.: Mackenzie 2871 (MO). Ocean Co.: Churchill s.n. (MO). Salem Co.: Meredith s. n. (MO). NortH CanoLINA: Cherokee Co.: Fox 4034 (MO). Cumberland Co.: Fox t$ God- .: Fox 4080 (MO). arn E x Da MA: Caddo Co.: Goodman 2342 (MO). Choctaw Co.: Houghton 4025 (MO). Comanche Co.: Stevens 1411 (MO). Johnston Co.: Palmer 6438 (MO), 6438 (MO). McCurtain Co.: Houghton 3788 (MO). Payne Co.: Stratton 188 (MO). PENNSYLVANIA: Bedford Co.: Berkheimer 3458 (MO). Lancaster Co.: Eisenhower s.n. (MO). Montgomery Co.: Redfield 1524 (MO). Philadelphia Co.: Eby s.n. (MO). Schuylkill Co.: Collector Unknown (MO). York Co.: Glatfelter s.n. (MO). SourH Carouina: Aiken Co.: Eggert s.n. (MO). Anderson Co.: Davis 8441 (MO). Dillon Co.: Godfrey SC49040 (MO). Georgetown Co.: Godfrey 146 (MO). Laurens Co.: Davis s.n. (MO). Oconee Co.: Duncan 11255 (MO). Pickens Co.: Rodgers 385 (MO). TENNESSEE: Cocke Co.: Kearney 675 (MO); Kearney 674 (MO). Texas: Bowie Co.: Eggert s. n. (MO). Brazos Co.: Palmer 11750 (MO). Dallas Co.: Reverchon 193 (MO). Harris Co.: Boon 185 (MO). Lee Co.: Eggert s. n. (MO). Nacogdoches Co.: Parks RX2287 (MO). Parker Co.: Tracy 8025 (MO). Travis Co.: Buckley s. n. (MO). Waller Co.: Hall 153 (MO). 16. STYLOSANTHES GUYANENSIS (Aubl) Sw. Svenska Vet. Akad. Handl. 11:296. 1789. Stems herbaceous or mostly suffruticose, erect or rarely scandent, to 1 m. tall, glabrous to puberulent to pilose or often hispid or setose, the setae often purplish. Leaflets green to purple, lanceolate to elliptic to nearly ovate, acute or occasionally. obtuse at the apex, glabrous to puberulent to tuberculate-bristly, particularly on the margins and the costa beneath, sometimes spinulose-toothed, with 2—7 pairs of nerves, the terminal usually 5-30 mm. long and 2-10 mm. broad (rarely to 45 mm. long and 20 mm. broad), the lateral ones slightly smaller than the terminal; peti- oles 1-10 mm. long, sometimes canaliculate above, glabrous to puberulent to setose, the rhachis 0.5-1.5 mm. long; sheath of the stipules glabrous to pilose to setose, 2-15 mm. long, 3- to many-nerved, the subulate teeth 2—10 mm. long. Spikes small, narrow, and 2-flowered to large, globose, and about 40-flowered; bracts mostly unifoliolate, the leaflet sometimes much reduced and often spinulose- toothed, the sheaths green to purple, glabrous to puberulent to bristly, 3—7 mm. long, 3- to 9-nerved; outer bracteole 1, lanceolate or ovate, 2.5—3.0 mm. long, acuminate, glabrous to pilose, occasionally purplish at the apex; axis rudiment none; inner bracteole 1, 2-3 mm. long, acuminate, glabrous to pilose, colored like the outer bracteole. Calyx tube 4-8 mm. long, glabrous to sparsely pubescent, the 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 331 lobes 3-5 mm. long, obtuse to acute, ciliate or sometimes pilose. Standard 4-8 mm. long, 3—5 mm. broad, suborbiculate; wings 3.5—5.0 mm. long, auriculate, spurred within at the base; keel petals 3.5—5.0 mm. long, falcate. Loment 1.5-2.5 mm. broad, reticulate; only the upper articulation fertile, glabrous to minutely short- pubescent near the apex, or occasionally with sessile glands near the apex, ovoid, plump, 2-5 mm. long; beak minute, 0.1—0.5 mm. long, strongly inflexed. KEY TO THE SUBSPECIES a. eur green, usually 10—60 mm. long, sometimes tuberculate-bristly but not spinulose- n internodes usually much pn than 1.5 cm. long; gp mostly pe red. Central America, South America . guyanensis ssp. Sen aa. aru See 4—12 mm. long, densely and shortly pe on the mE e costa, ADAE B internodes 0.5— orescence 1- red. Mex 16 guyanensis ssp. ` disifors Stylosanthes guyanensis is a most variable species. The leaflets range from small and lanceolate to large and elliptic. The indument of the stem is diverse. The spikes are large and to 40-flowered to small and 1- to 2-flowered. Numerous sub- species, varieties, and forms have been proposed, but intergradations occur between almost all. Except for subspecies dissitiflora, there is no strong geographical cor- relation with any of the variants. A great number of the most southern specimens are 40-flowered and bristly and the northern specimens 5- to 20-flowered and more or less viscid, but this is far from constant. Great confusion as to the proper name for this species has arisen. Aublet de- scribed Trifolium guyanense as having pubescent loments and illustrated it thus. The type photograph, however, appears to coincide with many of the specimens examined in this study with glabrous or puberulent loments. Those considering Aublet's T. guwyanense as inapplicable to present known specimens employ the binomial S. gracilis for this broad species. Numerous other morphological extremes have been proposed: S. bispida Rich. with densely hispid stems; S. surinamensis Miq. with rather narrow leaflets; S. ruellioides Benth. with densely pubescent stems and broad leaflets; S. longiseta Micheli with the stems bearing long bristles; and S. pohliana Taub. with acuminate leaflets. Stylosanthes ingrata Blake from Vaca Falls, British Honduras, is without fruit but seems to be S. guyanensis ssp. guyanensis. Stylosanthes guyanensis ssp. guyanensis occurs from Central America through South America into northern Argentina, also in the Antilles. It has been planted in Australia in an effort to impede soil erosion. Stylosanthes guyanensis ssp. dissiti- flora is limited to a small area in southwestern Mexico. It occurs at altitudes between 3500 and 5000 feet. [Vor. 44 332 ANNALS OF THE MISSOURI BOTANICAL GARDEN 16a. STYLOSANTHES GUYANENSIS ssp. GUYANENSIS Trifolium guyanense Aubl., Pl. Guian. 776. 1775, ex char. (T: Aublet s. n., photo!) Stylosanthes hispida Rich., Act. Soc. Hist. Nat. Par. 2:112. 1792, ex char Stylosanthes gracilis HBK., Nov. Gen. et Sp. 6:507. 1823, ex char. Stylosantbes surinamensis M ae Minnie 18:567. 1844, ex char. St A Beier enth.,in Mart, PL Bras. 151:90. 1859, ex char. Stylosantbes guyan is B subviscosa Benth loc. cit. 92. 1859, ex char. Stylosanthes (ees Micheli, n Mem. Soc. Phys. Genéve 28:7. 1883, ex pe Stylosanthes pohliana Taub. in Verh. Bot. B 32:29. 1890. (T: Pobl 1 and. 9. !) Stylosantbes phim var. jubesctii Pilger, i in Engl. Bot. Jahrb. 30:160. "901, ex char. (T: Pilge Stylosanthes juncea f. intermedia Chod. & Hass. in Bull. Herb. Boiss. ser. 2, 4:884. 1904. $ D Ge mont E var. longiseta (Micheli) Chod. & Hass. loc. cit. 885. 1904. (T: . Hassler 457 Stylosantbes guyanensis var. genuina Hass. in Fedde, Rep. Sp. Nov. 16:220. 1919. (T: Hassler 7781! Stylosanthes t ciis var. genuina f. esetosa Hass. in Fedde, loc. cit. 221. 1919. (T: 34!) Hassler Stylosanthes pee var. intermedia (Vog.) Hass. in dex lbe cot 1919 Stylosanthes guyanensis var. intermedia ssp. anomala Hass. in Fedde, loc. cit. 1919, ex char. Siylosaathei $ Finition var. subviscosa f. viscosissima Hass. in Fedde, loc. cit. 1919. (T: Hassler 6454! Stylosanthes guyanensis var. longiseta (Micheli) Hass. in Fedde, loc. cit. 222. 1919. ieee guyanensis var. marginata Hass. in Fedde, loc. cit. 223. 1919. (T: Hassler Stylosanthes ingrata Blake, in Proc. Biol. Soc. Wash. 39:51. 1928. (T: Record s. n.!) Stylosantbes gracilis var. vulgaris Burkart, in Darwiniana 3:247. 1939, ex char. Stylosantbes gracilis var. subviscosa (Benth.) ) Burkart, loc. cit. 248. 1939. BRITISH West INDIES: TRINIDAD: St. Joseph savanna, Britton, Hazen t$ Broadway 976 (GH, a NY). Bn URAS: BELIZE: Belize, et 11250 (US, NY). cavo: Mt. Pine Ridge, pe 11706 (US, GH); Vaca Falls, Record s. n. (US). sTANN a Stann — Schipp 486 (MO, GH, NY, F); oe ann dE Gentle 1883 Seed joe Swasey Branch, Monkey tiem Gentle 3898 (GH, F, US, NY, M A Rica: cartaco: Juan Viñas, Rowlee t$ Stork 839 E ei. Ca artago, Léon Be (F). HEREDIA: Santa Barbara, Pittier 1656 (US). sAN José: San ntiago, Brenes 14326 (GH, US); San José, Tonduz 2169 (US, GH); El Giant Skutch 2462 (NY, US); San Pedro de la Calabaza, Tonduz ono (US); Santa Maria, Standley 41505 (US); between Aserri and Tarbaca, Standley 34031 (US). wiTHoUT PRECISE LocaLITY: Boruca, Tonduz 4707 (US); Río Colorado, Venio 278 dé l EL SALVADOR: LA LIBERTA n Finca Germania and Finca San Antonio, near ARTS S : vicin- tapan, ewe & Padil e dbi is 318 (US, H H. NY). la ' 3098 (F), 3261 (F). SAN VICENTE: San Vicen GUATEMALA: ALTA matin: Coban, Standley 70111 (NY, F); Coban, von Türck- beim 84 (US). c HIMALTENANGO: Alameda, Joknston E (F); yet Chimaltenango to San Martín eer Standley 57982 (G NY, F); San Martin Jilotepeque, Standley 64402 (NY, F); San Juan-Ma lagua, M. 1555 (E. emmer. EF Rincón, Soa Río Guacalate, Stendley 49369 (P), 78200 (Fi ; near Escuintla, wer dd 63577 (GH, F). GUATE- ala and Fiscal, Standle- ey 59690 (F); near Finca La Roje 47 (US); exact locality unknown, 529 a Bretafia, between Aurora, Araile 18 (F); Gille 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 333 (US). HUEHUETENANGO: along Rio Cuilco, between Cuilco and San Juan, Steyermark 50838 (F); Rio Pucal, Standley 65833 (F); east of San Rafael Pétzal, Standley 83051 (F), 83076 (F); 14 miles south of Huehuetenango, Standley 823098 (F). 1zaBAL: between Los Amates and Izabal, Kellermann 7556 (NY, F), 7257 (NY, F); between Milla and Cadai, Steyermark 38355 (F), 38653 (F); Gualán, Blake 7691 (US); Cristina, Blake 7585 (US); Quiriguá, Standley 24223 (GH, US) ; near Quiriguá, Standley 72470 (F). JALAPA: vicinity of Tina. Standley 76661 (F); ihia of Jalapa, Standley 76803 (F). JUTIAPA: vicinity of Jutiapa, Taa Zë 75511 (F). QUEZALTENANGO: Colomba, Skutch 2048 (US, GH, n Cotzal and San pees ae 45109 (F); "eh locality un- knows, healer erg (F ). RETALHULEU: near uleu, — 88622 (F). SACA- TEPÉQUEZ: Barranco Hondo, Standley 60232 GE "er Santa María de Jesús, Standley 59401 (GH, NY, F); Ciudad Vieja, Tejada 273 (US); near rod age? 6172 3 jj ). SANTA ROSA: DECH E of Bar n Standley 77879 (F); Buena Vista a, der? ux 752 (F); La Vega, Heyde & Lux 4457 (US, GH). SUCHITEPÉQUEZ: Finca Las ire eh mark 35442 T » Tinca Moti. Skutch 1463 (GH). ZACAPA: nh de las Minas, Steyermark 29647 (F HoNDURAs: CH cinity of San "us rcos de Colón, ene? 15706 (US). COMAYAGUA: Bees ‘Stone 56264 (US, F); vicinity of Comayagua, Standley & aere 5972 (F); EL A t , near Ser: Standley 56148 (F). p ee Espiritu Sancto, Blake AUS). MORAZÁN: El dus rano, Molina 61 (F, US); San Antonio del Oriente, Bedden 668 (F), 687 (F); Piedra pepe Standley 11 037 (F); near Joya Grande, Standley & Molina 4466 (F) ; Guaimaca, Molina (F). orANCHO: fro Catacamas to Loma Pelona, Standley 1828 3 (F). EL ‘pana: Quebrada de Cessna am 2 169 Merrill & Williams 15713 (F); Ma anzaragua Road, Williams & Molina 11484 (MO, GH, F); Ofo de Agua, Williams & Molina 12027 (GH, F); Güinope, Rodríguez 1678 (F). SANTA BARBARA: Los Dragos, Standley & Lindelie 7405 (F). HIAPASs: exact locality unknown, Doyle 144 (US). JALAPA: Hacienda hs h Kënns Schiede 631 (NY). yatisco: San Sebastian, Nelson 4075 (US, GH). AXACA: Ja lapa, Pringle 9174 (GH). puesta: Metlatoyuca, Goldman 57 (GH, US). VERACRUZ: Minatitlan, sed s. n. (NY); Zacuapan, Purpus 1888 (NY, MO, US, GH, F); t genie Botteri 152 (GH). WITHOUT PRECISE LOCALITY: Lodiego, Palmer 1900 (NY, S, F); Monte Geer Matuda 1958 (US, NY, GH); Vallecitos, Montes a Oca, Hasson 11456 (US). NICARAGUA: GRANADA: near Granada, Grant 756 (GH). JrNoTEGA: San Rafael del Norte, Miller t$ Griscom 66 (US); southwest of Jinotega, Standley 9534 (F), 10082 (F), 10094 (F). MA: CANAL ZONE: Sosa Hill, Balboa, Standley 26443 (US); eriam Fort SoA ton and Corozal, KE 29159 (US); Summit, Standley 30068 (US); along Las Cru Trail, Hunter & Allen 753 (GH, MO); Rio Puente, Dodge, Steyermark & Allen "Sp oy) ; Pie NE to Mandinga, Piper 5129 (US). CHIRIQUÍ: Kalmin Terry 1264 (GH, US, M OUS cocti: Penonomé, Williams 167 (US, NY), 148 (US, Y); Natá, Allen 838 (GH, US, MO, NY). Panamá: Sabanaz, Standley 25837 ay Taboga Island, Sie ger 3 (US); between Matías Hernández and Juan Diaz, Standley 32077 (US); Chépo, Hunter & Allen 75 (MO, GH); Pacora, Bro. Maurice 790 US). ARGENTINA: -e Sierra Chicas, Estancia la Reduccion, iber 7397 (F). CORRIENTES: Estancia, Santa Teresa, Pedersen 79 (US). MISSIONES: Santa Ana, Rodri- guez Lx "de gen Ekman 1742 (NY); San Ignacio, Ekman 1740 (US). : : Guanai-Tipuani, Bang 1459 (MO, US, GH, NY, F); Nord Yungas, Coripatar as to (US, NY); Tipuani, Bücbtien 5423 (US, GH, MO, NY, F); ungas, San Bartolomé, near Calisaya, Krukoff 10536 (US, MO, NY, F); Tampa, Bücbtien 1787 (US); H Ege Tate 1031 (NY); Mapiri, Tate 466 (NY), 471 (NY); near La Paz, gf 8 (NY); Guanai, Rusby 2324 (NY). SANTA CRUZ: Velasco, Kuntze s. n. (NY "gs Sará, Buena Vista, eenig 6916 (GH, NY, F, MO); San José, Williams 354 a WITHOUT PRECISE LOCALITY: Eskia, White 612 (NY). [Vor. 44 334 ANNALS OF THE MISSOURI BOTANICAL GARDEN BAÍA: Mucugé, Serra Sincorá, Froes 20128 (NY). CEARÁ: Lagoa Porangaba, Syne uve 2726 (GH ee Us: e DISTRITO FEDERAL: Praía de Sernambetiba, L. B. pee 6359 (US). co1íás: Corgo d o Jaraguá, Pohl 949 (NY); Quexada, Maceil 2132 MINAS GERAES: São o do Paraiso, Vidal 3059 (GH); Caldas, Regnell Gre? (US); Do MA = uer, Ilheos, Moricand s.n. (GH); Viçosa, Mexia 4523 (NY, GH, MO, F); near Lagóa Pampulha, William: ms t$ Assis 6084 (GH); Estacao de enis Jue & Assis 7122 (GH); Belo Horizonte, Williams t$ Assis 5832 (GH, US); Bento Pires, Williams & Moreira 5219 (GH); Tijuco, Riedel 1316 (US); 168 ( ; i Chapeu de Sol, Archer & Barreto 5015 (US); Jaboticatubas, L. B. Smith 7078 (US) Diamantina, Archer 4084 (US); Serra do Taquaril, Barreto 5771 (F); Ituiutaba, Macedo ni (US). pará: Vigia, Campina do Palha, Pires 4083 (NY); Gurupá, Ducke 16182 US); Colares, Poeppig 2030 (NY); Belém, Baldwin 4557 t PARAÍBA: Areia, de exiis 885 (NY). Paraná: Villa Velha, Dusén 14781 (GH, F); Morungaba, Dusén 16568 (GH); Tibagy, Reiss 75a (NY, F); Jaguariaiva, Dusén 16381 (GH), 16506 (MO). PERNAMBUCO: Tapera, Pickel 2403 (US). RIO DE Bi lesa Cabo Frio, Rose & Russell 20693 (US); Tijuca, Riedel 1256 (US). s&o PA litatis Hoebne 137 (NY, GH); Iparanga, Luederwaldt 13323 ANY); Capivari, Riedel 1316 (US). WITHOUT PRECISE LO- CALITY: Serra do Cipo, Conceicao, Barreto (F); Rio Claro and Poracutu, Pohl 2107 Go Ser Mikan s.n. (NY); Corrego Padre, Pobl 1997 (NY, F). H GUIANA: Pirara, Schomburgk 381 (US); basin of Gre nuni River, near mouth of mie? Creek, A. C. Smith 2376 (MO, NY, US, F); Orealla savanna, Berbice, Irwin 330 (US); exact locality unknown, Schomburgk 240 (US). OLOMBIA: ANTIOQUÍA: Hayo Rico, Corres & Barkley 18A200 (US); Medellin, Archer 644 (US); Bello, Archer 436 (US); San Antonio, Bro. Daniel 530 (US, F); Fre- donia, Bro. Tomas 957 (US); El Prado, Bro. Tomas 1009 (US); El Carmen, Bro. Daniel 2310 (US). BOGOTÁ: Caqueza, CK 4327 (US). Boyacá: Mt. Chepon, Lawrence 332 um GH, hes AUCA: SE Lebman id (US); El Tambo, von Sneidern 1414 eta Pennell 1111. (NY); San Augustin, Bro. Daniel 27 (US, F). MAGDALENA: Pueblo Loge Valle Dupar, Bro. Angel 673 (US); "ee Marta, Jordan, H. H. Smitb 58 (US, NY, H, F); Jagua, Allen 688 (MO). meta: Cabuyaro » Sprague s. n. (US); Las Lagartija, southwest o ribe, Fosberg 10414 (US); Villavicencio, Bro. Apollinaire s. n. (US); El Mico airstrip, Verde Idrobo & s 1310 (US). NORTE DE SANTANDER: del Sarare, Río Cubucón, Cuatrecasas 18608 (U (US, F). saNTANDER: Río Casanare, Esmeralda, Los Llanos, Cutters se TOLIMA: iia. Pennell 3351 (NY) ; west of San Lorenzo, Pennell 34901 (NY, US, GH); Mariquita, Pennell 367; (US, NY, GH). EL VALLE: amaria, on Río Timba, near Timba, Fosberg 20518 (NY, US); Las Neives, west of Cali, Killip, Cuatrecasas Sg D ryander 39200 (US, F); Pavas, Pennell 5505 (NY, GH), 5532 Ga GE SE E east of Zarzal, Pennell, Killip & Hazen 8461 i : anta Rosa, Penne í i, Pichindé 150 (US); Come RETRO ae ip 6073 (US, NY); Rio Cali, Pichindé, Duque chilla de Santa anc S iiim 22966 (US); El Forge, near Buenaventura, Killi Cuatrecasas 38938 (U AUPÉS: Río Guaviare, San José del Guaviare, Cuatrecasas SCH 4585 ner? e LOCALITY: no location given, Mutis 2345 (US), 2976 ee exact locality unknown, Espinosa 590 (US). FRE UIANA: vicinity of Cayenne, ber Cowan E) Made aloe (NY); exact locality un eng all LINT D, ec PARAGUAY: AMAMBAY: Sierra de Aubeibey Hassler 1 NY, : Yhú, Hassler 9532 ( GH). cENTRAL: Cerros de EE ARS T : ae NY. LA CORDILLERA: Cordillera de Altos, Fiebrig 405 virer Cete eda eda stanislio, 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 335 Hassler 6014 (GH, NY), 5996 (NY, GH), 7030 (NY, GH), I (NY). WITHOUT PRECISE LOCALITY: Centurion, [I ie 4397 (GH, US); Santa i Hassler 2834 (MO, GH); Berghang, San Luis, Fiebrig 4453-4413 (GH). Peru: ayacucHo: Aino, Killip & Smith 22605 (NY, US). cuzco: San Pedro, Vargas 2576 (GH); Valle de Santa Ana, Herrera 3614 (F). Junín: Chanchosmayo Valley, Schunke 1410, 1412, 1518 (F). sAN MARTIN: near Moyobamba, Klug 3638 (NY, i US, MO, F); Tarapoto, Woytkowski 35097 (MO, F). TUMBES: La Palma, Anderson 948 (US). SURINAM: Republiek, Kuyper 18 (US); Coppename River, haa hat Pulle 279 (MO). VENEZUELA: ANZOÁTEQUI: Los Caños, Pr T 14472 (US); Río Cani, Garroni 62 L an Sa T Bolívar and El Cristo, Killip 37217 S, Es fio Geh Cardona 614 (US). coPrprs: San Carlos, Rudd 331 (US). DISTRITO FEDERAL: Turmerito, GC 371 38 (US, NY); Caracas, Bailey & Bailey 97 (US); El Limon, , Wiliams 10540 (F). MéRwa: Tovar, Fendler 297 (GH, MO); Monga de Zerpa, Lossen 314 (US). MoNacas: between Soen and San Anemia, Seyer , OI ( SZ KEE Carretera Trujillo-Boconó, Tomayo 846 pbc ie cage, Pittier 13150 (NY, MO, US). zuria: Mene Gra x Pittier 70588 (US). ISE LOCALITY: Tobay, Gebriger 358 (MO); Entrada, Warming 100 ( (US) mg mmm de T “@ s. GUYANENSIS SUBSP. GUYANENSIS f| = E -© S. GUYANENSIS SUBSP. DISSITIFLORA e 40 ^ SCALE ^ 1000 1300 mes 2 1000 2000 3000 E ansuctene - [Vor. 44 336 ANNALS OF THE MISSOURI BOTANICAL GARDEN 16b. STYLOSANTHES GUYANENSIS ssp. dissitiflora (Robins. & Seat.) Mohlenbrock, stat. nov. et (cA ake Stylosanthes dissitiflora Robins. & Seat. in Proc. Am. Acad. 28:105. 1893. (T: Robinson I72! Stylosanthes purpurata Blake, in Proc. Biol. Soc. Wash. 33:52. 1920. (T: Rose 2942!) Stylosantbes eciliata Blake, in Univ. Calif. Publ. Bot. 10:409. 1924. (T: Purpus 9246!) CO: JALISCO: near Guadalajara, Rose & Painter 7478 (US, GH); southwest of Autlán, McVaugb 14222 (US); near Etzatlan, Rose & Painter 7543 (US, NY); Río Blanco near Guadalajara, Pringle 11433 (US, GH) ; near Guadalajara, Pringle 5172 (GH). 17. STYLOSANTHES ANGUSTIFOLIA Vog. in Linnaea 12:63. 1838, ex char. (T: Sellow s. n.) Stylosantbes angustissima Klotzsch, in Schomburgk, Faun. Fl. Brit. Guin. 1200. 1840, ex char. Astyposantbes angustifolia (Vog.) Herter, in Rev. Sudamer. Bot. 7:209. 1943. Stems usually woody below, erect, much branched from the base, to 1 m. tall, slender, densely long-hispid or glabrescent. Leaflets linear to linear-lanceolate, acuminate, to 40 mm. long and 2.5 mm. broad, glabrous or rarely with scattered bristles on both surfaces, the margins often involute; petioles 2-16 mm. long, shortly bristly or nearly glabrous, the rhachis 1.5—3.5 mm. long; sheath of the stipules 3.5-7.0 mm. long, glabrous or usually with long tuberculate bristles and often villous, 5- to 9-nerved, the subulate teeth 2.5—4.5 mm. long. Spikes elon- gate, narrow, to 45 mm. long, 3- to 20-flowered; bracts unifoliolate, the leaflet often reduced to a small laminal projection, the sheath villous and usually with long bristles, 3—4 mm. long, 9- to 11-nerved, the teeth 2.0—3.5 mm. long; outer bracteole 1, 2.5-3.0 mm. long, ciliate near the apex; axis rudiment none; inner bracteole 1, 2.5-3.0 mm. long, ciliate. Calyx tube 3.5—5.0 mm. long, the lobes 1.5-2.5 mm. long and usually ciliate. Standard suborbiculate, 3.5—4.0 mm. long; wings 3—4 mm. long, auriculate at the base, shortly appendaged within; keel petals falcate, 1.5-2.5 mm. long, densely white hairy. Loment 0.8—1.5 mm. broad, 2-3 mm. long, faintly nerved, minutely pubescent, with only the small upper articula- tion fertile; beak 2.5—4.0 mm. long, strongly uncinate to circinate, usually over twice as long as the upper articulation. Stylosanthes angustifolia seems to prefer low sandy areas in woods. It is known from British Guiana, French Guiana, Surinam, and northern Brazil (fig. 7). t ANHAO: exact locality unknown Gardner ARÁ: Ilha do Mosqueiro, Killip & Smith 30423 (NY, US); near Santarem, Spruce s.n. (GH, ); exact sia. Teme. 1 paul: exact locality un n, Gardner 20904 (NY). Rio BRANCO: Caracara y, Froes 23664 ; Surumu, Ule 816r (U RIO GRANDE DO NORTE: Angicos, Swallen 4715 (US)- WITHOUT PRECISE LOCALITY: tropics, Burcbell H BritisH Guiana: Roraima, R. Schombu ee Ne 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 337 18. STYLOSANTHES LEIOCARPA Vog. in Linnaea 12:64. 1838. (T: Sellow 4235!) Astyposantbes leiocarpa (Vog.) Herter, in Rev. Sudamer. Bot. 7:209. 1943. Stems herbaceous to somewhat woody at the base, simple to much branched, ascending to suberect, nearly 0.5 m. tall, often viscid, densely and shortly bristly to glabrous near the base. Leaflets usually oblong, mucronulate at the apex, the terminal to 15 mm. long and 3 mm. broad, the lateral to 12 mm. long and 3 mm. broad, with short tuberculate-based bristles on both surfaces, with 2—4 pairs of very obscure veins not forming any submarginal nerve; petioles 3.5—4.5 mm. long, hispidulous, the rhachis 1.5-2.0 mm. long; stipular sheaths 4—6 mm. long, with 9—11 distinct veins, long-bristly, the subulate teeth 5-6 mm. long. Spikes 2.5—4.0 mm. long, 8- to 10-flowered; bracts unifoliolate, the leaflet hispid to short-bristly, the sheaths bristly, 3.5—4.5 mm. long with 7, rarely 9-11 veins, the teeth 2.5—3.5 mm. long; outer bracteole 1, 2.5—3.0 mm. long, ciliate near the tip; axis rudiment none; inner bracteole 1, 2.0—2.5 mm. long and ciliate. Calyx tube 2.5—3.0 mm. long, the lobes 2.0-2.5 mm. long and ciliate. Standard suborbiculate, 4—6 mm. long; wings 3.5-5.0 mm. long, auriculate, spurred within at the base; keel petals 3.5—4.5 mm. long, falcate. Loment reticulate with 1 or 2 additional longitudinal nerves besides the reticulate venation, about 2 mm. broad; both articulations fertile, or the upper sometimes abortive, or the lower abortive; upper articulation when fertile 2-3 mm. long and glabrous, the lower, when fertile, 1.5-3.5 mm. long and glabrous; beak somewhat uncinate or nearly straight, 2.0-2.5 mm. long, glabrous. The species is known only through a few collections from Uruguay, Paraguay, southern Brazil, and southern Colombia. A specimen collected by Arsène from Morelia, Mexico, is obviously S. leiocarpa, but this extreme extension of the range seems unlikely. A note penciled on the sheet by S. F. Blake in 1924 states that this is "probably one of the sheets distributed by Otto Weigel and is badly mixed." BRAZIL: MINAS GERAES: Osorio, Morro Grande, Rambo S1747 (US). RIO GRANDE DO sUL: São Leopoldo, Leite 1909 (GH); São Joño do Monte Negro, Bornmüller 657 (GH). SANTA CATARINA: Palhoca Massiambu, Reitz & Klein 653 (US). URUGUAY: MONTEVIDEO: Pooitos/ Carrasco, Het TDI (GH, NY, MO, F); T video, Baratini s. n. (MO); Carrasco, Osten (US). saw José: Barra, Herter (MO). WITHOUT PRECISE LOCALITY: Cabo Santa Maria, Rosengurtt B2471 (NY). 19. SrvLosaNTHES cayennensis Mohlenbrock, sp. nov. Herba erecta ad 30 cm. alta caulibus parce ramosis glabris vel puberulis. Foli- ola elliptico-lanceolata supra glabra subtus punctulata puberula setosa nervis 4—6-gemmatis; foliolo terminali ad 12 mm. longo 5.5 mm. lato; petiolo 2-4 mm. longo glabro vel puberulo; rhachide 1-2 mm. longa; stipulae striatae vagina 3—6 mm. longa saepissime puberula processibus subulato-mucronatis setosis 3-5 mm. longis. Spicae densae oblongoideae floribus 3—10; bracteis plerumque unifoliolatis stipuliformibus vagina puberula vel setosa 3-6 mm. longa nervis 5; bracteola exteri- [Vor. 44 338 ANNALS OF THE MISSOURI BOTANICAL GARDEN DER Ae ^ ISBN \ 7 T9 K * o $00 1000 1500 2009 MILES LE 1000 2000 3000 KILOMETERS Fig. 7 ore 1, 2.5-3.0 mm. longa apice ciliata; axis rudimento nullo; bracteola interiore 1, 2.5-3.0 mm. longa apice ciliata. Calycis tubus 3.5-5.0 mm. longus glaber lobis 1.5- 2.5 mm. longis. Corolla lutea; vexillo suborbiculato 4—5 mm. longo; alis falcatis 3.5- 4.0 mm. longis; carina falcata 3.5—4.0 mm. longa. Lomentum 1.5-2.0 mm. latum reticulatum, articulo superiore 2 mm. longo glabro vel raro puberulo, articulo inferiore 1.5-2.0 mm. longo glabro, rostro circa 1 mm. longo valde uncinato vel circinato glabro. ; Known only from wet habitats in the vicinity of Cayenne, French Guiana. It is related to S. guyanensis because of the glabrous loments and the short strongly uncinate beak but differs in its bi-articulate loments. FRENCH GUIANA: vicinity of Cayenne, Broadway 231 (US, NY, GH), 672 (NY; GH), 972 (GH, HOLOTYPE). y 231 (US, ), 67 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 339 20. SrvrLosaNTHES hippocampoides Mohlenbrock, sp. nov. Herba erecta ad 60 cm. alta caulibus parce ramosis setosis. Foliola lineari- lanceolata supra setosula subtus setosa nervis 3—4-gemmatis; foliolo terminali ad 15 mm. longo 2.5 mm. lato; foliolis lateralibus ad 15 mm. longis 2 mm. latis; petiolo 2-4 mm. longo puberulo vel setosulo; rhachide circa 1 mm. longa; stipulae striatae vagina 3—8 mm. longa glabra vel setosula processibus subulatis mucronatis setosis 3-5 mm, longis. Spicae densae ovatae floribus 2-6; bracteis plerumque unifoliolatis stipuliformibus vagina setosa 3—6 mm. longa; bracteola exteriore 1, 2-3 mm. longa apice ciliata; axis rudimento nullo; bracteola interiore 1, 2.0—2.5 mm. longa, apice ciliata. Calycis tubus 3.0—5.5 mm. longus glaber lobis 1.5—3.0 mm. longis. Corolla lutea; vexillo suborbiculato 3.5—5.5 mm. longo; alis falcatis 3—4 mm. longis; carina falcata 3.0-3.5 mm. longa. Lomentum circa 2 mm. latum valde reticulatum, articulo superiore 2.5—3.0 mm. longo glabro, articulo inferiore abortivo, rostro 2 mm. longo uncinato glabro. Stylosanthes hippocampoides closely resembles S. montevidensis and S. macro- soma. Stylosanthes hippocampoides has conspicuously reticulate, glabrous loments about as broad as long and a 4- to 8-flowered inflorescence; Stylosanthes macrosoma has obscurely nerved, glabrous loments about twice as long as broad and a 2- to 6-flowered inflorescence; Stylosanthes montevidensis has a reticulate-nerved, pubes- cent loment about as broad as long and a 4- to 40- flowered inflorescence. The specific epithet is derived from the loment which superficially is shaped like the sea-horse (Hippocampus). The species is infrequent in Uruguay and northern Argentina. Some specimens were included by Burkart in his S. gracilis var. rostrata. The presence of a distinct beak on the loment clearly distinguishes this species from S. gracilis of Burkart. ARGENTINA: CORDOBA: Unquillo, Cabrera s.n. (F); Sierra Chica, Kurtz s.n. (NY); Sierra Achala, Kurtz 8303 (NY, HOLOTYPE); Potrera de Loza, Galander s.n. (NY). rs: Ituzaingo, Puerte Loreto, Schuranz s.n. (US); exact locality unknown, c Burkart 6743 (GH). MisioNEs: Concepción, Clos 2023 (GH). i Urucuay: coLónIA: Riachuelo, Herter 15562 (GH, US, NY, MO, F); exact locality unknown, Clos 3231 (GH). 21. STYLOSANTHES MACROSOMA Blake, in Proc. Biol Soc. Wash. 33:52. 1920. (T: Morong 255!) Stylosantbes montevidensis £. arenosa Hass. in Fedde, Rep. Sp. Nov. 16:223. 1919, ex char. Stems ligneous near the base, erect, branched, to 2 dm. tall, with soft white pubescence and some scattered tuberculate bristles. Leaflets narrowly elliptic, mucronate, acute, nearly glabrous above but with scattered bristles beneath, with 3—4 pairs of conspicuous veins forming a submarginal nerve; terminal leaflet to 14 mm. long and 2.5 mm. broad; petioles 3-5 mm. long, with scattered short bristles, the rhachis 1 mm. long or less; sheath of the stipules 4.5—5.5 mm. long, with scattered bristles, 11- to 15-nerved, the subulate teeth 3—4 mm. long. Spikes oblongoid to ovoid, 2- to 4-flowered; bracts unifoliolate, the leaflet 4-7 mm. long, [Vor. 44 340 ANNALS OF THE MISSOURI BOTANICAL GARDEN was y RC CHECA \ Se ae ( eS -. Les m [| $ 1099 2009 3000 RILOMETERS ` Fig. 8 the sheath 2.5-3.5 mm. long, 4.0 mm. broad, with scattered bristles and about 9 nerves, the teeth 3.5—4.0 mm. long; outer bracteole 3.0-3.5 mm. long, ciliate at the apex; axis rudiment none; inner bracteole 1, 1.0—2.5 mm. long, ciliate at the apex. Calyx tube 2.5—3.5 mm. long, the lobes 2-3 mm. long and ciliate. Standard suborbiculate, 4-5 mm. long; wings 3.5—4.5 mm. long, obovate, auriculate at the base and shortly appendaged within; keel petals 3—4 mm. long, subrostrate. Loment 2.0-2.5 mm. broad, weakly nerved; only the upper articulation of the loment fertile, 3.5-5.0 mm. long, glabrous except for some short hairs on the margin near the summit; beak strongly uncinate to nearly coiled, 1.5-2.5 mm. long, less than half as long as the upper articulation. The type was distributed as S. &uyanensis although Morong noted on the col- lection label that it was not quite like other specimens of S. guyanensis which he had collected. Hassler named this species as a form of S. montevidensis. PARAGUAY: ASUNCIÓN: exact locality unknown, NY). WITHOUT PRECISE LOCALITY: Hassler 7606 (GH, NY) Morong 255 (US, MO, GH, 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 341 22. STYLOSANTHES VISCOSA Sw. in Prod. Veg. Ind. Occ. 108. 1788, ex char. Hedysarum bamatum B L. Syst. Nat. 10:1170. 1759, ex char. Stylosantbes glutinosa HBK. Nov. Gen. et Sp. 6:507. 1823, ex Stylosanthes viscosa B acutifolia Benth. in Mart. Fl. Bras. 15i: 91, 1859, ex char. Stylosanthes viscosa f. typica Hass. in Fedde, Rep. Sp. Nov. 16:220. 1919, ex char. Stylosanthes prostrata M. E. Jones, in Contrib. West. Bot. 15:135. 1929. (T: Jones 24002!) Stems ascending and spreading or prostrate and matted, much branched, to 1 m. long, densely pubescent with short viscid hairs. Leaflets to 25 mm. long, 5 mm. broad, usually much smaller, acute or obtuse, punctate beneath, shortly hairy or hispidulous, with 2-4 pairs of conspicuous veins; petioles 2.5-5.0 mm. long, hispidulous, viscid, the rhachis 1-2 mm. long; sheath of the stipules 3.5—5.5 mm. long, 1-2 mm. longer than the teeth, hispidulous and viscid on the back, 3- to 5-nerved. Spikes small, crowded, ovoid, 2-to 5-flowered; outer bracts usually trifoliolate, inner unifoliolate, the sheath equaling or slightly exceeding the teeth, hispidulous and viscid on the back, 5- to 7-nerved; outer bracteole 1, 2.5—3.0 mm. long, ciliate at the apex; axis rudiment none; inner bracteole 1, 2.5—3.0 mm. long, ciliate at the apex. Calyx tube 3-7 mm. long, glabrous to puberulent. Standard suborbiculate, 4-7 mm. long; wings 4-5 mm. long, auriculate, spurred within at the base; keel petals 3-4 mm. long, falcate. Loment to 2.5 mm. broad; only the upper articulation of the loment fertile, 2-4 mm. long, shortly hairy, reticulate- nerved; beak short, less than half as long as the upper articulation, usually about one-third to one-fourth as long, shortly hairy, strongly uncinate, often coiled. Most specimens possess subacute or obtuse leaflets 5-15 mm. long. Some have acuminate leaflets 15—25 mm. long, and these have been segregated as var. acuti- folia. This variation occurs throughout the entire range of the species, however, although most of the specimens with longer leaflets are found in the southern part of the range Stylosanthes prostrata of Baja California is merely a heavily viscid, prostrate form of S. viscosa. Stylosanthes viscosa has been reported several times as an adventive along the western coast of Africa. CAMAGUEY: Camagiiey to Santa Ana, Britton 1960 (NY, US); south of Sa SEH , Shafer E (NY). HABANA: pee m Baker 2930 (NY) ; Gran Sierra, Clemen 2078 (NY); Loma de las Yatas, Ekman 574 (US); Loma de la Juta, W Wilson & León 11643 of Sancti Spiritus, Shafer 12120 (GH, NY, MO, US, F); 12 km. east of Cascajal, Howar 5578 (GH), 65534 (GH, NY); Cienfuegos, Combs 407 (GH, US, NY, MO, F); La Mag- ena, Cayamas, Baker 2503 (NY, US); e Jack 8713 (NY : El Cobre, Britton, Cowell & Shafer 12866 (NY, M US); Guantanamo Bay, Britton 1960 (NY, US); Ge Vue unknown, Wrigbt 122 GH. NY, MO). WITHOUT PRECISE LOCALITY: Linden s. 2 o 1i. Z = Domin REPUBLIC: bie El Rubio, Jimenez 1058 (US). SANTO DOMINGO: Monte Crist ee Pes 12633 (US Two-Mile-W ood, KS Catherine, Harris e f , NY, MO, US, F); Lititz TUE, Harris 12899 (GH, NY, MO, US, F); Abbey G e. Orcutt 2 2869 (US); Kingston, Hansen 1897 (NY); Spring Hill, P. Portland, Harris 6630 (NY, US, F). [Vor. 44 342 ANNALS OF THE MISSOURI BOTANICAL GARDEN EL Sis near Chalchuapa, Calderón 1008 (US, GH). Gua JAL etween Monjos and Jalapa, Steyermark 32209 (F). JUTIAPA: Cerro Colorado, Standley 7618 3 (F). GUA: su of ofge Standley t$ Chacón 5385, 5403 (F); geng Standley pose (F, US). EL cavo: exact locality unknown, Gentle 2245 mies Swallen 10793 (US); west oi Chaquito, near Río Ye- geb Standley eme? (US); San Antonio, Rodríguez 667 (F). EL paraiso: Quebrada de Dantas, Standley, Williams & Molina 1289 (F); Yuscaran, Molina 1658 (F). O: BAJA CALIFORNIA: San José del Cabo, Jones 24002 (GH, NY, MO, US, F), 27236 — US); Todos Sege Jones 24769 (MO, NY, Pc Cape San Lucas, Rose 16364 (NY, US). cHiAPAs: Hacienda Monserrate, Purpus 9126 (US); Cen Ge Tapana, Oaxaca, and Tomala, Nelson 2045 (US), 2950 (US). corma: east side of Manzanillo Bay, Ferris 6199 (US). GUERRERO: 4 miles north of Acapulco, Barkley, Rowell t$ Westlund 30 (F); Acapulco, Palmer 10 (GH, . JALISCO: San Sebastian, northeast of Hacienda del Cura, Mexia 1373 (GH, MO). srnatoa: Río sin p bt ee 1525 (US); Santa Lucia Con- cordia, Debesa 1632 (US). sonora: Alamos, Ros andley & Rassi 12693 (GH, US, F); Alamos, Río Puerte, Gentry 2034 (GH, MO, U CM TABASCO: Achotal, Balancán, Matuda d ug (F, NY), 3047 (NY, F). zacatecas: near San Juan Capistrano, Rose 2467 (GH, US). Nicanacua: Volcan El Viejo, Oersted 4740 (F). PANAMA: PANAMA: Vicinity of San Carlos, Allen 1143 (GH, US, MO, IVI NY). HABAM Ayo = Sailopata, Cardena 3186 (F). WITHOUT PRECISE LOCALITY: São Yungas, Bichtien o (US); Urupana, White 942 (NY). BRAziL: Bafa: Mt. Peludo, cone 391 (GH, US, F); Machado Portela, Rose & Russell 19956 (US); Ilha de Itaparica, Pires 3421 (NY); Busca Vida, Bondar 2598 (F). DISTRITO FEDERAL: Ip a, Barreto 5 MINAS GERAES: ance, Pirapora, Cochran s. n. (US co, Vautbier 163 (GH); Ilheos, Riedel 211-212 ia i PERNAMBUCO: Prazeres, Pickel 980 (GH, US, NY) ; Tapera, "Pickel 186 (F). DE JANEIRO: near Rio de no des asd aber (GH), ATH, 174 Se » NY); Copacabana, Geier 5817 (NY). SANTA CA eitz (US). são PAULO: Campiña, Krug & Zagatto 2i (0$); peso cars 2523 ). BriTIsH Guiana: Berbice E Demerara) Co., Abrabam 89 (NY); t lacelicr oa known, Schombarek 178 E Y). ) abam 59 (NY); exact locality COLOMBIA: CUNDINAMARCA: Melgar, äech 3608 (GH, US, NY, F). HUILA: east of Neiva, Rusby & Pennell ue 3 (GH, US, NY), 1 NY). MAGDALENA: Rincón — Allen 405 (MO). SANTANDER: Le brija, der 76 (US). toma: Doima, Haught 2444 (MO, US, F, NY); dese Pennell 3608 ( GH. Us, NY, M PARAGUAY: LA CORDILLERA: Ypacaray, Hassler 11576 (GH NY, MO). WITHOUT SC LOCALITY: nee Río An. and Río Aquidilun-mí, S pipe (GH, SurtNAM: Zanderij Island, Maguire & Stahel 2 NY, US, F); Sectie Oeest, Stabel E Guess 523 = eria River, Rom m aa 4 xe a: BOLÍVAR: Island of Ma argarita, os quito, Johnston 249 (GH); een Ciudad Bolivar -i 5 ers Killip 37234 (US); ng J Upata d is p Semet [7563 (M NUEVA ESPARTA: Cocke Island, dae 6 (US, GH). a de ie near Valera, Pittier 10721 (GH, US, N 23. SrYLosANTHES figueroae Mohlenbrock, sp. nov. Herba humistrata vel suberecta basi ramosissima ad 20 cm. longa caulibus glabris vel puberulis. Foliola elliptico-ovata puberula nervis 3—5-gemmatis; foli- olo terminali ad 10 mm. longo 5 mm. lato; foliolis lateralibus ad 10 mm. longis; petiolo 1-3 mm. longo glabro vel puberulo; rhachide 0.5—1.5 mm. longa; stipulae striatae vagina 3-7 mm. longa glabra vel puberula processibus subulatis mucronatis 3-5 mm. longis. Spicae parvae oblongoideae floribus 2—7; bracteis plerumque uni- 1957] > MOHLENBROCK—REVISION OF STYLOSANTHES 343 foliolatis stipuliformibus vagina puberula 3—6 mm. longa nervis 3-7; bracteola exteriore 1, 2.5—4.0 mm. longa apice ciliata; axis rudimento nullo; bracteola in- teriore 1, 2.0-3.5 mm. longa apice ciliata. Calycis tubus 3.5-5.5 mm. longus glaber lobis 1.5—2.5 mm. longis. Corolla lutea; vexillo suborbiculato 4.0—5.5 mm. longo; alis falcatis 3.5—4.5 mm. longis; carina 3.5—4.5 mm. longa. Lomentum pubescens circa 2 mm. latum reticulatum; articulo superiore 2.5-3.0 mm. longo puberulo; articulo inferiore abortivo; rostro 2.0-2.5 mm. longo uncinato puberulo. Stylosanthes figueroae is distinguished from other members of its section by the pubescent loments with only one fertile articulation. It most nearly resembles S. humilis from which it differs in having the beak of the loment one-half to two- thirds the length of the upper articulation, rather than two to six times as long. This species is known only from the vicinity of Cali, Colombia. CoLoMBIA: cauca: Cali, Figueroa 897 (US, HOLOTYPE); Cauca, Triana s. n. (NY). 24. STYLOSANTHES MONTEVIDENSIS Vog. in Linnaea 12:67. 1838. (T: Sellow s.n.!) Stylosanthes juncea Micheli, in Mem. Soc. Phys. Genève 28:7. 1883, ex char. Stylosantbes juncea var. setosa Chod. & Hass. in Bull. Herb. Boiss., ser. 2, 4:884. 1904, ex char Stylosanthes montevidensis var. juncea (Mich.) Hass. in Fedde, Rep. Sp. Nov. 16:223. 1919. Stylosanthes montevidensis f. glabrata Hass. in Fedde, loc. cit. 1919. (T: Hassler 2709!) Stylosanthes montevidensis f. setosa Hass. in Fedde, loc. cit. 1919, ex char. Stylosanthes montevidensis f. typica Hass. in Fedde, loc. cit. 1919, ex char. Stylosanthes linearis Blake, in Jour. Wash. Acad. Sci. 14:287. 1924. (T: Jörgensen 2693!) Stylosanthes gracilis var. rostrata Burkart, in Darwiniana 3:251. 1939, ex char. Stylosanthes montevidensis var. heterophylla Burk. loc. cit. 260. 1939, ex char. (T: Du- sén s. n. Astyposantbes montevidensis (Vog.) Herter, in Rev. Sudamer. Bot. 7:210. 1943. Stems usually erect, to 1 m. tall, slender to robust, simple or often sparsely branched, glabrous to puberulent above or rarely throughout, usually with scattered long bristles. Leaflets lanceolate to linear, acute and mucronulate, glabrous to shortly hairy or occasionally with bristles along the costa beneath, with 2—4 pairs of conspicuous veins forming a submarginal nerve; terminal leaflet to 4 cm. long and 3 mm. broad, the lateral ones to 3 cm. long and 2.5 mm. broad; petioles 2.5— 7.0 mm. long, glabrous, hispidulous, or short-bristly, the rhachis about 1 mm. long; sheath of the stipules 5-11 mm. long, sparsely bristly or densely short-hairy or rarely glabrous and with 13-21 often conspicuous nerves, the subulate teeth 3.0- 5 mm. long, averaging about 4 mm. shorter than the sheath. Spikes globose, very small and few-flowered to large and over 40-flowered; bracts unifoliolate, the leaflet 2-5 mm. long, the sheath 2.5—4.5 mm. long, 2.5-5.0 mm. broad, ciliate along the margins or occasionally with a few bristles near the apex, usually 7- nerved, the teeth 1—4 mm. long; outer bracteole 1, 3—4 mm. long, ciliate at the apex; axis rudiment none; inner bracteole 1, 2-3 mm. long, often bifid, ciliate at the apex. Calyx tube 2-5 mm. long, the lobes 2-3 mm. long. Standard suborbiculate, 4.0— 6.5 mm. long; wings 3.5-5.0 mm. long, auriculate at the base; keel petals 3.5—4.5 : [Vor. 44 344 ANNALS OF THE MISSOURI BOTANICAL GARDEN mm. long, falcate. Loment 2-3 mm. broad, reticulate-nerved; only the upper articulation fertile, 3.0—5.5 mm. long, densely short hairy, at least near the apex; beak strongly uncinate, densely short hairy, 1.5-2.5 mm. long, one-half to one- third as long as the upper articulation. The variability of the species is attested to by the many varieties and forms which have been proposed. Many of these are based on tenuous characters, and there is no line of demarcation separating any of them. Micheli proposed S. juncea as a new species on the basis of the stem usually 4 mm. or more in diameter as opposed to 2 mm. for S. montevidensis. Hassler divided the S. montevidensis complex into two forms of "typical" S. montevidensis and two forms of var. juncea, recogniz- ing the latter as a variety, even though stating that it was merely a robust form. Blake described S. linearis in $ srYLosaNTHES because of the presence of an axis rudiment 0.3—0.8 mm. long. This character was not discovered in the present study of the type. Blake also noted the presence of 2 bractlets (— inner bracte- oles), but there is actually only 1 deeply bifid inner bracteole. The size of the inflorescence is exceedingly variable, specimens at one extreme having spikes 2- to 5-flowered while those at the other extreme have spikes over 40-flowered. Stylosanthes montevidensis is recognized from other members of its section by the one fertile pubescent articulation with the circinate beak about one-half as long as the articulation. A specimen collected by Calderón from El Salvador in all probability is adven- tive although the species does extend from Argentina to Colombia. Er Satvapor: near Chalchuapa, Calderón 1051 (NY, US). ENTINA: CHACO: — Margarita Belen, Aguilar 1163 (MO); Las Palmas, Iris 2693 ee US, MO). cónmpoBa: Sierra Chico, Lossen 173 (GH, MO, F). CORRIENTES: San Martin, La Cruz, Ibarrola 1770 (US); Estancia, Santa Teresa, Pedersen nt (US); Em a ri Pedersen 908 (US). cruz ALTA: Cam pi Alegre, Schreiter 4078 (F). JUJUY: E enda de Chañi, Schreiter 11083 (GH). MISIONES: cm , I9II ( pas, mania, Venturi 9957 (CH, US, MO). SANTA FE: Reconquista F.C.S.F Ae 5883 (GI TUC ei io Burru — Venturi 2616 (GH, US) Las Cejas, Venturi 2117 (US); pa Alegra, Venturi s. n. (U WITHOUT PRECISE LOCALITY: Cafi, Cancillo, Venturi pons US); La ër? ranja, Ekman po palis US, MO). PAZ: La Paz, Bang s. n. (US). SANT Sara, B , Stein- bach Lee (GH, oun E 5243 (GH), 3219 (B), 6614 h Sa nrc. Serrinka, Dusén pei I GH, 13718 GA Ms ). RIO GRANDE DO A 6 F). A CATARINA: ollector Unknown C1397 (GH). são ie ore Capital, Pickel 5499 (US). eieae PRECISE LOCALITY: 7 ) META: San Cal Dryander 3038 (US). RAGU NTRAL: Caballero, Morong NY, US, M NTE HAYES: suse Elisa, Hasle 2709 (GH, MO). ere Ke CISE EPI Primera, Jórgensen 4817 (NY, US, MO, F); Coro de Altos, Hassler 1798 (NY), 1799 (NY). RUGUAY: ALONES: Atlantida, Zosen B2830 (NY). cerro LARGO: Río SE Palleros, Rosengurtt B79 (NY). ÓNIA: Arroyo de Pintos, Artilleros, near erto Platero, Bartlett 21194 (GH). SE ÓN: Guinta del Colegio, Lorentz s.n. (US). sarro: San Antonio, Osten 5375 (US); San Antonio, "Resengerti eg (GH, F). 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 345 ibis ARR Nea (ah L LS Six a de e hj] eA ke / : SCALE 3 e.. s 1000 1500 2000 wiLES Se: |, a i Fa 2000 3000 KILOMETERS 2 s ` ` ` Fig. 9 25. STYLOSANTHES HUMILIS HBK. Nov. Gen. et Sp. 6:506. 1823, ex char. Stylosantbes sundaica Taub. in Verh. Bot. Brand. 32:21. 1890, ex char. Astyposantbes humilis (HBK.) Herter, in Rev. Sudamer. Bot. 7:209. 1943. Stems usually ascending or sometimes prostrate, herbaceous to subligneous at the base, to 0.5 m. tall, branched, usually with short white hairs along one side of the stem, often with scattered short bristles, rarely nearly glabrous except near the base, the pubescence usually more dense immediately beneath each node. Leaflets lanceolate or sometimes elliptic, often mucronate, acute, short bristly-ciliate to nearly glabrous, with 3—4 pairs or usually conspicuous veins; terminal leaflet to 15 mm. long (rarely to 30 mm. long) and 3.5 mm. broad, the lateral ones to 11.5 mm. long and 3.0 mm. broad; petioles 2.5—5.0 mm. long, shortly hairy and often with scattered bristles, the rhachis 1-2 mm. long; sheath of the stipules 3—5 mm. long, [Vor. 44 346 ANNALS OF THE MISSOURI BOTANICAL GARDEN short-bristly and with 5—7 nerves, the teeth 2.5—3.5 mm. long, averaging about 1 mm. shorter than the sheath. Spikes short, ovoid, crowded, 3- to 4-flowered; bracts 1- to 3-foliolate, the sheath 2.5—4.5 mm. broad, nearly always bristly, rarely only ciliate, with 5—9 nerves, the teeth 2-3 mm. long; outer bracteole 2.5—3.0 mm. long, ciliate at the apex; axis rudiment none; inner bracteole 1, 2.0-2.5 mm. long, ciliate at the apex. Calyx tube 4-5 mm. long, the acute lobes 1.5 mm. long. Standard suborbiculate, spurred within at the base, 3-4 mm. long; wings 3-4 mm. long, clawed and auriculate at the base; keel petals 3.0—3.5 mm. long, falcate. Loment 1.5-2.5 mm. broad, reticulate-nerved; only the upper articulation fertile, 1.5-2.5 (—4.0) mm. long, usually only puberulent, rarely pilose; beak strongly uncinate to coiled, 1.5—3.5 (—5.5) mm. long, equaling to greatly surpassing the upper articulation. Stylosantbes bumilis is variable in size and shape of the leaflets idi in size of the loments. Most specimens have leaflets less than 1 cm. long; others, to 3 cm. long. The size of the loment including the beak ranges from 3.5 to 9.0 mm. long. These differences are not correlated geographically. The species ranges from central Mexico through Guatemala and Panama to Colombia, Venezuela, and Brazil; also occurs in the Antilles and is adventive in Malaysia and Australia. Specimens from Malaysia described by Taubert as S. sundaica are actually S. humilis. Taubert included S. sundaica in $ sTYLOSANTHES but with the axis rudiment “most caducous.” In the specimens examined, no trace of an axis rudiment could be found. CuBA: CAMAGÜEY: La Gloria, Shafer 239 (NY). Las vitLAs: Lomas de Bafiao, Luna 730 (NY). EL SALVADOR: geen of San Salvador, Standley 19399 (GH, US, NY). Gua : near Chiquimula, Standley 3735 gr ae SE sermar 1 (US F). o Guacalate, fene 58237 (N F). ATEMALA: ES ~ Fiscal, ‘Standley 59788 (F). APA: vicinity id Dr Standley Mer (F). juTI- between Jutiapa and Plan de ye Standley 75455, 74910, 75652 (F); between rat Blanca and Amatillo, Steyermark 30457 (F), 3045 O (F). QUICHE: exact locality unkn Aguilar 1407 (F). SANTA Rosa: plains Li Entero, southeast of Chiqui- mulilla, Standley 78796 (F); south of Goce? Standley 79431 i near Cuilapa, Standley 77942 (F); Chupadero, Heyde & Lux 4162 (GH, US). APA: vicinity of m Sondi) 74588, 73589 (F); between La Fragna and Zeien prise po RAS: MORAZÁN: Río Yeguare, yaw & Molina 16967 (GH, US, F); Agua PARAÍSO: exact ec? unkn St ob Be ghee FN "Seel gy ee : own, Stan I , Stondley 55048 (US ey I3 (NY, U S, F). yoro: near Progreso MEXICO: CHIAPAS: ue Nos 9138 (US). corma: Colima, Palmer 108 (US). GUERRERO: Cutzamala, Cayuca, Hinton 6985 (NY, US, F); Acapulco, Palmer 25 (GH, : mo 4447 (MO, GH, F); Barranca de Portillo, MÉXICO: Temascaltepec, Tejupilco, Hinton 1932 (US), sg (GH), 4767 (GH, US). wicHoacáN: Apatzingán, Apuli, Hinton 5289 (US, ). OAXACA: San Gerónimo, Mell 2117 (NY). CRUZ: Acasonica, Purfus 8874 (Gti, US, US, Sea Zacuapán, Purpus 2331 (GH, US, NY, MO, F); Minatitlan, J. vi 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 347 NAL ZONE: Corozal, € 27403 (US); between Fort SE and eee "Standley 20191 (US). cocrÉ: Aguadulce, Pittier 48 37 (US), 4983 (GH, NY, US). PANAMA: Las Sabanas, Standley Ae (US); near Panamá, Standley 26704, 27730 (US); near Punta Paitilla, Standley 26293 (US); between Panama and Chépo, Dodge, ee Sire k & Allen 16700 (MO). vERAGUAS: west of Sof. Allen 1054 (GH, US, M seas BRA Baia: (data not legible) (MO). CEARÁ: Acude São Bento, Maranguape, eeng pe (GH, NY, MO, US, F). Pará: Boim, Rio Tapajoz, Kuhlmann 18223 (US). ERNAMBUCO: Tapera, Pickel 3036 (GH, US, F). RIO GRANDE DO NORTE: Angicos, Swallen 4737 (US). COLOMBIA: ANTIOQUIA: exact locality unknown, Gutierrez, Klevens t$ Barkley 1463 (US). ELA: GUARico: between La Encrucijada and Misión Abajo, Pittier 14925 (US); Mesa de El Sombrero, Pittier 12489 (MO). Mera: near Tovar, Fendler 1793 (GH, MO). TRUJILLO: Dividine, [pes 10823 (US). zuria: Maracaibo, M 10698 US). enee QUEENSLAND: Townsville, White 8818 (NY, US) LAYSIA: BALI: Wed cer. deVoogd 2061 (GH). NOESA PENIDA: Sampelan near Tangled, deVoogd 2397 (GH). Java: Horsfield or (GH). EXCLUDED OR DUBIOUS SPECIES Stylosanthes rigida Spreng. Syst. 3:310. 1826. This species is described by a single phrase which could well apply to a number of species of Stylosanthes. Stylosanthes tenuifolia G. Don, Gen. Syst. 2:281. 1832. Don distinguishes this species, which is known from the “Isle of Maranham, Brazil,” by its narrow leaflets. While this likely applies to S. angustifolia, it could apply to narrow- leaflet forms of S. guyanensis. Stylosanthes suborbiculata Chiov. Ann. di Bot. 13:381. 1915. The type for this species (Paoli 239), collected in 1913 in Somaliland, has not been seen and is therefore excluded from the systematic treatment. Chiovenda describes it as with- out an axis rudiment which distinguishes it from all other African species of Stylo- santhes except S. viscosa which is adventive along the western coast of Africa. A photograph of S. suborbiculata may be found in Result. Sc. Miss. Stefan.-Paoli Somal. Ital. #. 8 (1916) but this is of little value in determining the exact status of the species. Stylosanthes rupestris Stuck. ex Seckt, Rev. Univ. Nac. Córdoba 17?-75:145. 1930. This is a nomen nudum with only the location given by the author—Sierra (Malaguena), Argentina. 348 [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN INDEX TO ExsICCATAE talicized numbers refer to collectors’ numbers, s.n. ( date parenthetical numerals refer to the numbers assigned to the species in this aph. monogra sine numero) to unnumbered Abbott, W. L. 142, 937a, 1767 (13). Abraham, A. A. 89 E ue gredo x E ). Allard, H. A. 13087, 13107, 13114, 13313, 13699, 13725, 14007, 14281, 14350 (13). Allen, C. 405 i. 688 (16). Allen, P i. I054 (25), 1143 16). 2237, 2412 (7); 2302, 3015 (16). Arbeláez, E. P. & J. Cuatrecasas. 5753, 6018 qe; ge a2; 8328 (6). ie? : 644, 4084, 4717 (16); (1 dë En 2816 (22 Pardi. W. A. & M. Barreto. E 5 SS: — W.A.& riri Gehrt. 105 (2 riste-Joseph, Bro. A994 (1 phis E. 5086, 5610, 5650 (8). Baikie, W. B. s.s. (9). nen L.H.& E. 2 Bailey. 97,1349 (16 Mi C. F. 05, 2930 (22). : asselbring, 7202 (22). 4557 (1 Lu ; 1459 (16); 2150 (24); Baratini, E $. n. y Barkley, F. A., * 3 Paxson & G. L. Web- ster. 2569 (16 "i ap DR de Rowell & B. L. West- 2). Barle, J. 42 (9). Barnes, C. R. & W. J. G. Land. 240 (25). Barreto, M. 1713, 5495, qd 5770, 5771, $994 (16); SSC ): 5773 (6). 342 Barter, C Bartlett, 10612 (4); 11250, 11706, 11836 des '11405, 21194 (22). aur, G. 89, 90, 92 (8 Beardslee, Cae A 'Kofoid. s. n. (15). Bels, 43 (5). Bernid É. A. 20 (6); 21 (16); 24 (7). Bermüdez, L. A. & F. A. Barkley, 17Có45 7565, 7691, s. "n. (16); 7792 Blanchet, J. 1831 (22); B nm. (16). Blankenship, J. W. 13). . 2598 (22) Bonpland, A. s.n. (22) Borell, J: R. 2ór (13) ox, H. E. 900, 947 (13). , 444, 3532, 3934, 4635, 4678, 5245, EN ogee Brandegee, T. S. 22}; Brenes, A. M. pe (6); 5775, 14326 (16). Britton, E. G. 3336 (7). Tainio; Ni L 93, IQOI, 1941 ( 13); 1960, 2417 (22). Britton, N. L. & E. G. Britton. 208 (13). Britton, N. L., E. G. Britton & A. Brown. 4506, 5189, oe Un 1756 a». Cedo 2). Britton, N. L. & J. F. Cowell. 729 (13). Britton, N. L., J. F. Cowell & J. A. Shafer. 12866 (22). Britton, N. L. & W. C. Fishlock. 1007 1 (13). Britton, 2: L., T. E. Hazen & W. E. Broad- f 16 E & C. F. Millspaugh. 2454 Britton, N. L. & J. A. Shafer. 3013 (13 Broadway, Y. E. (19); 709 (16). Brown, S. xp SN Bruc V" $5 (20 Büchtien, O. 210 (a): ; 1787, 4105, 5423; 8100, 8101 Ale Burchell, W 4, 4109 ru 4439; 74 6); pA (6); 9004 (17 5883 art, (24); 6743 (20); p 7397 (16). 232, 667, s. n. (13); 231, 672, 972 1957] MOHLENBROCK—REVISION OF STYLOSANTHES Bush, B. F. 1048, 3265, 6271, 15604, s. n. (15). Calderón, S. foot (22); 1051 (24); 1133, 1189, 1287 (16). Campbell, E. 6orọ (13). 338 (9). M. Alexander & L. Kellogg. 22 E Carvalho, A. 201 (16). Castaneda, R. 252 (13). Cattelan, A. 250, 253 (16). Chapman, A. W. 1898 (15). Chase, N. C. Cornman, 2132 (5). Correa, J. & P. A. Ga 16A200 (16). Cow ien = 5. TRE (16 Cowan, R. S. Maguire. 3 (16). Cuatrecasas, J epo 3824A, e 76534, 215, 16608, SC 38038 SC N. M. 397 (2 Curtis, A.G. bg (9). Curtiss, A. H. 9, r18 (13); 6458 (15). Daniel, Bro. 530, 2310, 4127 (16). Darko, K. O. 56 | 44 (22). > 495, 17463, 17682, 17794, 19347, 19618, 20059, 21195, 21270, 22- 360, 23340, 23350 15). Dodge, C. W., W., J. 835 (16). Do yle, C. B. SI (13); 144 (16). Prod F. 72 Ui an (25); 2366, pana (17); 409. (16 ander, E. 2112 (16); 2028 (24). 349 Ducke, A. 16182 Geer Dugan, A. 4008 (13 Dugand, A. & R. Seite 2864 (13); 3631 (16 ). Duncan, W. H. dig C (15). Duque, M. J. Se (16 eec P. 147 (22); 9083, 14781, 16381, r2 (16); 9147, 13718 (24); ds ' (2) Duss, P. 3017 (13). Dyer, J. J. & F. Verdoorn. 3426 (9). Earle, F. S. 9 (13 Earle, F. SS & C «LN TD. 2). gg IO, 209, 382, 1883, 1920, s. n. (13); 14708 (8). co S T 5). pmo EL nes DI 1742 eT 1743 (24); N6or3 (13); 16). Baier I 517 (8), 590 (16). Fairchild, D. m (13). Faris, J. A pe 33s Fawcett, v." 0 (13). Fendler, A. E A6), 1793 (25). Ferreyra, R. 5960 (8). Ferris, R. S. 6199 (22). Fiebrig, K. 392, 4384-4082 (22); 205, 4397, 4410 (16); 4453-4413 (10); 5028 Zeg 897 (23). Fishlock, W. C. 163 (13). Flores, R. S. 7,7 (3). - DEW. (15). Fo sets, T. R. 109414, 20518 (16); 22001 Fox, W.B. Hd 4034, 4080 (15). Fox, W. B. & R. K. Godfrey. 2559, 2600 (15). are A. 3054 (13). RL Ee (16); 23664 (17). Fu LM SIT,t m. ( Funck, N. & L. J. Schlim. 262 (16). Galan dec Dh Me Garcia, H. 4685, pera (16); B6328 (6). ardner G 2826 > 27 (22); 972, 2826 (16); 973 (6); 2093 (1); 2094, 6000 (17). Garroni, P. 62 (16). Gaumer, G. F. 908, 992, 1955, 1956, 2044 3). Gehriger, W. 358 (16). 350 Fas 25239 (22). 1050, 1883, 3898 (16); Men c (22). Gentry, H. S. 2934 Geib Glatfelter, N. M. s. 15). 1371 3); 5817 (22); 6509 Glaziou, A. Godfrey: R. K. 146, SC49090 (15). Goldman, E 57 (16). Goll, G. P 31 (13) oodman, G. J. 2342 (15) Grant, V. 756 (16). Graves, E. W. r Greenman, J. M. 3863 (15). Grisebac pate Ay PA OUS) UB). . J. Klevens & F. A. Bark- (25). Haenke, T. 1552 (16). Hammarlund, C. FE Hansen, O. 123 (13); 1697 (22). Harper, R. M. 740 (15). eec Ww. “630, 11764, 12406, 12733, 12 (22 ee W. 809 (22). Hartweg 958 Hassler, E. 4632 (2) ; 1798, 1799,2700, 5825 24); 1803, 2834, 2 006, 3082, 3330, 6014, 6454, 7030, 708, 7761, 9532, 12043 (16); 7575 (18); 7606 (21 Haught, O. F44, 134 (8); 2444 4434 avard, V. 2 (13). Heller, A. A. E (13); 4126, 14005 (15 Heller, C. H7 (16 bes Heriberto, Bro. 70 (13); 297 (25). ee F.L. 30404 t 8906, er (18); 85562 Heyde, E. T (16). Heyde, E. T. & E. Lux. - 3752, ; 6r20 (16); rn SE GEN Hinton, G. I, 1932, 4767, 5289, 6085 ( (25); 1j (16 A - Hitchc AS : EE n 127 die); r (2). Hoehne, F. A. Gehrt. 17665 (6). n. W. Gehriger. 48 (22 Holton, LE. 22 e 955 E ior — we 822 (3): 1018 (17); Set H. W. 3788, 4025 (15 — C. W. 56 i ag i gie R. A. See 4478 (22); 2788 [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN Howard, R. A. & E. S. Howard. 8413, 8795, 9293, 9659 (13). Howell, J. T. 8815, 9599, 9948 (8); I0- 88 (22 (22). Hunnewell, F. W. R 19781 (13). Hunnewell, F. W. & L. Griscom. 14324 (13). Hunter, A. A. & P. H. Allen, 75, 222, 753 16). Hutson, LR s. z. (13). Ibarrola, T. S. 2310 (12). Idrobo, J. M. & A. Fernández. 76, 1310 (16). Irwin, H. S. 330 (16). Jack, LG. 8713 (22 ldrobo e Z4 6). Jimenez, J. e J- KS (13); 1058 (22). A Johnson, P, 1 Ce 9 (22); 346 (16). Jones, M. E. i PLUS 27236 (22). Jórgensen, P. 2077 (2); 2693, 3628, 4203, 4817 (24). Kattah, e, J. A. Molina & F. A. 5 (13). 870 2); IIIj3.(17). 7257, 7556 (16). (3); 3I. (7); mam (6); 11563, 34509, 37138, 37217 (16); 37234 Killip, E. P. & J. Cuatrecasas. ip, E. P., J. Cuatrecasas & R 39200 (16). Killip, E. P., A. Dugand & R. Jaramillo. 3). 38164 (1 Killip, E. P. eg EC 3343 (16). hmann. V39798 Killip, E. P. (6). Killip, E. P. & A. C. Smith. 14067 (13); um ipn 30423 (17). Kings, V. GC69, LC75 (13). ee 4638 (16). SH. & O; Zagatto. 2191 (22). pes B. A. 10536 (16). Kuhlmann, J. G. 18223 (25). Kuntze, O- 148 (13); jer (16); s. n. (6, Se og Lf. (6). ryander. Mol. E (20). Key: J. 18 ( Langlasse, É. 518 (25). Lankester, C. H. 307 (16). 1957] MOHLENBROCK—REVISION OF STYLOSANTHES. Lawrence, A. E. 332 (16). Leavenworth, W. C. 135 (4). F. C. BT362, 2855, 5958 (16). León, J. 115 (22); 186, 375, 1028 (16); 663 (135 GE (es Leonard, 863, u 3401, 7037, 3797, 5708, 04s, e Leonard, E. C. M. E 11224, 12925, 13448, aes (13). IIOOI, n, G. W. Linden, J. s.n. [o5 Linder, D. H. 97 (17). Lorentz, P. G. s.n. (24). Lossen, T. 314 (16) Lossen, V. 173 (24). Macs gp de 13323 (16). Luna 730 (25). Gg c. L. 6885 (22); 6922 (16). Lützelberg, P. von. 35 (22); 112 (1). McCree, J. 822 (15) McFarland, F. T. & W. A. Anderson. 24I (15). McVaugh, R. 14222 (16b). —— e p met 4993 (16). Tem 0, 1241 (2); 1670, 2226 eters K. we 347 (15): Madrana, D. s.n. (13). Maguire, B. & G. Scal 23733 (22). Malus. GO A. be = (24). Mandon, G. Marsh, E.G. 969 (d) D mire Ji ss Martius, P. von. 1123 (16); 1124 (6). Monde: E. 1:958 (16); 3041, 3047 (22). Maurice, Bro. 790 (16). 790 . 1956, 2230 (9). Mearns, E. A Mell, C. x AS (25); s. n. (16). Merrill, . Williams. 15713 (16). Mexia, Y. ML ee 4523 (16). Miers, J. 2 Miller, O. O. & f CR Johnston. 68 (13); (22). Miller, W. & L. Griscom. 66 (16). Mills paugh, c. F. i (6); 153, 391, 1077, C M. ee Moldenke, H. M. Molina, A. 1681, 4079 Molina, J. A. 61, ie. y (16); 376, 1658 (22). 351 Molina, J. A. & F. A. Barkley. Boo4ó6 (13). e och s. n. (13). raes, J. C. En 885 E (6). Morel, L 957». "uu €: T. s .(21); 399 (24); 399b Mueller, C. H. & M. T. Mueller 325 (4). eener W.C. 12417 (16). utis, J.C. 2345, 2076 1 4683 (16). Nash, G. V. 678 (13). Na ih, G.. V. & N. Taylor. 1069, I1I04, 1545, 3852 (13). elson, E. W. 2865, 2045, 2950 (22); 4075 (16). cifaro, H. Ni n. (7). Northrop, J. L k A. R. Northrop. 37 (13). Oersted, A. e fie E: Oliveira, J. O’Neill, 5 $640 (22 2865. (22); 4447 (5). Osten, C. 5375 (24); 6498 (18). 3137 (13); Palmer, E. 10 (22); 25, 108 (25); 490 (4); 1900 (16). Palmer, P H 6a, 19, 132, 461, 2253, 5888, 6438 (1 S M. & J. H. rt Aar (1). Pammel, L. uu s. n. Po: Pock M E” T. = Pennell, F. W. 1469, 1582, 3351, 3491, 5505, 5532, 5623 (16); ip. 6073, 6352, 3, 44463 (9). Pickel, D. B. 186, 980 (22); 2403, s.n. (16); 2481, 2556 (6); 3036 (25); 5499 Piper, C. V. 5129, 5130 (16); 6089 (13). 342 (16). ires, J. M. I (22); 4063 Pires, J. M. & C. Black. 1994 ittier, H. IIO, 1656, 2165, 4707, 5098, 10553, 10558, ^. nq 14472 (16); 4837, 4983, 10698, 10833, 12489, 14925 (25); 7319, 9679 a 254 (7); 10721 (22); Pis V. A. G72 Plowes, D. C. ad "iu Poeppig, E. 149 (17); Aë IQOI, 2930 (16). 352 Pohl, J. E. 336 (22); ; 507 (6); 949, 1997, 2039, 2107, s. n. (16 Pollard, C. L., E. Pale, & W. Palmer. 206 (3). Pringle, 5782, 6721 (11); 2253, 165 (4i fu (16); 4540, 5172, 11433 (16 Pulle, A. 139 (17); d à us, C. A. 416, 1 (16); 466, 9126, 9245 (22); 2331, $594, 9138 (25); 4928 (4); 7152 (10). Questel, A. 116, 217, 1459 (13). Quintas, F. 6 (9). Rambo, B. SI747 (18). Rapp, S. 4 (13 Raunkaier, C. 1137 (i3): Record, S. La (16). Regnell, A. F. 428, 430 (16); 429 (22); £75 (2 Reiss, R. 75a (16); s. n. (2). Reverchon, $E 3009, e ei (15). 16). Ricksec Ricksecker, Mrs. J. J 13). Gë f 1316, 1877, Riedel, 122,:2 ee 8 " DH 3b vids (22). Ried, L.& f Tes 162 (16). Ries, S. von. E Robertson, N. F. 32 15). 78, 509, 973, 3662 (10); 667 , 687, 756, 1062, 1678, 3762 (16); 1257 (25). Rojas, R. T. 47,192,217 (16) ombouts, H. E (22 Rorud, B. 144, 145 (8). Rose, J. N. 2467, Upon (22). Rose, J. N., W. R. Fitch & P. G. Russell. 4028 Rose, J. N. & w. Hough. 4584 (10). MU b N. & J. H. Painter. 7478, 7543 16b ose, J. N., J. H. Painter & J. S. Rose. EEO ll 6 ose, J. N. . G. Russell. 19956 (22); 20693 (16). Rose, J. N., P. C. Standley & P. G. Russell. 12693 (22). S B79, B1070, B2830 (24); e ; s: n. (17, 22). Rothrock, J. T. 140 (13). Rowlee, W. W. & H. E. Stork. 839 (16). Rudd, V. 33t (16). [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN Rusby, H. H. es 2324 (16). Rusby, H. H. . W. Pennell. 325 (1); 1042, 1083 = IIII (16). Russell, C. 5). Ruth, W. A. 338 (15). Rydberg; P. A. & R. Imler. 454 (15). Saer, J. 631 (12). St. Hilaire, A. de. s.n. (6). — E 529 16). npn P 52150 16 22); Tonio. A. 7339 (16). Samuels, J. A. 260 (22). Beer J. 380, 638 (6); 382, 502, 923 (16 Sergerit, F.H. E340. Schomburgk, R. J. 240, 381, s.m. (16); 530 ( Schott, H. Y. 16 (1 ier Schreiter, R. 4078, 11083 (24). Schunke, C. ab The "Ae (16). 16). 24). ee Je A. , 104, 1551, 2499 (13); 239 Geck 305, 12120, 12866 (22); 863 (3); 2 Sharp, A. E 45169 (16). Shreve, F. 7256 (22). Sieber, F. W. 37 (5). Sintenis, P. 620, 1092, 4756 EN Skutch, A. 1463, e 2462 (16 Small, J. K.& & J. J. Carter. 3382 in Small; J. K., C. A. Mosier & G. K. Small. 643 (3). Smith, A. C. 2376 (16) Smith, J. G (25). Smith, L. B. 6359, 7078 (16). Sneidern, K. von. 215, 1414, 2311, 4348 (16). Snodgrass, R. E. & E. Heller. ; I71, 450; 673, 767, 836 (8). eRe CITY: »G. & —. Gonggrüp. 523 (22). Standley, P. C. 1077, 12110, 12260, 14213; 14503, 14862, 14983, 21222, 28654 (10) 5 25126 (13); 1702a, 9534, 10082, 06, 18283 en Get 2 40. 56278, 57982, 58200, 59401; 65833; , 60232, 61723, 63572, 1957] MOHLENBROCK—REVISION OF STYLOSANTHES 72470, 74734, 75042, 7551I, 764 si Standley, 538 (22); 5400, 5866 (10); "397216124 (18) Standley, P. C e. 7405(16). Me iuge P. C. kx x em ga a. 4406 (16); 31 Standley, P. C. & E. Padilla. 3008, 3261 es Standley, P. C. & L. O. Williams. 483 (25); 1563, 1598, 1634 (10). ee O. Williams & A. Mo- t. . Wi : on (10); 2143 (16). K. etherford & G. D. Steinbach, J. 3219, 52 43, 6614, qd (24); 5397 (12); 5306, 6143, 6916 ( 16, 22 Stevens, E. P. Stevenson, J. A. 13). Stewart, 1695, (8). Steye ; J- A. H217, 143985, Maer 15373, 15948, 23879, 24436, 2 30457, 31372 (25); 32209 (22); 35442, 38653 do: 51654 (3); 56234 (6); 57584, 61783 (16). Stratton, R. Ej (155. Svenson, H. K. 151, 11250,11301,11400(8). Swallen, g R. 4715 (17); 4737 (25); 10793 (22). Tate, G. H. H See d I03I (16). omson, 3.9. (9). Thorncroft, J. 1913 (9). Tomas, Bro. 957, 1009 (16). Tomayo, e E 077 (13); 1503, 1846, 2736, 3 Tonduz, ^u Gees 4707; de (16). Toro, R. A. 257, 468, 565 (16). 353 Townsend, C. H. T. 7 (6). I 4323, 4327 ees n. (23). Trelease, W. 7,215 (15). Tü he ge H. von. 3337, 3711 (13). Tweedie, J. 11667 (22). 84 (16); 2529, 2804, Ule, E. 8761 (17). Underwood, L. M. & F. S. Earle. 1682(13). Underwood, L. M. & R. F. Griggs. 658(13). Valerio, J: 278, 1238 (16). Valeur, E. J. oy pr (13). SE 257 Velasco, L. V. Go Bang (16). Vélez, I. (13 Venturi, S. I 519, 2117, 2616, 5992, 9957 (24); 5136 ( Vidal, N. J. A 3659 (16). Viereck, H. W. 300 (4). Vogel, J. R. T. Vogl, C. óór gd, —. de. Ph Dk 1 (7). 2061, 2397 (25). Waby, J. R. 62 (13). arming, E. 100, s.m. (16); IOI (6); iw OQ Lois A. 5935 (8); 6215 (12); 457 White, C. 942 (22). White, C. T 8818 (25). Wight, A. E. 8I, 86, i Hi 814 (9). Williams, L. O. 10540 Williams, L. Y See 201 (6); 5646 5834, 5863, 6473 (1); 5832, 6084, 6146, 7103, 7122 cot 6850 (6). A. Molina. 11027, 16967 (25); 11261, 13220, 14133 (10). Williams, L & J. I. Moreira. 5219 (16). 28, 118, 167, 384, 612 (16); 98 (13). Wilson, G. 7359, 7637, 7821 (13); 8405 re Wilson, P. & Bro. León. 11643 (22). Woytkovski F. "3509 Wright, C. 122 (22); SE (13). Wright, C., C. C. Parry & H. Brumm mel. 77 (13). Xantus, L. J. 26 (22). Yeshoda, Mrs. K. 415 (9). 354 ANNALS OF THE MISSOURI BOTANICAL GARDEN INDEX TO NAMES [Vor. 44 n type Me accepted, preexisting names; italics, synonyms; bold face, names or ar conblisd tio Aracbis aprica Asty posanthes angustifolia montevidensis Hedysarum hamatum var. M Ononis deeg Stylosan an seier 2 angustissima aprica Sect. Astyposanthes bangii a elatior B guianensis var. hispidissima elatior B bispidissima erecta var. acuminata esi guineensis Sect. gr Ge ssp. dissi var. genuina f. esetosa g var. intermedia .. ssp. anomala . — var. longiseta ma var. hispidissima var. nudiuscula ilis hum bumilis ingrata juncea f. intermedia aal a a > SÉ D a f. typica var. heterophylla var. juncea var. longiseta cronata riparia f. ocbroleuca var. setifera ruellioides rupestris new 19571 MOHLENBROCK—REVISION OF STYLOSANTHES 355 scabra 2 BLS surinamensis 332 scoparia __ 2c 2 sympodialis 317 poo Es ^ ` ` tenuifolia . — 347 osa 318 tuberculata __ 316 suborbiculata EE A viscosa _ eso 40 12981 subserice des 320 B v EE ADIT SPART Y o Sect. Beete GAN THIS EE _ 308 f. LE Re EIEN esent rece E T a Sect. Styposantbes sec 308 rfl ‘foram gE PEO RS vol sundaica pt eee a EE See eee aS 332 GENERAL INDEX TO VOLUME XLIV New scientific names of plants and ES final members of new combinations are printed ha in bold-face type; synonyms and page italics; and all other matter, in ordinary E cgi aving reference to figures and plates, in A Adianthum cordifolium, 167; ternatum, 150 Ger tropical, Some lichens of, II. Usnea, Ase of genera in the acte in rela- n to their morphology, 10 tee andromedaefolius, ded 184; atro- purpureus, 144; bellus, 161; bracbyb- ferus, 162; Breweri, 135; cordaius, 166; flexuosus, 174; mucronatus, 158; myrt til- lifolius, 184; ornithopus, 158; Pringlei, Me. SS 168; subverticillatus, 50; ternifolius, 150 Aa 235; peru 244; foliosa, dee mesocbora, 245; occidentalis s, 245; parv ; qu uinqueflora, 244; poh aui 253; copelandi, 254; gur- S nd wegi, 260; heroes, 251; lemberti, 254; Memes zs mex microca Se 264; occidentalis, 245; SCH var. eier 237; wislizeni, 235; American wild hop, 271 Amstutz, Erika: A new genus of Iso- A new genus of the tribe e (C ae), 295 | PE DA pica £la Sa E 295, 207 Antidesma E 274 Antbopogon rella, i ete 213; Portus 230; brevifolium, 2 detonsa, 213; elegans, 2 inc e 225; macounii, 228; rum var. ton- . uniflorum, 228; procerus, 22321 thermalis, 217; ton- erc? aprica, 327; [eid 318 Aridity: and rapi evolution, 110; struc- tural trends in the Cleomoideae in rela- tion to, 8 Asplenium atropurpureu pel levet 306, ai pacers 336; humilis, 345; leiocer ta, 337; monteviden- sis, 343 B Bracts of hops, 282, 282; scar lengths of, 282, 284, 285, 280, 290; veins, 282, 282 C wem ri Androcalymma, a new of, ae GER Capparidaceae: Sadi in the, III. Evolu- tion mi phylogeny of the western North American Cleomoideae, 77; taxonomy of the 7 Capparis spinosa, 77 Cassebeera andromedaefolia, 179, var. gra- cilis, 179, var. pubescens, 179, var. rubens, atropurpurea, 144, v ar. minima, 141; 177; ternifolia, dud longimucronata, 155, var. mu 158, var. ornitbopus, 158, var. Sen Cassieae, Ge a new genus of the tribe, Cheilanthes GH He mucronata, 158, va igbtia bescens, 179; P MIA, Ti; Een 150 Chironia, 202 Chlora, Chromoso: bers: in species of Gen- tianella and related genera, 198; Pellaea Sect. Pellaea, 1 Cleom grapli a e 88, 96, 07, EN roo, 104; 81, 82, 9 stipules, SCH 88, CH phylogeny, 81; tax- — 77 sect. Peritoma, 77; isomeris, 78, 96; Q : platycarps, 7 siliculifera, Pes aids. 78, 358 Cleomella: distribution of, 89, 90, 96, 98, 99, IOI, 104, 106; PRSE prolate of, 77, 107; morphology, 78, of fruits, of i 87; of stipules, 87, 89; phylogeny, 81; onomy, Sedes Evolution and phylogeny of ; 77; age of genera in, 104, 111; dis- bai of the western North American genera of, 88, 90, 06; geographical evolu- of the, s 106, 107; morphology, 8, 81; taxonomy, 7 Climate, Pose den. and morphology of the Cleomoideae, 88 s, 275, morphology of, in vari- etal maigre ae Crawfurdia, 1 Crossopetalum, Cat gentianoides, 210 Crypteris divaricate, 179; pubescens, 179 Darinphytrum, 202 Dasystepbana CC DEE Davis, Edward L.: Mor phological complexes in hops (Humulus ere .) with spe- cial reference to the American race, 271 Denckea, 210; crinita, 223 Desert ide of the Cleomoideae, 91 Dicorynia Dodge, ced W.: at lichens of trop- ical Africa. IL Usne Dwyer, John D. “ain roc cibis: new genus XT " tribe Casies (Gaenlsitii- aceae), 2 E Ericala ey 253 4 Evolution and phylogeny of the western orth American Cleomoideae, 77, 104, 107 Environment, relation of, to morphology: of hops, 283; of the orbe 88 ouno relations hips of the western North American Cleomoideae, 114 F Fern up Pellaea, 125 Fr ne dps ZE series of the North Amer- an Cleomoideae, 80, 81, 82 [Vor. 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN G Gentiana, index to names, 267 Gentianella Moench, A revision of the North Weeer GE of, 195; chromo- some neric position, 196; geographic distribution, 206; history of, 2 01; in o , 267; ere es 204; taxonomi c ere ent, 208 Gentianella, 2 Geographic debo of the Cleomoideae, pe dl. relation of generic origin to, 97 Giles, John M : A revision of the North erican species of the genus Gentian- ills Moench, 195 H Halenia, 197 Hedysarum hamatum, 318, var. a, 324, . B, Hippion m 261; poor mee ue latum, 264; qui vinquefolium, Hop: the Ain wild um distri- bution of types, ; ëch Backa, 286, 289, Brewer’s Gold, 289, Fuggles, 2 eg 281, 288, Styrian, 290; Tettnang, Hops (Humulus lupulus L.), Morphological complexes in, with special reference axis 282, 284, 285; inflorescence, 275, 276; leaf, 279, 283, dentation, 279, pubes- cence, 281 Humulus, 271; aculeatus, 273; amarus, 273; americanus, 273, 274; communis, 273; cordifolius, 273; cordifolius, 273, 275; wire 275, 275, var. cordifolius, 275, va eomexicanus, 273, 274; scandens, 275, 274; volubilis, 273; vulgaris, 273 I Iltis, Hugh H.: Studies in the Capparida- ceae. IIL Evolution and phylogeny of the western North American Cleomo- ideae, 7 florescences in the Cleomoideae, 84, 84, > o DETUR a new genus of, 121 Isoetes oni. arbores 78 Ixanthus, 19 INDEX Jaeschkea, 197 E Leaves: in genera of Cleomoideae, 87; of 5 op varieties, scores of lobe and dentation mber Leian thus lanceolatus, Lupulus amarus, 273; scandens, 273, 274 M Martiusia, 2 Medicinal uses dot hops, 272 rtis Marphoingis I ie d in hops (Humulus lupulus L.) with special reference to the American race, 271 N North American Cleomoideae, and phylogeny of, 77 Nothochlaena andromedaefoli ia, 179; atro nM ine a, 144; cordata, 167; Keele, evolution O Ononis coriifolia, 314 Opsantha, 209 Orthoselection in genera of the Cleomoideae, Oxystylis: age of genus, 108, 111; SC of, 89, 90, 97, 99, ee morphology it, 80, 82, cone $5, 86, of stipules, 5 7, 88; phy- logeny, 81; taxonomy, 7 Oxystylis, 78; lutea, 78, do. 82 P ires = section Pellaea, A revision of the fern g ;129* cytology, 129; ecology, purea, wi SCH eM — 162; Pne Polrsofieitis; 125 359 Pteris andromedaefolia, 179; atropurpurea, 44, var. venosa, 144; ; spiculat 150; Gët 144; "triphylla, 150 Reyin, ei the fern genus Pellaea section Pellae Revision the North funt species of Gentianella ng I4 a of the genus Re 299 S Spiragyne, 2 Stipules, modifications of, in the Cleomoi- deae, 8 Studies in ge Capparidaceae, III. Evolu- tion and phylogeny of the western North American Gage té us of Isoetaceae, 121 stematic treatment, 306; uses, 305 Swertia, 197, 203 T Taxonomy: of the Cleomoideae, 77; of Gentianella, 208; of Pellaea section Pei. of Stylosanthes, laea, 153; on, Alice genus Pellaea section Pellaea, PME biflorum, 327; chic? 331, U , 1; acanthera, 63; aequatoriana, 27; strigos blepharoides e Re eri, pä escens, ameroonensis, 25; 50; P dier. 30; ceratina Se geen 360 4; cervicornis, 1; chloreoides, 67; chon droclada, i complanata, 11; co Kante ar. Mey 60; corrugata, 12; Dads 56; ge? e. ar. Biodenata, 27; dasy- poga var. Berg 23; dasypogoides rata, i ; di SE ls. 29; E ai 56; indigena, 73; isabellina, 55; ae 43; steinii, d livida issi 22; denn 52; E 8; perspinosa, 39; picta, 54; Pinkertonii, 24; bob Wislizenia: [fVor. 44, 1957] ANNALS OF THE MISSOURI BOTANICAL GARDEN 14; Promontorii, 65; residere 15, pulchella, 21; re ta, pul- 11; ek 47; xpo 50; trichodeoides, 75; tropica, 31; undulata, 37; usambarensis, 30; Vainioana, 20; KC es 11; vesi- bd 16 Ww age of genus, 108, 111; dis- tribution of, 89, 90, 97, 98, 99, 100; relationships, no X. morphology, 78, of fruit, inflorescence, 84, et 89, v Miele, '87, phylogeny, 8 Wislizenia, 78, 84; sin: 78,6 7. 89, var. 94, ar. palmeri, Zo, melilotoides, 83, var. refracta, Zo MISSOURI BOTANICAL GARDEN EDGAR ANDERSON, Curator of Useful Plants CanaOLL W. DODGE, Mycologist RoBERT E. Woopson, JR., Senior Taxonomist | PO N. ANDREWS, aleobotanist Tiron VON SCHRENK, Si Assistant Curator Museum of Economic Plants STAFF Acting Director Hoen C. CUTLER _Grorce B. Kan SCHAACK, Acti x of Herbarium JULIAN re STEYERMARE, Honorary Research Associate Jonn D. Dwyer, Research Associate NELL C HORNER, Bibliographer an and Editor - Oscar E. GLAESSNER, Controller | — .BOARD OF TRUSTEES President — Vice-President ` Danret K. CATLIN