Dodota JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor-in-Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR. IVAN MACKENZIE LAMB VOLUME 61 1959 The New England Botanical Club, Inc. Botanical Museum, Oxford St., Cambridge 38, Mass. ew REFERENCE LIBRARY MAR 3 0 1959 Dodora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor-in-Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR, IVAN MACKENZIE LAMB Vol. 6l January, 1959 No. 721 CONTENTS: Synopsis of the Genus Lepanthes in Mexico. Richard Evans Schultes and Gordon W. Dillon E 1 Notes on the Flora of a Gorge of Esopus Creek, New York. Heny h. Dunoan ERSTE re 21 Plants New to Illinois and to the Chicago Region. Julian A. Steyermark and Floyd A. Swink oe ME 24 A Milestone in Botanical Cartography [Review]. Askell Love D" een ev re id 28 The New England Botanical Club, Inc. Botanical Museum, Oxford St., Cambridge 38, Mass. RHODORA.—A monthly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid, in funds payable at par in United States currency in Boston; single copies (if available) 60 cents. Back vol- umes l-58, with a few incomplete, can be supplied at $5.00 per volume. Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes, relating directly or indirectly to the plants of North America, will be considered by the editorial com. mittee for publication. Articles concerned with systematic botany and cytotaxonomy in their broader implications are equally accept- able. All manuscripts should be double-spaced throughout. Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance, will be furnished at cost. Address manuscripts and proofs to Reed C. Rollins, Gray Herbarium, 22 Divinity Avenue, Cambridge 38, Mass. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill, Botanical Museum, Oxford Street, Cambridge 38, Mass. Application for transfer of Second Class Mail Privileges from Lancaster, Pa. to Boston, Mass. is Pending EATON PRESS, Inc. Specialists in Learned Publications 17 SPRING STREET WATERTOWN 72, MASS. Rhodora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB D ee ee re m Vol. 61 January, 1959 No. 721 SYNOPSIS OF THE GENUS LEPANTHES IN MEXICO RICHARD EvANS SCHULTES and Gorpon W. DILLON Prior to 1938, only four species of the orchidaceous genus Lepanthes were known to be native to Mexico. In 1938, Schultes described Lepanthes Rekoi from a Oaxacan collection and began a study of the numerous Mexican specimens of the genus in the Orchid Herbarium of Oakes Ames. The study was later carried out jointly by the two writers of this paper. Subsequent collec- tions referable to Lepanthes were made in Mexico by Dr. H. Emery Moore of the Bailey Hortorium and by Dr. Robert L. Dressler, now of the Missouri Botanical Garden; several un- described species were discovered in the Moore and Dressler collections and are herein described by Mr. Charles Schwein- furth and by Dr. Dressler. Many of the specimens of Mexican Lepanthes are from the valuable collection of the late Mr. Erick Ostlund of Colonia del Valle, D. F. and Cuernavaca, Mexico. Four new species were found in the Ostlund collection, and sundry sterile specimens also preserved in his herbarium (now a part of the Orchid Her- barium of Oakes Ames) indicate that botanical explorations in central and southern Mexico, especially in Guerrero, Oaxaca and Chiapas, may yet increase the number of Mexican species. In this paper, an enumeration of the Mexican representatives 2 Rhodora [VOL. 61 of Lepanthes is given, together with a key to the recognised species, of which there are sixteen. Nine are herein described as new. ‘Lhe material on which the paper is based is preserved in the Orchid Herbarium of Oakes Ames. Lepanthes tridentata Sw., a West Indian species, has been reported from Mexico, but we have omitted it in our enumera- tion, believing the report to have been based on the misidentifi- cation of a Galeotti specimen of Pleurothallis disticha (Linnaca 22 (18419) 820), a concept which, in this paper, is transferred. to Lepanthes. Lepanthes turialvae has sometimes been credited to the Mexi- can flora. This binomial refers to an extremely confused. con- cept, so confused that the rejection of the name as a nomen confusum has been suggested. (Schultes, R. E. in Rhodora 60 (1958) 97). No matter which of the several ways in which the binomial has been used be adopted, Lepanthes turialvae is defi- nitely not represented in our material from Mexico. Species of Lepanthes are found in both the Atlantic and the Pacific watersheds of Mexico in relatively high regions of the following states: Hidalgo, Mexico, Pueblo, Vera Cruz, Oaxaca, Guerrero and Chiapas. It is noteworthy that, so far as is known, all the Mexican species are endemic with the exception of Lepanthes acuminata Schltr., L. oreocharis Schltr. and J. Pris- tidis Rchb.f. In general, the Mexican species show relationships with the Middle American representatives of the genus. In some cases, however, West Indian affinities are apparent. It is very probable that additional species will still be dis- covered in Mexico. Lepanthes is well represented in Middle and South America and occurs throughout much of the West Indies. The number of species-concepts represented in Middle America is very large in comparison with the number in Mexico: in Costa Rica, for example, forty or more arc recognised. Ihe northern limit of the genus in Mexico appears to be the State of Hidalgo, where Moore collected Lepanthes avis and L. Moorei. 1959] Schultes and Dillon,—Genus Lepanthes 3 KEY TO THE MEXICAN SPECIES OF LEPANTHES 1. Peduncles conspicuously overtopping leaves. 2. Sepals conspicuously ciliattn ae. 2a. Sepals eciliate. 3. Leaves three or more times longer than wide 3a. Leaves suborbicular to ovate-elliptic. 4. Flowers large, almost equal to leaves (over 1.5 cm. long). Sepals, when spread, forming ellipse. . (16) L. tenuiloba. 4a. Flowers much smaller than leaves. Sepals, when spread, noticeably constricted at junction of dorsal (15) L. Schiedei. (9) L. oreocharis. with laterals. 5. Lip trilobate ....... nnn 5a. Lip bilobate. 6. Lobes of lip oblong and divaricate. 7. Petal-lobes very unequal, without an api- . (14) L. Rekoi. cule in sinus between them (3) L. avis. 7a. Petal-lobes subequal, with definite api- cule in sinus between them........ (4) L. congesta. 6a. Lobes of lip cuncate and incurved... (10 L. orizabensis. la. Peduncles shorter than leaves. 8. Lip trilobate (middle lobe small, often an apicule). 9. Petal-lobes conspicuously unequal. 10. Petal-lobes with an apicule in sinus between them. Anterior lobe of petal obliquely deltoid... . (2) L. ancylopetala. 10a. Petal-lobes without an apicule in sinus between them, Anterior lobe of petal semiorbicular....(6) L. Moorei. 9a. Petal-lobes subequal. II. Petal-lobes glabrous. Sepals ciliate (412) L. parvula. lla. Petal-lobes papillose, especially along margin. Sepals not ciliate. (11) L. papillipetala. 8a. Lip bilobate. 12. Sepals usually long-acuminate. Lateral sepals re- curved. 13. Petal-lobes subequal mE D. (i) L. acuminata. 13a. Petal-lobes unequal. 14. Posterior lobe of petal oblong; anterior lobe acuminate, shorter than posterior.. (8) L. Oestlundiana. 14a. Posterior lobe of petal falcate; anterior lobe suborbicular, larger than posterior. (13) L. Pristidis. 12a. Sepals acute or obtuse, not long-acuminate. Lateral sepals not recurved. 15. Petals ciliate; anterior lobe minute (7) L. nigriscapa. 15a. Petals eciliate; anterior lobe long. (5) L. disticha. l. Lepanthes acuminata Schlechter in Fedde Repert. 10 (1912) 355. Range: Guatemala. Honduras and Mexico (Chiapas). MEXICO: Dressler 1439. 4 Rhodor: [VOL. 61 2. Lepanthes ancylopetala Dressler, sp.nov. Herbae epiphyticae, caespitosae, usque ad 4.5 cm. altae. Folia ellip- tica, usque ad 18 mm. longa. Inflorescentiae folio breviores. Flores suc- cedanei, in racemis brevibus. Sepalum dorsale deltoideo-lanceolatum vel ovato-lanceolatum, acutum, 2-nervium: sepala lateralia usque ad medium vel ultra connata. Petala transverse bilobata, plusminusve sigmoidea; lobis posterioribus ligulatis vel lanceolatis, ca. 1 mm. longis, lobis anterioribus oblique deltoideis, quam posterioribus multo breviori- bus. Labellum cum apiculo infracolumnari, lobis lateralibus ovatis, acuminatis, columnam amplectentibus. Plant epiphytic, caespitose, 1.5-4.5 cm. tall. Secondary stem erect, monophyllous, 8-25 mm. tall, bearing 4—7 closely appressed sheaths, each with an ovate, acute, infundibuliform mouth: sheaths more or less hispidulose along nerves and margin. Leaves short petiolate, elliptic, tridenticulate apically, 3-5 mm. wide, 8-18 mm. long. Inflorescences I-several per growth, 3-10 mm. long (peduncle 1.5-5 mm.), rarely equalling the leaf, flowers alternate, 2-ranked, successive. Bracts of in- florescence infundibuliform, .5-1.1 mm. long, acute, carinate, more or less hispidulose or hispid-villous; pedicel .6-1 mm. long, jointed to ovary which is .5—.8 mm. long. Flowers rose-pink. Dorsal sepal deltoid-ovate, acute, somewhat concave, 1.2-1.6 mm. wide, 2.22.5 mm. long, 3-veined; lateral sepals 2-veined, obliquely deltoid-lanceolate to ovate-lanceolate, acute, .8-1.5 mm. wide, 2.2-2.8 mm. long, connate for 1.11.6 mm. Petals transversely bilobed, with a tooth more or less conspicuously developed at the junction of the lobes, the posterior lobe much the larger, each lobe somewhat bent near the base, so that the petal is often somewhat sigmoid in form; posterior lobe .8-1 mm. long, .25-.45 mm. wide, ligu- late to lanceolate, sometimes slightly falcate, acute to obliquely obtuse; anterior lobe obliquely deltoid, .2—5 mm. long, acute or obtuse. Mid- lobe of lip an upcurved apicule beneath column, .15-.25 mm. long, lateral lobes clasping column, ovate, acuminate and incurved anteriorly (more or less Comma-shaped) , 35-6 mm. wide, .7—.8 mm. long. Column .6-.85 mm. long, dilated above. MEXICO: Estado de Chiapas, Municipio de Ocosingo, Laguna Ocotál Grande, about 25-30 km. southeast of Monte Líbano. Alt. about 950 m. “Karst topography. Epiphytic on small mossy trees in pine forest, also (but less common) on trees along lake shore and in upper branches of tall tropical evergreen forest; flowers rose-pink." July 20-August 20, 1954. Robert L. Dressler 1475 (rype in Herb. Ames 69093) . This species is similar in habit to Lepanthes stenophylla Schltr. and L. oreocharis Schltr., but it is readily distinguished by its very unequal petal-Iobes; it is also a smaller plant. Other species, 1959] Schultes and Dillon, Genus Lepanthes 5 such as Lepanthes nigriscapa and the related L. inaequiloba A. & S., have very unequal petal-lobes; but the petals of these species are of a different shape, and the plants are of quite different habit. The specific epithet refers to the "bent petal" which is char- acteristic of the species. 3. Lepanthes avis Reichenbach f. Xen. Orch. 1 (1856) 144, tab. 50, II, 5-7. Range: Mexico (Vera Cruz, Hidalgo). MEÉxico: Moore 5288; Smith 341; von Kirchmar s. n. (Record from the Reichenbach Herbarium (TYPE) ) . 4. Lepanthes congesta R. E. Schultes, sp. nov. Herbae epiphyticae, caespitosae, usque ad 4 cm. altae. Caules secun- darii erecti, numerosissimi, vaginarum ostiis et nervis minute hispidulis. Folium rotundato-ellipticum, leviter marginatum, obtusum vel subacu- tum. Inflorescentiae folium multo supcerantes, congestiflorae. Flores minimi, succedanei. Sepalum dorsale valde cucullatum, late ovatum, acuminatum, 3-nervium. Sepala lateralia usque ad medium connata, ovata, apice subacuminata, 2-nervia. Petala transverse bilobata, inter lobos cum apiculo; lobi posteriores oblongi, margine integri; lobi anteri- ores minores sed similes. Labellum bilobatum, lobo utroque oblongo vel obovato; apice rotundato, incurvato. Columna superne dilatata. A caespitose, epiphytic herb up to 4 cm. high. Secondary stems erect, very numerous, concealed by several closely appressed hyaline sheaths, each of which terminates in an infundibuliform mouth; sheaths hispi- dulous along the nerves and on the thickened margin of the mouths. Leaves round-elliptic, marginate, obtuse or subacute, 8-9 mm. long, 6-7 mm. wide. Peduncles much longer than the leaves. Flowers many, borne in succession in very congested racemes. Lateral sepals about 2 mm. long, 1.1 mm. wide, connate for more than half their length, ovate but apically acuminate-acute, 2-nerved. Dorsal sepal cucullate, about 2 mm. long, basally 1.2 mm. wide, broadly ovate, 3.nerved. Petals transversely bilobed, with an apic ule between the lobes; the posterior lobes oblong, the margin entire, 1 mm. long, 0.5 mm. wide; the anterior lobes about 1 mm. long, 0.3 mm. wide. Lip bilobed, each lobe oblong or obovate with an incurved, rounded apex, 0.8 mm. long, 0.5 mm. wide. Column 1-1.2 mm. long. MEXICO: Estado de Vera Cruz, Orizaba. “On oak trees". May, 1905, C. A. Purpus 1355 (type in Herb. Ames 67044) . There has long been indecision as to the proper identification 6 Rhodora [VOL. 6i of Purpus 1355. Mr. A. A. Eaton, in the first decade of this century, studied the collection, made a sketch of the flower and wrote that the plant “agrees in essentials with Lepanthes avis but usually has more than one or two peduncles and there are discrepancies in detail of the flower; but as Reichenbach's draw- ings in Xenia are evidently poor, drawn apparently by somcone who did not know just what to represent, I think this may safely be called L. avis. Column agrees well." In 1959, Schultes studied the collection and expressed doubt that it could be referable to Lepanthes avis. There can be little uncertainty that, in Purpus 1355, we are dealing with a concept hitherto undescribed. There are a number of points of disagreement with Lepanthes avis, to which it seems to bear closest affinity. Lepanthes congesta has longer spikes, very much exceeding the leaves and usually has many more flowers. There are, furthermore, three or four inflorescences arising together, not, as in Lepanthes avis, singly or rarely in pairs. Lepanthes congesta has flowers which are yellow, not red and white as in L. avis. The lip of Lepanthes congesta is not divaricate but is, on the contrary, very strongly inturned at the tip of the lobes. Perhaps the difference of greatest degree is to be found in the petal, for while Lepanthes avis has. (according to Reichenbach’s figure) very unequal lobes, those of L. con- gesta are subequal with a definite apicule in the sinus between them. 5. Lepanthes disticha (Rich. & Gal.) Garay & R. E. Schultes, comb. nov. Pleurothallis disticha Richard & Galeotti in Ann. Sci. Nat.. sér. 3. 3 (1845) 16. For more than a century, this obscure concept was included in the genus Pleurothallis. Recent investigation, however, has shown it to belong definitely to Lepanthes. The accompanying plate of. Lepanthes disticha was prepared from Nagel et Monzón 6754. 6. Lepanthes Moorei C. Schweinfurth, sp. nov. Herba epiphytica, pusilla, caespitosa. Caules tenues, plusminusve 1959] Schultes and Dillon, Genus Lepanthes 7 erecti, vaginis pluribus arcte tubulatis omnino obtecti. Folium breviter petiolatum; lamina elliptica vel ovato-elliptica, subacuta vel subacumi- nata, basi cuneata. Inflorescentia saepissime unica, supra dense pauci- vel multiflora, quam folium brevior. Flos parvus, ut in genere mem- branaceus. Sepalum dorsale late ovatum, abrupte acutum. Sepala lateralia oblique ovata, abrupte acuta, per medium inferiorem connata. Petala transversa, in circuitu oblique ovato-lanceolata. Labellum trilo- batum, comparate magnum, lobis lateralibus triangulari-lanceolatis, in- curvis, columnam excedentibus. Plant epiphytic, small, slender, caespitose, up to 7 cm. high. Roots fibrous, glabrous, relatively stout. Rhizome abbreviated. Stems con- gested, very slender, erect or lightly arcuate, entirely concealed by 8 or less close tubular sheaths which are finely muriculate along the nerves and terminate in an ovate marginate hispid mouth, up to 5.5 cm. long (very variable in height). Leaf solitary, terminal, erect, spreading, shortly petioled; lamina elliptic or ovate-elliptic, sub-acute to short acuminate with a tridenticulate apex (when mature), cuneate at base, up to 2.1 cm. long and 1 cm. wide, chartaceous in the dried specimen. Inflorescences solitary (very rarely paired), more or less shorter than the leaf, densely 2- to many-flowered, up to 1.8 cm. long. Floral bracts congested, distichous, infundibuliform. Flower very small, membrana- ceous, column orange with red blotches. Sepals connate as in the genus, glabrous. Dorsal sepal broadly ovate, abruptly acute, 3-nerved below the middle, about 3 mm. long and 2.4 mm. wide. Lateral sepals obliquely ovate, abruptly acute, connate up to the middle, 2-nerved, about 2.8 mm. long (from the apex to the base of the column) and 1.8 mm. wide where broadest. Petals transverse, obliquely ovate-lanceolate in outline, minutely cellular-pubescent on the margins, about 1 mm. long and 2 mm. wide; posterior lobule relatively large. obliquely ovate-triangular, obtuse; anterior lobe very short, semiorbicular. Lip adnate to the lower part of the column, 3-lobed; lateral lobes relatively large, obliquely tri- angular lanceolate, incurved, obtuse, peltate, about 1.8 mm. wide above; mid-lobe minute, hirsute, apparently deciduous. Column arcuate- decurved, shorter than the lip, with an enlarged, subeuadrate, 3-dentate rostellum. MEXICO: Estado de Hidalgo, Distrito de Zacualtipán. “Slopes of ravine with pine-oak and moss-covered beech-magnolia woods. ‘To the left of road beyond Rancho EI Reparo, about 6.1 miles on road from Zacualtipán to Tlahuelompa.” Alt. about 2000 m. October 14, 1949, H. E. Moore, Jr. 5289 (rype in Herb. Ames No. 65879) . This species differs from Lepanthes inaequiloba A. & S. in its slender aspect, with differently proportioned petals and larger lip. 8 Rhodora [VOL. 61 7. Lepanthes nigriscapa R. E. Schultes & Dillon, sp. nov. Herbae epiphyticae, caespitosae, usque ad 15 cm. altae. Caules secun- darii erecti, plusminusve septemvaginati, vaginarum infundibuliformium ostiis hispidulis. Folia elliptica, obtusa, apice tridentata. Pedunculi foliis breviores vel subaequales. Flores pauci in racemis abbreviatis. Sepalum dorsale late lanceolatum, obtusum, 3-nervium. Sepala lateralia usque ad medium connata, late lanceolata, obtusa, 2-nervia. Petala late ellip- tica, transverse bilobata; lobi posteriores late elliptici, ciliati in margine exteriore; lobi anteriores minuti, obtuse triangulares. Labellum bilo- batum, lobo utroque dolabriformi, apice rotundato incurvo, in sinu incisura triangulari. Columna sursum dilatata. A caespitose, epiphytic herb up to 15 cm. high. Secondary stems up to ll cm. long, strong, erect, concealed by about 7-8 closely appressed, somewhat coriaceous sheaths, each of which terminates in an infundi- buliform mouth, Sheaths hispidulose along the prominent nerves and on the thickened margins of the mouths. The uppermost sheath en- closing the petiole of the leaf. Leaves 4.5 cm. long and 1.5-2 cm. wide, elliptic, obtuse, tridentate at the apex. Peduncles not over-topping the leaves. Flowers few, borne in succession in abbreviated racemes. Lateral sepals 2.3 mm. long, about 1-1.3 mm. wide, connate for more than half their length, broadly lanceolate, obtuse, ?-nerved, the inner nerve more prominent. Dorsal sepal broadly lanceolate, obtuse, 2.5 mm. long, 1.5 mm. wide, 3-nerved. Petals transversely bilobed, broadly elliptic in outline, 2.2 mm. wide; the posterior lobe broadly elliptic, 1.9 mm. wide and | mm. long, ciliate along the outer margin: the anterior lobe very minute, 0.3 mm. wide, obtusely triangular. Labellum bilobed, each lobe dolabriform with a rounded, strongly incurved apex, 1 mm. long, with a slight triangular indentation in the sinus. Column dilated upwards. The sepals are yellow, and the petals and lip are red. MEXICO: Estado de Oaxaca, Pacific slopes, northwest of Pluma Hidalgo. “Epiphytic on trees along river near the coffee plantation Copalita." Alt. about 1100 m. October 19, 1936, Nagel & Juan G[onzález] 6441. (type in Herb. Ames 51713). Future collections may indicate that Lepanthes nigriscapa should be treated as a variety of L. inaequiloba A. & S. The former has flowers which are only half as large as those of the latter and petals which are ciliolate, not glabrous. PLATE 1235. LEPANTHES NIGRISCAPA Schultes & Dillon. 1, plant, natural size. 2, flower, enlarged five times. 3, petal, enlarged ten times. 4, lip, enlarged fifteen times. — LEPANTHES OESTLUNDIANA Schultes & Dillon. 5, plant, natural size. 6, flower, enlarged five times. 7, petal, enlarged ten times. 8, lip, enlarged twenty times. DRAWN BY G. W. DILLON. LEPANTHES L Oestlundiana Schultes d Litton PLATE 1235. LEPANTHES. 10 Rhodora VOL. 61 Lepanthes nigriscapa scems to be related to L. hondurensis Ames of Honduras, differing from it in part by having smaller flowers with obtuse (instead of acuminate) sepals. The petals of the two species differ very markedly. ‘The structure of the column is similar in both species. While the lips are similar, that of Lepanthes nigriscapa lacks the apicule in the sinus which is found in L. hondurensis. Vegetatively, Lepanthes nigriscapa and L. hondurensis are very similar, but the latter has more numerous and more closely placed sheaths than the former, and its leaves tend to be acuminate. The lip of Lepanthes nigriscapa resembles that of L. cascaja- lensis Ames of Costa Rica in being deeply bilobed with inturned apices, but the lobes of L. cascajalensis are flat, bearing no thick- ened ridge. The petals of Lepanthes cascajalensis are. broader, while the sepals are broader and more acuminate. ‘The column differs markedly. Vegetatively, the two species are very distinct. With the exception of the lip, there is little to point to a rela- tionship between Lepanthes nigriscapa and L. cascajalensis. 8. Lepanthes Oestlundiana R. E. Schultes & Dillon, sp. nov. Herbae epiphyticae, laxe caespitosae, usque ad 14 cm. altae. Caules secundarii erecti vel leviter patuli, multivaginati, vaginarum infundibuli- formium ostiis hispidulis. Folia ovato-elliptica, acuminata, apice tri- dentata. Pedunculi foliis breviores vel subaequales. Flores pauci in racemis abbreviatis. Sepalum dorsale ovatum, valde acuminatum, 3- nervium. Sepala lateralia ovata, usque ad medium connata, acuminata, 2-nervia. Petala eciliata, transverse bilobata, late lanceolata, lobis pos- terioribus oblongis, lobis anterioribus multo minoribus. Labellum unguiculatum, bilobatum, lobo utroque malleoliformi, eciliato, ovato, apice acuminato. Columna tenuis, sursum dilatata. Plants loosely caespitose; epiphytic herbs up to 14 cm. high. Second- ary stems erect or slightly spreading, up to 9.5 cm. long, concealed by about 6-9 closely appressed, more or less coriaceous, blackish sheaths each one of which terminates in an infundibuliform mouth. Sheaths his- pidulose along the prominent nerves and on the thickened margin of the mouth; uppermost sheath enclosing the petiole of the leaf. Leaves broadly elliptic-ovate, abruptly acuminate, prominently tridentate at the apex, up to 4.5 cm. long and 2 cm. wide. Peduncles not overtopping the leaves. Flowers few, borne in succession in abbreviated racemes. Dorsal sepal 3 mm. long and 1.5 mm. wide, broadly ovate, strongly acuminate, 3-nerved. Lateral sepals 3 mm. long and about 1.2 mm. wide, connate 1959] Schultes and Dillon,—Genus Lepanthes 11 for about half their length, ovate, strongly acuminate, 2-nerved with the inner nerve more prominent. Petals 2.4 mm. wide, transversely bi- lobed, broadly lanceolate in outline, eciliate. Anterior lobe of petals 0.7 mm. long, 1.5 mm. wide, oblong. Posterior lobe much smaller, 0.9 mm. wide. Labellum 1 mm. wide, unguiculate, bilobed, each lobe mal- leoliform, eciliate, ovate in outline, with a prominent central ridge or fold, the apex acuminate and slightly incurved. Column very slender, dilated upwards. Sepals and labellum pale green; petals yellow and red. MEXICO: Estado de Vera Cruz, Zacuapam. “Epiphytic on trees in damp and shady places near a brook in virgin forest.” Alt. about 2925 m. February 12, 1932, Otto Nagel 2657 (TYPE in Herb. Ames, 51708) . Lepanthes Oestlundiana seems to be related to Schlechter’s L. scopula, a native of Middle America. There is a drawing of the type of Lepanthes scopula in the Ames Herbarium. ‘The leaves of Lepanthes Oestlundiana are more constantly acuminate, the sepals are more acuminate, the lip has broader lobes of a slightly different shape than the corresponding parts of L. sco- pula. The posterior lobe of the petals of L. scopula, furthermore, is strongly retrorse, while that of L. Oestlundiana is character- istically retrorse. However, the relationship between these two species, is close. Lepanthes Oestlundiana is related also to L. nigriscapa from the Pacific slopes of Oaxaca. The petals of both species have the same general pattern. The posterior lobes of the petals of Lepanthes Oestlundiana are oblong and eciliate, while those of L. nigriscapa are broadly elliptic and ciliate on their outer margin. The anterior lobe of the petals of the former is oblong, of the latter obtusely triangular. The sepals of Lepanthes Oest- lundiana are ovate and strongly acuminate, those of L. nigri- scapa are elliptic and obtuse. The lips, while of the same gen- eral pattern, differ in that the lobes of the lip of Lepanthes Oestlundiana bear a central ridge or fold while in L. nigriscapa the ridge is almost marginal. The apices of the lobes of the lip of the former are acute and eciliate, those of the latter are obtuse and coarscly ciliate. Lepanthes Oestlundiana has broadly ellipticovate leaves with prominently tridentate apices, while in L. nigriscapa the leaves are elliptic with the tridentation of the apices less marked. These differences, though constant and 12 Rhodora [VOL. 61 certainly of specific importance, nevertheless are not grcat, and it is clear that the two species are very closely related. ‘This close relationship assumes additional significance when it is empha- sized that Lepanthes Oestlundiana is a plant of the Atlantic watershed of Vera Cruz, while L. nigriscapa grows on the Pacific watershed of Oaxaca. We take pleasure in naming this species in memory of the late Mr. Erik Ostlund from whose large herbarium of orchids this collection came. 9. Lepanthes oreocharis Schlechter in Fedde Repert. 10 (1912) 483. Range: Guatemala and Mexico (Chiapas, Guerrero). MEXICO: Matuda 1688; Juan G[onzález] 1060; G. B. Hinton 1744; Wendt s. n. (TYPE). 10. Lepanthes orizabensis R. E. Schultes & Dillon, sp. nov. Herbae epiphyticae, caespitosae, usque ad 5 cm. altae. Caules secun- darii erecti, plusminusve septemvaginati, vaginarum infundibuliform- ium ostiis hispidulis. Folia suborbicularia vel elliptica, marginata, ob- tusa, apicibus tridentatis. Inflorescentiae folium superantes. Flores pauci, succedanei, in racemis brevibus. Sepalum dorsale late lanceolatum, acuminatum, 3-nervium. Sepala lateralia lanceolata, usque ad medium coherentia, apicibus acuminatis recurvis, 2-nervia. Petala eciliata, trans- verse bilobata, lobis anterioribus oblongis cum apicibus rotundatis, lobis posterioribus cuneatis, apicibus obliquis. Labellum bilobatum, lobo utroque triangulari, apice acuminato incurvato. Columna tenuis, sursum dilatata. Plants epiphytic, caespitose, up to 5 cm. high. Secondary stems erect, 1.5-3.5 cm. long, concealed by about 7 closely appressed hyaline sheaths, each terminating in an infundibuliform mouth; sheaths his- pidulose along the prominent nerves and the thickened margin of the mouth. Uppermost sheath enclosing the petiole of the leaf. Leaves suborbicular to elliptic, marginate, obtuse, 0.8-1.3 cm. long, about 0.7 cm. wide, apex inconspicuously tridentate. Peduncles overtopping the PLATE 1236. LEPANTHES REKOI R. E. Schultes. 1, plant, natural size. 2, flower, enlarged five times. 3, petal, enlarged ten times. 4, lip, enlarged sixteen times. — LEPAN- THES TENUILOBA Schultes & Dillon. 5, plant, natural size. 6, flower, enlarged two and one half times. 7, lip, enlarged ten times. 8, 9, petals showing variation, enlarged five times. — LEPANTHES DISTICHA (Rich. & Gal.) Garay & Schultes. 10, plant, natural size. 11, flower, enlarged five times. 12, lip, enlarged fifteen times. 13, petal, enlarged ten times. — LEPANTHES ORIZABENSIS Schultes & Dillon. 14, plant, natural size. 15, flower, enlarged five times. 16, lip, enlarged ten times. 17, petal, enlarged ten times. DRAWN BY G. W. DILLON. L. tenuiloba Schultes § 492 Ton RESchulles LEPANTHE-S Kee hoi bensis Oc Hu es delet lor L. origa L- RT oe : EA 4 S à XD Wi K o, D A j N > =; M. “5 oor — zi Garay ERC Schuttes PLATE 1236. LEPANTHES. 14 Rhodora [VOL. 61 leaves. Flowers few, borne in succession in abbreviated racemes. Dorsal sepal broadly lanceolate, acuminate, 4.5 mm. long, 2.25 mm. wide, 3- nerved. Lateral sepals coherent for almost half their length, lanceolate, narrowing to a recurved, acuminate tip, 4.5 mm. long, 1.5 mm. wide, 2-nerved, the inner nerve more prominent. Petals about 2.7 mm. wide, eciliate, transversely bilobed. Anterior lobe oblong with rounded apex; posterior lobe cuneate with oblique apex. Lip 1.8 mm. wide, enfolding the column, bilobed, each lobe more or less triangular with a central ridge, the apex acuminate and slightly incurved. Column slender, di- lated upwards. MEXICO: Orizaba, Vera Cruz, 1855, Müller s. n. (ver in N. Y. Bot. Gard.) ; Sierra de Agua [Vera Cruz), May 1854, Müller 982. Lepanthes orizabensis seems to be allied to L. avis, being simi- lar vegetatively, in the shape of leaf, in the lax inflorescence and 5 ) in general aspect. Lepanthes orizabensis is larger, however, and differs from Z. avis (according to Reichenbach’s rather sketchy drawings) in the shape of the petals, the shape of the lip and the relative size of the flowers. Another ally of Lepanthes orizabensis is L. Schiedei, the type specimen of the former having been originally determined as the latter species. Ihe absence of cilia on the sepals of Lepanthes orizabensis and the differently shaped lip, however, make these two species quite distinct. Lepanthes orizabensis is likewise very closely allied to L. fracti- flexa A. & S. of Cuba, from which it can be separated by its smaller flowers and sepals of a different shape. ll. Lepanthes papillipetala Dressler, sp. nov. Herbae parvulae, caespitosae, epiphyticae, usque ad 22 mm. altac. Folia petiolata; laminae orbiculares vel ellipticae, usque ad 12 mm. longae. Inflorescentiae folium breviores. Flores succedanei, in racemis brevibus. Sepalum dorsale deltoideo-ovatum, acutum, 3-nervium. Sepala lateralia oblique ovato-lanceolata, acuta, 2-nervia, usque ad medium connata. Petala transverse bilobata, pilis minutis papillatis ornata; lobis posterioribus ligulatis, apicibus obtusis, ca. 1 mm. longis; lobis anterioribus angustioribus ligulato-lanceolatis. Labellum cum apiculo infracolumnari, lobis lateralibus ovatis, acuminatis, hispidulis, columnam amplectentibus. Plant epiphytic, caespitose, 10-22 mm. tall. Secondary stem erect, monophyllous, 4-10 mm. tall, bearing 3-4 closely appressed sheaths, each with an ovate, acute, infundibuliform mouth; sheaths more or less 1959] Schultes and Dillon,—Genus Lepanthes 15 hispidulose along nerves and margins. Leaves petiolate, petiole ca. 2 mm. long, blade orbicular to elliptic, apically tridenticulate, with a pronounced submarginal vein, 3-5 mm. wide, 5-12 mm. long. Inflo- rescences one to several per growth, 3-5 mm. long (peduncle 1-2.5 mm.), flowers alternate, 2-ranked, successive: bracts of inflorescence in- fundibuliform, ca. 1 mm. long, acute, carinate, more or less hispidulose. Pedicel 1-1.5 mm. long, jointed to ovary, which is ca. 1 mm. long. Flowers light pink, dorsal sepal deltoid-ovate, acute, 1.75-2.2 mm. wide, 22.7 mm. long, 3-veined, midvein with few minute papillae externally, margin sometimes very minutely papillate-serrulate. Lateral sepals 2- veined, obliquely ovate-lanceolate, acute 1.1-1.25 mm. wide, 2-2.9 mm. long, connate for ca. 1.8 mm., apices slightly divergent. Petals trans- versely bilobed, with a tooth (the apex of the petal) more or less devel- oped at the junction of the lobes; posterior lobe ligulate, obtuse, .8- 1.1 mm. long and .4—55 mm. wide, bearing minute papillate hairs, espe- cially on inner margin (these hairs ca. .04-.05 mm. long): anterior petal lobes narrower, ligulate-lanceolate, obtuse, with a tuft of minute papil- late hairs at tip, .2-.3 mm. wide, .65-1 mm. long. Mid-lobe of lip an apicule beneath column, ca. .2 mm. long; lateral lobes clasping column, ovate, acuminate and incurved anteriorly (more or less comma-shaped) , hispidulose, especially on posterior margin, .75-.8 mm. long, .3-.5 mm. wide. Column ca. .75 mm. long. dilated above. MEXICO: Chiapas, Municipio de Ocosingo, Laguna Ocotál Grande, about 25-30 km. southeast of Monte Líbano. Alt. about 950 m. "Karst topography. Flowers light pink, epiphytic on small mossy trees in pine forest, also (but less common) on trees along lake shore and in upper branches of tall tropical evergreen forest." July 20-August 29, 1954. Robert L. Dressler 1476 (rype in Herb. Ames 69092) . This dwarf species is at once distinguished from all its allies, except L. parvula, by the papillate hairs of its petals. Various species, such as the Lepanthes Lindleyana Oerst. & Reichb. f.— complex and L. Rekoi have ciliate petals, but papillate petals are not common in the genus. In habit, this species resembles Lepanthes avis, but the lateral lobes of the lip in the latter species are divergent rather than closely enclasping the column as in the former. Lepanthes papillipetala is closely related to L. parvula but is distinguished by its entire sepal-margins, its hispidulose lip and its larger flowers. 12. Lepanthes parvula Dressler, sp. nov. Herbae parvulae, caespitosae, epiphyticae, usque ad 22 mm. altae. 16 Rhodora [VOL. 61 Folia orbiculari-elliptica, usque ad 10 mm. longa. Inflorescentiae folio breviores. Flores succedanei, in racemis brevibus. Sepala ciliata. Sepal- um dorsale deltoideo-ovatum, acutum, 3-nervium. Sepala lateralia oblique ovata, acuta, 2-nervia, usque ad medium vel ultra connata. Petala transverse bilobata, papillis minutis ornata, lobis falcatis, apici- bus obtusis. Labellum cum apiculo infracolumnari, lobis lateralibus ovatis, acuminatis, minutis, papillatis, columnam amplectentibus. Plant epiphytic, caespitose, 7-22 mm. tall. Secondary stem erect, monophyllous, 3-12 mm. tall, bearing 3-5 closely appressed sheaths, each with an ovate, acute, infundibuliform mouth; sheaths more or less hispidulose along nerves and margins. Leaves petiolate, petiole ca. ! mim. long, blade orbicular-elliptic, apically tridenticulate, with a some- what pronounced submarginal vein, 2.5-6 mm. wide, 3.5-10 mm. long. Inflorescences one to several per growth, 2-5 mm. long (peduncle .5-2 mm.), flowers alternate, 2-ranked, successive; bracts of inflorescence infundibuliform, ca. .7 mm. long, carinate, acute, hispidulose. Pedicel ca. .8 mm. long, jointed to ovary, which is ca. .5 mm. long. Flowers light pink, sepals ciliate on margins and slightly papillate on exterior of midveins; dorsal sepal deltoid-ovate, acute, 1.3-1.6 mm. long, .9-1.1 mm. wide, 3-veined (lateral veins very weakly developed) ; lateral sepals 2-veined (only inner vein well developed) , obliquely ovate, acute, ca. .6 mm. wide, 1.3-1.6 mm. long, connate for ca. .9 mm. Petals minutely papillate, transversely bilobed, with a deltoid tooth more or less devel- oped at the junction of the lobes; lobes more or less equal, falcate from a deltoid base, obtuse, .45-.6 mm. long, whole petal lunate with central, deltoid tooth on concave side. Mid-lobe of lip a pubescent apicule ca. 15 mm. long beneath column, lateral lobes minutely papillate, ovate, acuminate, ca. .6 mm. long, clasping column, Column ca. 1.6 mm. long, dilated above. Capsule ca. 2 mm. long. MEXICO: Chiapas, Municipio de Ocosingo, Laguna Ocotál Grande, about 25-30 km. southeast of Monte Líbano. Alt. about 950 m. “Karst topography. Flowers light pink, epiphytic on small mossy trees in pine forest, also (but less common) on trees along lake shore and in upper branches of tall tropical evergreen forest." July 20-August 20, 1954. Robert L. Dressler 1477 (rv in Herb. Ames 69094) . In its ciliate sepal-margins and general habit, this species re- sembles the larger Lepanthes Schiedei Reichb. f. but differs in several characters: the short inflorescence, the much smaller flowers and the more or less papillate petals and lip, which are differently shaped. From Lepanthes papillipetala Dressler, this species may be distinguished at once by its ciliate sepals, its less strongly papillate petals, its papillate lip and its smaller flowers. 1959] Schultes and Dillon, Genus Lepanthes 17 These species grow together in pine forests and, without the aid of a lens, can be separated only on the basis of a marked differ- ence in flower size. It is interesting that three distinct new species of Lepanthes should be found growing together in the same habitat, especially so when two of the species appear to be rather closely related. 13. Lepanthes Pristidis Reichenbach f. in Linnaea 22 (1849) 820; Xen. Orch. 1 (1856) 151, t. 50, figs. IV, 11-14; in Walpers Ann. Bot. 3 (1852-53) 156. Range: Cuba and Mexico (Vera Cruz). MEXICO: Liebold s. n. (Rec- ord from the Reichenbach Herbarium) (TYPE); Sartorius s. n. (Record from the Reichenbach Herbarium); Schaffer 51 (Record from Reichen- bach Herbarium). Reichenbach's drawings of this concept are so inadequate that it is not clear from them just what are the affinities of L. Pristidis. A study of the type drawings and of Reichenbach's description, however, indicate that it is distinct from the other species of Mexican Lepanthes. It is probably related to Lepanthes disticha. Lepanthes Pristidis occurs also in Cuba (Acufia, “Cat. descr. orquid. Cub." (1939) 145). 14. Lepanthes Rekoi R. E. Schultes in Bot. Mus. Leafl. Harvard Univ. 6 (1938) 193. Range: Mexico (Puebla and Oaxaca) . MEXICO: Juan González] 2415; Schultes 817; Schultes & Reko 336 (TYPE). This exceptional little orchid, known only from one station in northeastern Oaxaca and one station in northeastern Puebla, shows unmistakably clear relationships with Lepanthes Dawsonii Ames of Honduras. The two have in common a trilobate lip which is apparently of rare occurrence in the genus. It has been thought that Lepanthes is a genus in which con- siderable latitude should be allowed in the identification of speci- mens, because intraspecific variation has been thought to be rather great. Since colonies of Lepanthes are usually small and frequently isolated, collections of species of this genus seldom contain sufficient material to permit a study of intraspecific varia- tions. Fortunately, Schultes was enabled to study a number of 18 Rhodora [VOL. 61 living specimens for variation. within the species at the type locality of Lepanthes Rekoi. In August 1938, the type of Lepanthes Rekoi was collected on an oak tree on the eastern slopes of the great Cerro de los Frailes in the District of Teotitlän in northeastern Oaxaca. Diligent search yielded only one flowering and one fruiting specimen. ‘The characters of this species were so striking and sharp, how- ever, that Schultes described the plant as new. A collection by Juan G[onzález] from Puebla, though differing in some slight details from the type, was cited with the type. In July, 1959, Schultes revisited the type locality and found in full flower the colony from which the type came. Again, search within a 150-yard radius failed to reveal another colony. Nearly thirty flowers were closely examined before pressing and were compared with a drawing of the type. Very little deviation from the type was noticed. More flowers were critically examined in the laboratory with the type at hand; almost no variation in floral parts and no variation in specific characters was in evidence. 15. Lepanthes Schiedei Reichenbach f. in Linnaca 22 (1849) 820; Xen. Orch. 1 (1856) 144, t. 50, figs. I, 1-4; in Walpers Ann. Bot. 3 (1852-53) 524. Range: Mexico (Vera Cruz and Chiapas). mexico: Matuda 1595; Schultes 1012; Schiede s. n. (Record from Reichenbach Herbarium (TYPE). l6. Lepanthes tenuiloba R. E. Schultes & Dillon, sp. nov. Herba parvula, caespitosa, epiphytica. Caules secundarii erecti, tenues, monophylli, plusminusve trivaginati, vaginarum infundibuliformium ostiis hispidis. Folium ovatum vel ovato-ellipticum, coriaceum, apice tridenticulatum, basi cuneatum. Pedunculi folium superantes. Flos grandis, usque ad 15 mm. longus, purpureo-ruber. Sepala connata, glabra, in circuitu elliptica; sepalum dorsale elongato-triangulare, acu- tum, prominenter 3-nervium; sepala lateralia triangulari-lanceolata, acutiuscula, 2-nervia. Petala parvula, glabra, transverse bilobata; lobis anterioribus lanceolatis longo cum angusto, abrupte acuminato, lineare apice; lobis posterioribus oblongo-lanceolatis. Labellum bilobatum: lobis lateralibus in circuitu oblique oblongo-ellipticis, in sectione 1959] Schultes and Dillon, - Genus Lepanthes 19 abrupte malleoliformibus. Columna tenuis, apice dilatata, inter labelli lobos circumdata. Plant very small, 2.5-3 cm. tall to the tip of the leaf, caespitose, epi- phytic. Roots fibrous, short, stout. Secondary stems slender, erect, monophyllous, from 3-10 mm. long, concealed by about 3 hyaline, close- tubular sheaths which end in infundibuliform mouths; sheaths pubes- cent on the longitudinal nerves and on the margins of the mouth; the uppermost sheath largest, enclosing the petiole of the leaf. Leaf ovate or ovate-elliptic, 8-12 mm. long, 4-7 mm. broad, tridenticulate at the apex, cuneate at the base, coriaceous, marginate, short-petiolate. Pe- duncles conspicuously overtopping the leaf, very slender, bearing near the middle one tubular sheath with an infundibuliform mouth, up to 30 mm. long, few-flowered (usually only one expanded at a time). Flowers very large for the plant, when spread out up to 1.5 cm. long from the tip of the dorsal sepal to the apex of the lateral ones, borne in succession on a raceme. The sepals glabrous, purplish-red, when spread out forming a perfect ellipse in outline; lateral sepals connate to about the middle, asymmetrically triangular-lanceolate, acutish, 2- nerved, about 7 mm. long, 2.5 mm. wide at the base; dorsal sepal elon- gate, triangular, acute, prominently 3-nerved, about 7 mm. long, 5 mm. wide at the base. Petals very small, purplish-red, glabrous, transversely bilobed; the anterior lobe lanceolate, 0.7 mm. wide at the base, with a long narrow, abruptly acuminate, linear apex; the posterior lobe oblong- lanceolate, 2 mm. long, 1 mm. wide or less. Labellum bilobed; the lateral lobes obliquely oblong-elliptic in outline, 1.6 mm. long, 0.5 mm. wide, in section abruptly malleoliform, tightly enclasping and partialiy obscuring the column. Column slender, slightly dilated at the apex. MEXICO: Estado de Chiapas, Siltepec, August 9, 1937, E. Matuda 1595a. (rype in Herb. Ames, No. 46731; Duplicate Type in Herb. Univ. Michigan). Lepanthes tenuiloba is not closely related to any known Cen- tral American species of Lepanthes, unless it be with L. guatema- lensis Schltr. 'The two species agree in floral size and in general habit, but there are significant structural differences in the flowers. It is easily distinguished from other species by the flowers which are exceedingly large, considering the size of the plant, and by the connate sepals which, when spread out, form a perfect ellipse in outline instead of being constricted at the junction of the dorsal with the lateral sepals. The relative minuteness of the petals in comparison with the large sepals is a feature which cannot be found in many species. Distinctive also is the shape 20 Rhodora [VOL. 6i of the petals with their curiously attenuated anterior lobes and the shape of the labellum which tightly enfolds the very slender column. Lepanthes tenuiloba has the largest flowers among the Mexican species, although several species are larger vegetatively. Matuda 1595 was a mixed collection. It contained several ex- cellent specimens of Lepanthes Schiedei Reichenb. f. The half dozen or more specimens of the much larger flowered Lepanthes tenuiloba have been renumbered Matuda 1595a. — BOTANICAL MUSUEM or HARVARD UNIVERSITY AND AMERICAN ORCHID SOCIETY, CAMBRIDGE, MASS. NOTES ON THE FLORA OF A GORGE OF ESOPUS CREEK, NEW YORK Henry F. DUNBAR The object of this paper is to record evidence of a changing flora: (1) remnants of a flora of a colder climate which is pre- sumed to have preceded the present time but followed the last glacial period, (2) presence of species characteristic of a warmer climate and which may be the vanguard of a gradual plant mi- gration into this area. DESCRIPTION OF THE AREA This study covers the slopes immediately adjacent to Esopus Creek and creek bed, Ulster County, N. Y. from the Ashokan Reservoir to the flood plain at Marbletown. Through most of this area the creek flows rapidly with very little quiet water over a stony or gravelly bottom. The shores, mostly heavily wooded, rise steeply, and, in some places consist of sheer cliffs of Devonian sandstone often 100 ft. high. Frequent sand terraces and gravel bars occur with a flora characteristic of such places. ‘The eleva- tion of the study area lies between 200 and 500 feet above sea level. The average rainfall is about 43 in. per year and the temperature range is from a minimum of -25? F to a maximum of about 100° F. The predominating trees on the sand, gravel or silt shores are poplars (Populus deltoides and P. grandidentata), sycamore (Pla- tanus occidentalis), and both the gray and speckled alders (Alnus serrulata and A. rugosa). The steep slopes adjacent to the creek are heavily covered with hemlock and white pine (Tsuga cana- densis and Pinus Strobus), the three birches (Betula lenta, B. lutea, B. papyrifera), and the maples (Acer rubrum, A. saccharum. A pensylvanicum). On the higher slopes various oaks (Quercus alba, Q. Prinus, Q. rubra, Q. velutina, and Q. coccinea) complete with the pines and the hemlocks. ‘The more conspicuous shrubs along the shore are the ninebark (Physocarpus opulifolius) , the dog- wood (Cornus Amomum), and various willows (Salix spp.) . 21 22 Rhodora [VOL. 61 Of outstanding interest is the occurrence of species usually found further north or at higher elevations. Their presence here in cool, shady, moist situations suggests that they may be relics of a time when a colder climate favored their more general dis- tribution. Such a condition doubtless followed the last glacial period, and, as warming up progressed, plants requiring a damp and cool environment failed to survive in the territory surround- ing the gorge. SIGNIFICANT PLANIS OF THE AREA The species listed below scem to be noteworthy, judging from published ranges available to me — chiefly those of the 8th edi- tion of Gray's Manual, which, except where otherwise indicated, is the basis for remarks on ranges. The nomenclature used fol- lows that of the Manual. Specimens of all of the plants mentioned have been collected by me, unless otherwise indicated, and are deposited in the herbarium of the New York State Museum in Albany, N. Y. HIGH NORTHERN BryuM MUHLENBECKIIL This moss, of circumpolar distribution, has previously been reported in the northeastern U. S. only from northern parts of New York and New England.! MYURELLA JULACEA. Grout describes this species as a “subalpine moss, south to Conn." ketchledge collected both of these mosses on the gorge. He also reports the Myurella from northern New York.? LYCOPODIUM SELAGO, var. PATENS. Only three or four plants of this clubmoss were found in the gorge growing on rocks close to the creck. This boreal species occurs in the Catskill Mountains in Greene Co., about 25 miles away, and recently I found a large stand of it in the Shawangunk Mts. about 15 miles south of the study area. WoopsiA ALPINA. This fern of “arctic regions", has not previously been reported from New York State south of Essex Co. in the northern Adirondacks,? but it has been recorded from both northern Vt. and Me. About 25 plants were found scattered along a northeast-facing cliff. CRYPTOGRAMMA STELLERI. A circumboreal species, reaching locally in northeastern North America into Penn. and N. J. In New York State ! GROUT, A. J. Moss Flora of North America. Vol. 11. p. 234. ? KETCHLEDGE, EDWIN H. Checklist of Mosses of New York State. New York State Museum Bulletin number 363. ? Map files of the Botany Office, New York State Museum, Albany, N. Y. 1959] Dunbar,—Esopus Creek, N.Y. 23 there are four known stations in Greene Co. and one in Delaware Co. Thus this station becomes the southernmost one reported in the state. SEepUM RoskA. The occurrence of this plant in the gorge is the fifth New York station. Usually growing in arctic regions, it has been found in Madison, Yates, Greene and Schuyler Cos., N. Y. It has also been collected in northeastern Penn. and on Roan Mountain in North Caro- lina. Uhl has reported 11 gametic chromosomes in plants from the Esopus Gorge.* ACHILLEA BOREALIS. (A. Millifolium L. var. nigrescens E. Mey.) Identified by S. J. Smith who believes that it should be given varietal status, following the treatment of Cronquist in the Britton and Brown Illustrated Flora. According to Smith, these specimens are identical with those found in some of the Adirondack gorges, near Elmira and in the Delaware Water Gap. CANADIAN ZONE The following plants exhibit more or less isolated extensions of range from the cooler valleys of the Catskills. PorvsricHUM Braunu. This fern common in the Catskills at ele- vations from 1500 to 2200 feet, is rare at other altitudes. One plant was observed on a wet, north-facing cliff. It was subsequently washed away during a flood and hence was not collected. PICEA RUBENS AND ABIES BALSAMEA. ‘Three or four depauperate speci- mens of each were found on the wooded slopes. Abundant stands of both occur in the Catskills above 3500 ft. some 15 or 20 miles to the west. ‘TRISETUM SPICATUM var. MOLLE. This grass was collected in the gorge by S. J. Smith. EPILOBIUM CILIATUM. Smith identified this species also growing in the same area. LoseLiA KALMII. Fernald reports this plant as growing from New- foundland to central Maine, northern New Jersey and southeastern Pennsylvania. Here it occurs in several locations on wet cliffs. The following might be designated as Appalachian rather than Canadian. ILEX MONTANA. But one (male) plant has been found in the area. VIOLA ROTUNDIFOLIA. This violet occurs sparsely in moist, shady ravines. VIBURNUM ALNIFOLIUM. This shrub, growing in similar places to the violet, is even more scarce. SOUTHERN The following plants suggest that they may have moved into this 4 Heteroploidy in Sedum Rosea (L) Scop. Evolution 6; 81-86. 1952. 24 Rhodora [VOL. 61 area as the climate warmed up. It is true that in each case there are instances of their occurrence even further north, but they are scarcely expected to be found associated with those listed above and are offered as possible evidence of post-glacial northward migration. CORALLORHIZA ODONTORHIZA. ‘The range for this orchid is from southwestern Maine to Georgia and Alabama. TEPHROSIA VIRGINIANA. This species is recorded from Florida north to southern New Hampshire. GALIUM PILOSUM. This bedstraw occurs in the Finger Lakes region of New York and sparsely in the Hudson Valley, but is not reported from the mountain regions of New York State.? SOLIDAGO ODORA. With a distribution much like that of the Tephro- sia, this plant is equally unexpected in the gorge. ACKNOWLEDGMENTS I wish to express my appreciation of help in preparing this paper to the following persons: Dr. R. T. Clausen, Dept. of Botany, Cornell Univ, for data on Sedum; E. H. Ketchledge, Department of Botany, N. Y. State School of Forestry for data on mosses; Stanley J. Smith, New York State Museum for permission to examine the Botany Office files and for many helpful sugges- tions; Professor G. S. Torrey of the Univ. of Conn., and Charles Schweinfurth of the Botanical Museum of Harvard University, for a critical reading of the manuscript and many constructive thoughts. PLANTS NEW TO ILLINOIS AND IO IHE CHICAGO REGION JULIAN A. STEYERMARK AND FLOYD A. SWINK Since the last publication by the authors and Dr. Thieret!, the following new records have come to light. All specimens are deposited in the herbarium of the Chicago Natural History Museum. PLANTS NEW TO ILLINOIS BurOMUS UMBELLATUS L. ‘This species was found growing spontane- ously and aggressively spreading in a small pond south of 87th Street and east of Kean Avenue near Buffalo Woods Forest Preserve, Cook Co., Aug. 6, 1957, Swink 3007. BETULA PUMILA L. var. GLABRA Regel. This variety is not given for Illinois in the 8th Edition of Gray's Manual. The following collection belongs to this variety: Low mound in swale, north of Waukegan, Lake Co., June 8, 1908, Frank C. Gates 2500. PopoPHYLLUM PELTATUM L. f. DEAMH Raymond. This color form of 1959] Steyermark and Swink,—Plants New to Illinois 25 the May Apple, previously known in this country only from Indiana and Missouri, has recently been collected in the Chicago area, as follows: Tinley Creek Woods along bridle path east of Harlem Avenue, 114 miles south of Highway 143, May 15, 1957, Herbert Pahnke s.n. PopOPHYLLUM PELTATUM L. f. POLYCARPUM Clute. This form was re- ported by Pepoon? from the Chicago area but no specimens are extant. The data for the only herbarium record for Illinois are as follows: Woods bordered by Mill Creek Road, Kelly Road, and Dilleys Road, on property of Dr. and Mrs. L. F. Yntema, northwest of Wadsworth, about 5 miles southwest of Zion, Lake Co., August, 1956, Jean Yntema s.n. RUBUS OCCIDENTALIS L. f. PALLIDUS (Bailey) Robins. The data for the collection of this form with cream- or amber-colored fruit are as follows: Open borders of Quercus alba-Quercus rubra-Carya ovata woodland bordering swampy meadow, on property of Dr. and Mrs. L. F. Yntema, west of Highway 41, northwest of Wadsworth, about 5 miles southwest of Zion, Lake Co., July 18, 1956, Steyermark 81982. Rosa MICRANTHA Sm. This adventive species has become established at the following locality: Natural prairie overlying glacial gravel moraine on crest of slope just east of Highway 173 and north of Forest Hill Golf Course, T44N, R2E, Sect. 8, 0.8 mile northeast of junction of Highways 51 and 173, just northeast of Rockford, Winnebago Co., Oct. 1, 1955, Steyermark 80289. RHAMNUS FRANGULA L. var. ANGUSTIFOLIA Loud. This unusual narrow- leaved variety is not recorded in any of the principal floras of the United States; it was found growing with typical Rhamnus frangula. The nar- row elongate Salix-like leaves have irregularly undulate-scalloped mar- gins. The data are as follows: Near a small pond west of 104th Avenue south of Longjohn Slough approximately 14 mile south of 95th Street in Palos Forest Preserve south of Willow Springs, Cook Co., July 2, 1957, Swink 3001. OXYDENDRUM ARBOREUM (L.) DC. Three trees of this species, of dif- fering sizes and ages, have been located near each other in southern Cook County. They have certainly become naturalized at this locality, where they were growing with the native flora. The data are as follows: In a quaking aspen thicket at the north end of Sweet Woods Forest Pre- serve south of Thornton, Cook Co., Sept. 29, 1956, Swink 2787. In addi- tion, there is a specimen of this species in the Chicago Natural History Museum Herbarium collected by Robert Ridgway on Sept. 1,5. 1920: with the following location: 914 miles northeast of Olney, on Corn Ford Road, Richland Co. In his Ligneous Flora of Richland County, page 30. Ridgway states that the specimen "had long been misplaced and was with a lot of herbaceous material sent to the Missouri Botanical Garden, where identified by Mr. J. M. Greenman. ‘There is some reason 26 Rhodora [VOL. 61 for suspecting that the original label had slipped from the corner and had been replaecd by one which may not have belonged to the specimen in question, a memorandum to that effect having been inserted. How- ever, whether this suspicion is correct or not, there can be no question that the specimen was collected in. Richland County, the only doubt being as to the locality and date”. This tree probably was introduced. VERBENA STRICTA Vent. f. RoskirLORA Benke. The first collection for this rose-colored form for Illinois is as follows: Dry pasture, Starved Rock State Park, La Salle Co., July 27, 1940, George D. Fuller 2299. GALEOPSIS LADANUM L. f. PARVIFLORA (Lam.) Koch. This is a small- flowered form with corolla only 3-1 mm. long, equalling or only slightly exceeding the calyx. ‘This form is not mentioned in any of the prin- cipal manuals of the flora of the central or eastern states. The data are as follows: Along Midway Plaisance, Chicago, Cook Co., Aug. 10, 1956, Juan V. Poncho 7020. The only other species of the genus recorded for Illinois by Jones et al. is Galeopsis tetrahit, which is recorded from Boone County, where collected by E. W. Fell, and from Cook County, on the basis of the report by Higley & Raddin*. The following addi- tional specimens of Galeopsis tetrahit are to be found in the herbarium of the Chicago Natural History Museum: H. N. Patterson, from the vicinity of Oquawka, Henderson Co., in 1897; and W. C. Ohlendorf, from Lawndale, Cook Co., in August, 1885. VIBURNUM LANTANA L. This shrub was collected in woods of York Forest Preserve south of Elmhurst, Du Page Co., November 5, 1955. It was found in a secluded area of the preserve far from any habitation and growing among native species; F. A. Swink 2782 SOLIDAGO ULIGINOSA Nutt. var. LiNOIDES. (T. & G.) Fern. Solidago uliginosa var. linoides has not been recorded for Illinois either in the 8th Edition of Gray’s Manual or in Jones et al The following collection may be cited: Lake Co., natural swampy meadow fringed by Salix spp., Cornus stolonifera, and Betula pumila, on east side of Highway 59, north of Miller Road, T43N, R9E, Sect. 12, 4 miles north of Barrington, October I, 1955, Steyermark 80305. PLANTS NEW TO THE CHICAGO REGION Horcus LANATUS L. This Old World grass, reported from Champaign and Crawford counties by Jones et al? has been collected in a lawn at 9800 Willow Springs Road in the Palos Forest Preserve area, where it was growing spontaneously, F. A. Swink 3004 PARONYCHIA FASTIGIATA (Raf.) Fern. var. PALEACEA Fern. ‘This species ! Steyermark, J. A., Swink, F. A., and Thieret, J. W. Plants New to Illinois and Indiana and the Chicago Region. RHODORA 59:31—33. 1957. " Pepoon, H. S. Flora of the Chicago Region. Chicago Academy of Sciences. 1927. 1959] Steyermark and Swink,—Plants New to Illinois 27 has been reported by Jones et al.? from central and southern Illinois only, but is found to be rather frequent in several of the Cook County forest preserves. Collections were made from two of these localities, as follows: Open woodland just east of Hidden Pond in Hidden Pond Woods Forest Preserve, August 14, 1957, F. A. Swink 3009; and open woodland near picnic area of Wolf Road Woods, F. A. Swink 3010, August 15, 1957. STELLARIA PUBERA Michx. var. SILVATICA (Béguinot) Weath. This plant, previously known in the state only from Cook County, was col- lected in shaded ground of Warrenville Forest Preserve, Du Page Co., June 2, 1956, F. A. Swink 2789. AGRIMONIA ROSTELLATA Wallr. This species is locally abundant in open wooded areas of Cantigny Woods Forest Preserve near La Grange, Cook Co., as the following collection shows: F. 4. Swink 3011. Previous localities, as recorded in Jones et al.*, show its distribution in central and southern Illinois. Lorus coRNiCULATUS L. On a sandy knoll beside a small creek in a cow pasture 2 miles south of the Kankakee River, 14 mile west of Ilinois-Indiana state line, 2 miles north of Hopkins Park, Kankakee Co., August 4, 1957, Karl E. Bartel 6. Previously known in Illinois only from Champaign, Iroquois, and Jo Daviess counties (Jones et al.%, p. 286). HYPERICUM SPATHULATUM (Spach) Steud. In a roadside ditch 14 mile south of the Will-Kankakee county line, 3 miles northeast of Essex, Kankakee Co., August 2, 1953, F. A. Swink 2380. Previously known only from the southern and west-central portions of the state. EPILOBIUM HIRSUTUM L. In rich black soil along the Santa Fe Rail- road south of Argonne National Laboratory at the southwest end of Rocky Glen Forest Preserve, Du Page Co., July 21, 1957, Karl E. Bartel 7. Previously known only from Cook Co., where first reported by Steyer- mark and Swink (RuoponaA 51: 147-149, 1949). SWERTIA CAROLINIENSIS (Walt.) Ktze. At edge of woods near picnic area, Pioneer Woods Forest Preserve, 107th Street and. Willow Springs Road. Cook Co., July 29, 1956, J. W. Thieret 2137. This plant was re- ported by Pepoon?, who stated that the station had been destroyed in the construction of a paved highway. Although reported and accom- panied by a photograph on page 427 of Pepoon's Flora, there has appar- ently been no authentic specimen kept of this original discovery. The collection by Dr. Thieret represents a new and different location for this species. Jones, G. N., Fuller, G. D., Winterringer, G. S., Ahles, H. E., and Flynn, A. A. Vascular Plants of Illinois. 1955. t Higley, W. K. and Raddin, C. The Flora of Cook County, Illinois, and a part of Lake County, Indiana. Bull. Chi, Acad. Sci. 2, no. 1: 94. 1891. 28 Rhodora [VOL. 61 A MILESTONE IN BOTANICAL CARTOGRAPHY.! — Geobotanical investiga- tions in general and studies on the distribution of taxa in particular were instrumental in forming one of the main pillars on which the theory of evolution rests. Likewise, this approach has greatly affected the development of taxonomic botany during the past one hundred years. It has also strongly influenced the understanding of the need for extensive collections and large herbaria wherever taxonomical and geo- botanical studies are to be performed. It has been said that over-con- fidence in some of the hypotheses, advanced by this important school of thought has sometimes resulted in unnecessary splitting of species because of some geographical distinctiveness; a typical case may seem to be the thick volume XII of the magnificent Flora SSSR with its 849 species of Astragalus. There are also instances when too strong adherence to an originally fruitful hypothesis has counteracted further research on details seemingly offsetting these ideas. As a whole, however, the geobotanical approach to taxonomy and evolution has been one of the most prolific ones in botany in the past and it will certainly continue to be so for a long time to come. In studying the distribution of species and their past history, differ- ent approaches have been tried. They have developed from the very schematical descriptions of areas given by Linnaeus and his predeces- sors, through the more elaborate outlines of Willdenow, Wahlenberg, and von Humboidt, to the more or less detailed maps of present-day publications. The distribution maps seem to have originated with DeCandolle, who used a few such outlines in his "Géographie botanique raisonné” in 1855. They were, however, developed further by the in- fluential Austrian school of geobotany, and then notably by its greatest representatives, Kerner von Marilaun and his son-in-law, Richard Wettstein. In later years, Scandinavian followers of this school im- proved considerably the methods of mapping and stressed the necessity of greater exactitude. As a direct result of this, the so-called dot-maps, on which small dots represent every collection or locality, have been employed in Scandinavian geobotanical and taxonomical works for many years. ‘This scientific art has recently reached fulfillment in the well-known “Atlas of the distribution of vascular plants in N.W. Europe,” worked out and published by Professor Eric Hulten of Stock- holm in 1950. That Atlas gives exact maps of the known distribution in Fennoscandia (in its wider sense) of almost all the species occurring within the area. It is based on records in literature and on detailed studies in the main herbaria in these countries, which certainly are better known botanically than any other comparable area in the world. ! ERIC HULTEN: The Amphi-Atlantic Plants and Their Phytogeographical Connections. Kungliga Vetenskapsakademiens Handlingar. Fjarde Serien. Band 7. Nr. 1. Stockholm 1958. pp. 1-340. 1959] Lóve,—Botanical Cartography 29 In addition, more schematical maps show the general distribution of the species, but these are in no way detailed and are sometimes incorrect; they were never intended to be more than approximative. Although Professor Hultén knows more about the distribution of the higher plants in the entire northern hemisphere than does anybody else, his interest has been focused on certain problems of past dis- persals involving plants of Scandinavia and of the Beringian region. His studies on the flora of the latter area effected his coining of the now universally accepted theory of equiformal progressive areas, which may perhaps be regarded as an outgrowth of the age and area hypoth- esis. He was able to demonstrate, in 1937, that this theory could explain most distribution areas of plants on the continents, and also that it could give distinct indications as to the place of origin, or rather place of survival, of the different species. At that time, Professor Hultén tried to press this hypothesis to explain the distribution of all plants in the boreal zone, and especially those confined to a limited area in northwestern Europe and to a larger area in eastern North America. These are the truly amphi-Atlantic plants in the restricted meaning of the term, but Professor Hultén stressed that they could only be prop- erly studied in connection with plants having larger areas, or what others have named bis-Atlantic distribution. Scandinavian botanists had long regarded the strictly amphi-Atlantic distribution as an indi- cator of a former trans-Atlantic land connection, but Professor Hultén maintained that if they were seen from his wider point of view these areas could more appropriately be explained as being remnants only of a formerly circumpolar area. Despite several indications to the con- trary, recently reviewed in a good article by Dr. Eilif Dahl in the Nor- wegian journal Blyttia, Professor Hultén has vindicated this point of view in several articles and also in his Atlas, but the lack of adequate maps of the total extent of the species under discussion has made it almost impossible to consider the entire problem on a fully scientific basis. In a recent book on "Ihe amphi-Atlantic plants and their phyto- geographical connections," Professor Hultén has published a group of 279 maps of species which apparently are selected with the above-men- tioned explanation in mind. ‘These are the most accurate maps of general distribution ever published of so many species at the same time, and their reproduction and exactness are such that most remarks on them must be regarded as vague comments only. When all the boreal flora has been so mapped with the same care as given to these maps for Europe, this method of indirect inquiry into the history and dispersal of the boreal species will be exhausted and the hypotheses based on them will then have to be tested by aid of other and more exact approaches. 30 Rhodor: [VOL. 61 In the new book, Professor Hultén starts with a very readable and concise introduction, stating in a nutshell the present knowledge of the distribution of boreal plants in general and of his opinions as to the explanation of their areas in particular. This fascinating chapter is followed by information about the material on which the maps are based, and also about the plants included and excluded. In this con- nection it must be said that comparing the reasons given for the exclu- sion of some species, and the maps given of some others to which the same principles could have been applied. indicates that the selection has been somewhat arbitrary, but this cannot be avoided as long as all the species are not mappable. And the reader must keep in mind that there is a distinct tendency behind the selection, as stated in the introductory chapter. The bulk of the book consists of a short text about the individual species, on the left hand page, and detailed maps of two species or taxa, in dark-brown and greyish blue, on the right hand page. The book concludes with a list of references, a comprehensive, though far from complete, bibliography, and an index. A detailed description of the maps is not possible; they have to be seen and studied to be fully appreciated. “There is no doubt that future geobotanists will long draw upon the wealth of information here col- lected, and the book is likely to influence plant geographers in such a way as no other recent publication can do. American readers will prob- ably be able to find a lot of “mistakes” or "omissions" in the areas of many species, whereas European botanists will have less opportunity to add to the information collected by the author. Ihe reason for this discrepancy between the American and European parts of the distribu- tions mapped is evidently caused by the fact that our present knowledge of the American flora is considerably more restricted than that of the European plants. Not only can this be seen by comparing the vast number of flora lists and treatments (both past and present) of small regions in Europe with the low number of such articles printed in America (and the difficulties in getting such treatments printed here), but it is perhaps best seen by comparing the number of specimens in European and American herbaria. European herbaria contain over 77 million specimens, whereas the herbaria in the United States and Can- ada, an area twice as large as Europe. contain only about 32 million specimens. Still more relevant to the present study is the fact that less than 2 million specimens of herbarium plants are available in Canada as a whole, and vast regions here are almost devoid of collections. Therefore, criticism of the maps for possible mistakes in this very essential part of the area should be made with this in mind, at the same time as caution in the hypothesizing is required just for the North American parts of the areas. A rather common misunderstanding may seem to have affected the 1959] Love,—Botanical Cartography 31 author when he tries to explain the fact that considerably more species have been introduced into North America from Europe than the reverse way. He regards this as connected with the much greater disturbances by cultivation in the New World. Certainly, some weeds may have got an unexpectedly wide distribution in a short period because of this fact, whereas ecological and climatical conditions may counteract this to a certain extent. In order to get weeds established, however, even the most favourable conditions for dispersal within the country are of no significance if seeds are not carried over the ocean, and the main reason for the difference in number of introduced species must be the fact that considerably more seeds were transported westward than eastward. All the settlers brought with them effects of different kinds and seeds from their homelands, including many weeds. In addition, the fishing vessels carried fish only towards Europe but ballast westwards, and this ballast was very often soil which was carried ashore in the new country. The whole problem of introduction of plants and animals has recently been excellently studied in a book by a compatriot of Professor Hultén, namely in “The faunal connection between Europe and North America” by Professor C. H. Lindroth, the eminent ento- mologist. In that book the elimination of introduced species, before discussing possible dispersal routes of other plants or animals, is done more thoroughly than plant geographers have ever done with their material. Professor Hultén has excluded most introduced species, and he has mapped some others to show areas of no significance to the geobotanical problems aimed at. But although there is no doubt that the author has been very critical and correct in his selection in most Cases, others are occasionally somewhat irrelevant, since only palvnologi- cal studies can. demonstrate with certainty the age of some of these plants in the flora. It is hardly very logical to regard a species as intro- duced in North America solely because it is introduced in New Zealand, but it is to be hoped that this kind of reasoning has not been employed on other species than Juncus subnodulosus. Naturally, the species concept of Professor Hultén is based on that of the classical geobotanical school, but it may sometimes seem to be un- necessarily wide and other times it is unduly narrow. Because of his careful differentiation of what he believes are intraspecific races, this does not matter, since taxa which others may regard as species are usually shown with different signs. In a few cases, however, the usual caution has broken down, with worthless maps as a result. Examples of such inadequate maps are, e.g., those of Molinia coerulea and Myrica Gale. It does not matter that the map of Eriocaulon septangulare shows no distinctions between the endemic British species and the American E. pellucidum with which it has long been misidentified, but the map of what the author calls Sisyrinchium montanum is a complete mess, 32 Rhodor: (VOL. 61 including a mixture of S. albidum, S. Bermudianum, S. montanum s.str.. and a still undescribed endemic species from Ireland. ‘This last mixture is, perhaps, the very best demonstration of the fact that the morphological-geobotanical method, though excellent, is by no means sufficient for exact studies of this kind, since it cannot always prevent evolutionary heterogeneity being included into a single taxon. Professor Hultén should be excused for ignoring recent biosystem- atic data which in fact would have added considerable strength to the arguments the maps are intended for, though not always in support of his own ideas. Also, unfamiliarity with cytogenetics alone is responsi- ble for his attempts to use terms and explanations from this branch of botany to enhance his views on some species: his "cytogenetical" ex- planations of Antennaria Porsildii, Draba fladnizensis and D. lactea, and a few other groups are not supported by available cytogenetical evidence. Only in a very few cases docs this affect the maps slightly, and the data in the maps are much more important than the good or bad judgments we may pass upon them now or later. It is pity that all new combinations proposed, on pp. 16, 38. 52. 74, 96. 106, 146, 174, 204, 246, 262, 294, and perhaps elsewhere, are illegiti- mate since the author does not follow the old-fashioned rule for refer- ences to the basionyms in the text, as required by the International Code. It is to be hoped that at least some of these combinations will be legitimized elsewhere. There are a few printing errors, none of them essential. The dot in Iceland for Luzula pallescens ought to be erased, and some mix-up has caused a dot on the same island for Carex Hartmani — the same dot is correctly placed for Carex adelostoma based. on the same original information. Léve 1951 on p. 96 should be Love & Love, and Rousseau on p. 152 is probably an error for Rouleau. The specific epithet of Carex macloviana is everywhere capitalized by mistake. And several of the references mentioned in the text are not met with in the bibliograhy, which, nevertheless, is rather comprehensive and invaluable as a source for further literature on boreal geobotany. In conclusion it must be said that this most recent one of the many outstanding contributions to boreal geobotany by Professor Hultén is a worthy addition to his long list of books and the most valuable map collection ever published. It is the very best basis available for discus- sions on the amphi-Atlantic. plants. However, the problem of the origin of these areas cannot be properly discussed before maps are available of all the boreal and arctic flora, a task nobody can fulfill better than. Professor Hultén himself. Even when such maps are avail- able for the entire flora of the northern zone, they will have to be com- plemented by detailed evolutionary studies of each species and bv extensive palynological investigations before all doubts are removed. 1959] Lóve,—Botanical Cartography 33 Although the present maps by no means can be regarded the final step towards the solution of this delicate and important geobotanical prob- lem, they are undoubtedly a very important step in the correct direction. In this connection it does not matter if the explanation these maps are intended to support will prove to be right or wrong, since the main object of the compiler is, after all, not to vindicate his own old opinions but to get the correct explanation of the facts expressed in the peculiar areas of the relatively few species of plants with true amphi-Atlantic distribution. ASKELL LOVE, INSTITUT BOTANIQUE DE L'UNIVERSITE DE MONTREAL. Volume 60, No. 720, including pages 289 — 344, was published 30 January, 1959. ow REFERENCE LIRRARY MAY 11 1959 Dodota JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor-in-Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR, IVAN MACKENZIE LAMB Vol. 61 February, 1959 No. 722 CONTENTS: The Present Status of the Recently Described Genus, Teuscheria. Leshe A. Garay iier irt en a Fehlen e aE 35 A Revision of Bartonia and Obolaria (Gentianaceae). 00 /// ors LN E 43 Elsholtzia ciliata in Essex County, Massachusetts Stuart K. Harris. sess e S 63 The New England Botanical Club, Ine. Botanical Museum, Oxford St., Cambridge 38, Mass. RHODORA.—A monthly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid, in funds payable at par in United States currency in Boston; single copies (if available) 60 cents. Back vol- umes l-58, with a few incomplete, can be supplied at $5.00 per volume. Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes, relating directly or indirectly to the plants of North America, will be considered by the editorial com- mittee for publication. Articles concerned with systematic botany and cytotaxonomy in their broader implications are equally accept- able. All manuscripts should be double-spaced throughout. Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance, will be furnished at cost. Address manuscripts and proofs to Reed C. Rollins, Gray Herbarium, 22 Divinity Avenue, Cambridge 38, Mass. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill, Botanical Museum, Oxford Street, Cambridge 38, Mass. Second Class Postage Paid at Boston, Mass. EATON PHESS, Inc. Specialists in Learned Publications 17 SPRING STREET WATERTOWN 72, MASS. QTRbooora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 61 February, 1959 No. 722 THE PRESENT STATUS OF THE RECENTLY DESCRIBED GENUS, TEUSCHERIA LESLIE A. GARAY It was only recently that I established the genus Teuscheria in the American Orchid Society Bulletin on the basis of a collec- tion from Ecuador by Mr. Henry Teuscher. While this paper was in the proof-reading stage, another new species, T. vene- zuelana which is described herewith, came into flower in Caracas in the collection of Mr. G. C. K. Dunsterville. This second species at once reminded me of a taxon described by Schlechter as Bifrenaria Pickiana. When I examined critically our material of B. Pickiana in the Orchid Herbarium of Oakes Ames, I was surprised to find that it also has the characters of my new genus. Thus within a few months our knowledge of the monotypic genus Teuscheria, through the discovery of two additional species, was enlarged both as to number of species and as to their distribution. All species of Teuscheria are epiphytic in habit. The primary stem or rhizome is elongated and completely beset with imbricat- ing sheaths which at an carly stage brcak down into numerous, fibrous segments. The pseudobulbs are either aggregated or more or less remote and are hidden by three layers of scarious cataphylls. "These cataphylls tightly enclose the developing pseudobulbs and, as was observed by Mr. ‘Teuscher, they expand together with the pseudobulbs. During the process of expansion, the outer two layers soon wither and also break up into fibres, 35 36 Rhodora [VOL. 61 giving a brush-like appearance to the pseudobulbs. The inner- most layer scems to remain persistent and it encloses not only the pseudobulb but the basal portion of the petiolate leaf as well. The pseudobulbs are relatively small, pyriform in shape and are terminated by a single leaf. It is noteworthy to mention that the pseudobulb is not articulate with the petiole of the leaf but is continuous with it. The persistent leaves are rather thin in tex- ture, elongate with a prominent petiole. The blade itself is dis- tinctly but not prominently plicate, oblong-ligulate, with a sharply acute or subacuminate apex. ‘The inflorescence is always one-flowered and borne laterally at the base of the pseudobulb, either of the recent year’s growth or that of the previous year. The slender peduncle is either erect or pendent and adorned with a number of remote sheaths. The relatively large flowers are exceedingly interesting because they are not resupinate, i.c. the lip is uppermost in position during anthesis. The non- resupinate position is attained through a 45° bend of the ovary, an unusual method seldom seen in the orchid family. The flowers, which are open only for 24 hours, show a few morphologically interesting points. The fleshy sepals and petals are either con- nivent or spreading; the lateral sepals are shortly connate just above the column-foot and form a small spur-like structure where the free tip of the column-foot, to which the base of the lip is firmly adnate, is enclosed. This free prolongation of the column-foot is apparently very prominent in T. venezuelana, while in T. Pickiana it is hardly produced. The large lateral lobes of the flabellate lip enfold the short, cylindrical column. The four pollinia are unequal in size, without a stipe, and are held together by a small amount of viscose matter at the tip of the rostellum. When Teuscheria was published, a colleague wrote me that he considers T. cornucopia to be synonymous with Bifrenaria Pickiana, and that Teuscheria does not belong to the Phajeae, TEUSCHERIA CORNUCOPIA GARAY. Fig. 1. general habit of plant, slightly less than natural size; fig. 2. front view of flower, x 134; fig. 3. side view of flower with dorsal sepal, petals and one lateral sepal removed, x 134; fig. 4. lip expanded, x 2; fig. 5. the apex of the column, much enlarged; fig 6. dorsal sepal, x 134; fig 7. petal, x 315; fig. 8. pollinia, much enlarged. cornucopia TEUSCHERIA_. G aray 38 Rhodora [VOL. 61 as I proposed, but rather to the subtribe Lycasteae. In the generic description of Teuscheria, of course, I gave a clear description of the nature of pollinia: "Pollinia 4, cerea, inaequalia, visco parco cohaerentia, sine glandula". Bifrenaria is known to have a well developed stipe and gland. When I examined our material of B. Pickiana, I was unable to locate any stipe or gland. In one collection from Ecuador (Drew E-654), however, I found the reason for Schlechter's decision to refer this taxon to the genus Bifrenaria. Bifrenaria Pickiana has a large, triangular rostellum which breaks off very easily. When one examines Schlechter's original drawing, the prominent, triangular gland of the pollinia is clearly seen. The stipe, as depicted by Schlechter, is surely imaginary, and represents, very probably, only the viscid matter. Bifrenaria Pickiana has always been perplexing to orchid tax- onomists. It was described from cultivated material and was assumed to be native to Colombia. For a long while, nothing more was known about this species; then, unexpectedly, it ap- peared in a collection from Honduras. Later it was collected on a number of occasions in Costa Rica as well as in Mexico, Colom- bia and Ecuador. Dr. L. O. Williams, having seen living material of it from Costa Rica, decided that it should be included in the genus Xylobium, and he made the indicated transfer. ‘This transfer, however, did not solve the placement problem of the species, because one of the generic requirements of Xylobium is that the inflorescence is always several-flowered. If the inflores- cence is one-flowered only, the species would ordinarily be refer- able to the genus Lycaste. However, both Xylobium and Lycaste have pollinia with a very prominent stipe. Bifrenaria Pickiana was, therefore, still misplaced. With the detection of Teuscheria cornucopia and the additional species described in this paper, we at the Ames Herbarium believe that the proper disposition of B. Pickiana is in the genus T'euscheria. Teuscheria Garay, in American Orchid Society Bulletin 27: 820, 1958. Sepalum posticum inferum, a sepalis lateralibus omnino liberum; sepala lateralia lata, basi pedi columnae adnata, mentum fornicatum vel erectum calcariforme formantia. Petala sepalis inclusa et angustiora, Labellum sepalis inclusum, superum, involutum, trilobum, ad apicem 1959] Garay — Present Status of Teuscheria 39 pedis columnae incurvum adnatumque et cum co non articulatum. Columna brevis, crassa, basi in pedem longissimum apice incurvum producta; clinandrium vix evolutum, marginatum; rostellum valde pro- ductum, late triangulum. Pollinia 4, cerea, inappendiculata, visco parco cohaerentia, sine glandula. Ovarium genuflexum. — Herbae epiphyti- cae rhizomate abbreviato, nunc elongato. Pseudobulbi oblique pyri- formes, unifoliati. Folium plicatum, angustum, petiolatum. Inflorescen- tia singula, lateralis, uniflora. Flos satis conspicuus, non resupinatus, a latere visus nonnunquam cornucopiam in mentem revocans. CLAVIS SPECIERUM 1 Lobus intermedius labelli lobis lateralibus aequilongus vel brevior; pseu- ODUM Dip APO Tee Atte 2 ³ĩðꝗ 86 5 oe 1* Lobus intermedius labelli lobis lateralibus duplo vel triplo longior; pseu- 0 meee nt ee anc EM S nde T. Pickiana 2 Petala oblancolata; margo labelli antice crenulata .............. T. venezuelana 2* Petala oblongo-lineares; margo labelli antice integra .......... T. cornucopia Teuscheria cornucopia Garay, in Amer. Orch. Soc. Bull 27: 820, 1958. Epiphytica, caespitosa, usque ad 20 cm. alta; rhizomate valde abbre- viato; pseudobulbis aggregatis, oblique pyriformibus vel ovato-pyriformi- bus, monophyllis, ca. 2 cm. altis; folio lineari-oblongo, plicato, promi- nenter 3-nervio, 15 cm. longo, ] cm. lato; inflorescentia singula, erecta vel paulo arcuata, gracili, in medio univaginata, 4 cm. longa; flore nutanti, satis conspicua, a latere visu cornucopiam in mentem revocante; sepalo postico elliptico vel obovato-elliptico, apice obtuso, 5-nervio, 14 mm. longo, 7 mm. lato; sepalis lateralibus obliquis, oblongo-ovatis, apice acutis, 5-nerviis, 16 mm. longis, basi 9 mm. latis; petalis lineari-oblongis, apice acutiusculis, 5-nerviis, 10 mm. longis, 4 mm. latis; labello cuneato- flabellato, antice 3-lobo, lobis lateralibus rotundatis, lobo terminali reniformi, in medio leviter retuso, disco supra medium callo 3-dentato ornato, toto labello 16 mm. longo, 15 mm. lato; columna humili, crassa, dorsaliter 4 mm. alta, basi in pedem longissimum producta; ovario cum pedicello ca. 1.5 cm. longo. Ecuador: Western slopes of the Andes. 100 km. from Guayaquil. Altitude about 3,000 ft., H. Teuscher s.n. (TYPE in the Ames Orchid Herbarium no. 69,265.) The plant from which the type was taken is cultivated in the collection of the Montreal Botanical Garden. ‘The above description was prepared from living material. Teuscheria venezuelana Garay, sp. nov. Epiphytica, caespitosa, usque ad 35 cm. alta; rhizomate valde abbre- viato; radicibus filiformibus, flexuosis, glabris; pseudobulbis oblique pyriformibus, unifoliatis, ca. 2.5 cm. altis; foliis lineari-oblongis, plicatis, apice acutis vel subacuminatis, basin in petiolum angustatis, usque ad 40 Rhodora [VOL. 61 Apex of Leaf Column, Lip (sectioned) & Loleral Sepal 2mm Anther & Potlinia 474. Teuscheria venezuelana Garay 35 cm. longis, 2.5 cm. latis; inflorescentia singula, pendula, uniflora; pedunculo satis gracili, leviter arcuato, vaginis bracteiformibus remotis obsesso, ca. 7 cm. longo; sepalo postico oblongo-obovato, apice acuto 1959] Garay — Present Status of Teuscheria 41 vel obtusiusculo, concavo, satis carnoso, 19 mm. longo, 7 mm. lato; sepalis lateralibus obliquis, oblongo-ovatis, apice acutis vel obtusiusculis, basi oblique decurrentibus, 25 mm. longis, 8 mm. latis; petalis oblique oblanceolatis, satis carnosis, apice acutis, 18 mm. longis, 5 mm. latis; labello cuncato-flabellato, antice trilobo, lobis lateralibus suborbiculari- bus, lobo intermedio haud producto, rotundato, disco a basi usque ad medium callo elevato, farinoso ornato; margine labelli antice crenulata; toto labello 22 mm. longo, antice 18 mm. lato; columna satis humili, crassa, basi in pedem longum producta. Venezuela: Choroni Road, seaward slopes. Altitude about 5,000 ft., G. C. K. Dunsterville 474 (TYPE in the Ames Orchid Herbarium no. 69,206.) This description was prepared from living material. The color of the fleshy sepals and petals is bronze with a light maroon tinge; the lip is white with a diffused pink flush near the margin. Teuscheria Pickiana (Schltr.) Garay, comb. nov. Based on Bifrenaria Pickiana Schltr. in Orchis 6: 8, fig. 1, 1912. See also Mansf. in Fedde Rep. Beih. 58: t.49, Nr.196, 1930. Xylobium Pickianum (Schlu.) L.O.Wms. in Ceiba 4: 271, fig., 1955; ibidem 5: 187, 1956. Epiphytica, usque ad 65 cm. alta; rhizomate lignescente, elongato, flexuoso, cataphyllis imbricatis, jam emarcescentibus, fibrosis omnino obtecto; radicibus filiformibus, flexuosis, glabris; pseudobulbis satis distantibus, ovoideis vel oblique pyriformibus, cataphyllis scariosis demum fissis omnino obtectis, unifoliatis, 1.5—3 cm. altis; folio oblongo- ligulato, plicato, apice acuminato vel acutiusculo, basi in petiolum angus- tato, petiolo incluso 15-60 cm. longo, 1-3 cm. lato; inflorescentia singula, uniflora; pedunculo gracili, leviter arcuato, interdum reflexo (valde post anthesin), ca. 3 cm. longo; sepalo postico satis carnoso, oblongo- elliptico, apice obtuso, 5—7-nervio, 12-15 mm. longo, 4—5 mm. lato; sepalis lateralibus oblique ovato-oblongis vel triangulari-oblongis, apice acutis vel obtusis, 5—7-nerviis, cum pede columnae mentum conicum, obtusum formantibus, 13-18 mm. longis, 5-7 mm. latis; petalis oblongo- ligulatis, obtusis, 3—5-nerviis, 12-16 mm. longis, 4-5 mm. latis; labello e basi cuncata trilobo, lobis lateralibus oblique oblongis, apice obtusis, lobo intermedio cuneato-subquadrato vel cuneato-flabellato, apice brevi- ter exciso, lobis lateralibus duplo vel triplo longiore; disco in medio callo 3-dentato ornato; toto labello 15-25 mm. longo, inter lobos late- rales 13-18 mm. lato; columna humili, crassa, basi in pedem longissimum producta; ovario cum pedicello ca. 4-5 mm. longo. Mexico: Chiapas; limestone area, near Laguna Ocotal Grande. Alti- tude about 950 m., R. L. Dressler 1629 (ames). Honduras: in moist forest near Nacimiento del Río Lindo, Dept. Cortés. Altitude about 600 42 Rhodora [VOL. 61 Teuscheria Pickiana (Schltr.) Garay m., Williams & Molina 14555 (EAr); Santa Cruz de Yojoa, Dept. Cortés. Altitude about 2000 ft., Edwards 541 (ames, Ny). Costa Rica: region between Río Esquinas and Palmar Sur de Osa, vicinity of Esquinas Sta- tion, Prov. Puntarenas. Altitude ca. 30 m., P. Allen 5584 (EAP); forested hills near Esquinas Station, vicinity of Río Esquinas, Prov. Puntarenas. Altitude ca. 60 m., P. Allen 5688 (Har); Río Blanco, North-east Cordil- leras. Altitude ca. 1000 m., F. C. Lehmann 1219 (Ames, G.). Colombia: Dept. del Valle, Cordillera occidental, Hoya del Río Digua. Altitude 900-1180 m. J. Cuatrecasas 14901 (ames). Ecuador: Prov. Imbabura, Cordillera occidental. Above Rio Chalguayaco. Altitude about 4300 ft., W. B. Drew E-634 (AMES) .— ORCHID HERBARIUM OF OAKES AMES, BOTANI- CAL MUSEUM, HARVARD UNIVERSITY. A REVISION OF BARTONIA AND OBOLARIA (GENTIANACEAE) ! Joun M. GILLETT I. THE GENUS BARTONIA MUHL. The generic history of Bartonia Muhl. (named in honor of Professor Benjamin Smith Barton, 1766-1815, of Philadelphia, as was Bartonis Sims of the Loasaceae), has been rather thor- oughly reviewed by Fernald and Weatherby (1932) in an amusing little paper entitled “Bartonia, a Comedy of Errors." Briefly, these authors argued in favor of retention of the name and pro- vided a list of generic synonymy with bibliography. Bartonia was subsequently conserved just twenty years later in order to avoid confusion with Sims’s genus. Willdenow credited Muhlenberg with the discovery of the genus and provided a description in 1801 but the single species B. tenella ascribed to it is usually credited directly to Willdenow. Michaux independently described Centaurella in 1803 with two species, C. verna and C. paniculata. Britton, Sterns and Poggen- berg found that B. tenella Willd. had been described previously as Sagina virginica by Linnaeus and made the combination B. virginica (L.) BSP. Persoon in 1805 erected the genus Centaurium (not Hill, 1756) describing C. autwmnale based on Centaurella paniculata Michaux. Muhlenberg then made the combinations Bartonia verna (Michx.) Muhl. and B. paniculata (Michx.) Muhl. in his Catalogue of 1813. Robinson then made a superfluous combina- tion B. paniculata (Michx.) Robins. when he decided that Small’s Bartonia lanceolata described in 1903 was synonymous with it. Pursh did not simplify the situation when in 1814 he recognized three species and one variety employing Michaux’s genus Cen- taurella. These were C. vernalis and a f uniflora, C. aestivalis 1 Contribution No. 1748 from the Botany and Plant Pathology Division, Science Service, Canada Department of Agriculture, Ottawa, Ontario. 43 44 Rhodora [VOL. 61 and C. autumnalis. The first two entities are perhaps segregates of Bartonia verna but their identities are still in doubt because the types have not been located. I have included them in the synonymy of C. verna based on the description. Again the type of C. aestivalis has not been located and Pursh’s description is inadequate to establish its position. The last, C. autumnalis, in- cluded both C. paniculata Michx. and Sagina virginica L. because the type of the latter, Clayton 649, is cited. Pursh made no men- tion of Persoon's earlier Centaurium autumnale. Grisebach in his Species Gentianearum of 1839 employed Michaux’s Centaurella containing three species, C. autumnalis Pursh, C. verna Michx. and a C. moseri Steudel & Hochstein. He also described a B brachysepala of C. autumnalis. Grise- bach's treatment is similar for De Candolle's Prodromus of 1845. Asa Gray in the Synoptical Flora employed Bartonia as the generic name recognizing two species B. tenella and B. verna. C. moseri he regarded as synonymous with B. tenella saying, sé an occasional form, with leaves or scales and branches mostly alternate.” Centaurella moseri was subsequently transferred to Bartonia, the combination attributed to Robinson & Schrenk by Gilg in the Pflanzenreich in 1895. ‘The Pflanzenfamilien treatment of the Gentianaceae by Gilg placed Bartonia between Canscora and Obolaria in his Erythraeinae, enumerating three species, B. ten- ella, B. verna and B. moseri. Three years after the Robinson & Schrenk transfer, Robinson described Bartonia iodandra from Newfoundland and promoted a rash of investigation by Fernald. In 1921 Fernald described B. iodandra var. sabulonensis from Sable Island, Nova Scotia, and in the following year he transferred both B. iodandra and his own variety sabulonensis to the status of varieties of B. pani- culata. Following a tremendous amount of field work in Nova Scotia, Fernald found so many specimens that could not be placed in either B. paniculata or var. todandra that he created a variety intermedia to take care of them. Except for a form of B. virginica 1959] Gillett — Bartonia and Obolaria 45 described by Victorin in 1919, the genus has remained relatively untouched until now. Chromosome counts have been made by Rork (1949) of Bar- tonia paniculata from her own preparations and a count was obtained for B. virginica by the same worker from a slide supplied by Uhl. A count of n=26 was obtained for both species. From material collected in the field by the author at St. Pierre de Howick, Quebec, R. J. Moore of this Division obtained a count of 2n=ca. 52 from a squash of ovary tissue. This material, of course, was B. virginica. No count has been obtained from New- foundland or Nova Scotian material. No chromosome count has been recorded for Bartonia verna. Flowering and fruiting dates taken from herbarium specimens are presented in Table 1. These are actually collection dates but since the species flower for long periods they are indicative of the flowering time for cach population. Evidently these dates coin- cide for B. virginica and B. paniculata but differ widely for B. verna. B. verna exhibits the typical distribution of a southern coastal plain species. B. paniculata and B. virginica are sympatric over a large portion of their ranges. The former extends further west in the southern portion of the coastal plain and further north in the northern portion than the latter. B. virginica on the other hand, tends to extend farther inland into the Great Lakes region. The disjunction in distribution between Maine and Nova Scotia can be explained by post-glacial flooding of the coastal plain so that migration northward must have followed in the wake of the receding glacier before this area was inundated. ‘This theme has been discussed by Fernald. Morphologically B. verna is a clear-cut species while the others are not. The paniculata and iodandra populations intergrade quite freely; occasional intermediates occur between B. paniculata and B. virginica. Although the two species are sympatric over a 46 Rhodora [VOL. 61 TABLE I. PHENOLOGY or BARTONIA Month Week B. virginica B. paniculata B. verna ssp. panic. ssp. iodandra July l FL — FL — 2 FL -— -— — 3 FI — -— — 4 FL FL -— — August l FL FL FL — 2 FLR FL FL — 3 FLR FL FL — 4 FLR FL FLR — September l FLR FL FLR — 2 FLR FL FLR — 3 FLR FLR FR — 4 FLR FLR FR -— October ] FR FR FR —- 2 FR — FR -— 3 FR FR — — 4 FR — — November l FR -— -- 2 FR — — FL 3 — — — FI 4 — — — — December l — — —- — 2 — — — — 3 — — — — 4 — —— — FI January l — — — FI 2 — — — — 3 — — — I 4 a ao — FL February 1 — — — FL 2 — — -—— FL 3 — — — — 4 — — — — March l — — — FL 2 — — -—- FLR 3 — — -- FLR 4 — — — — April 1 — — — FR 9 — — — — 3 — — — — 4 — — — — FL — in flower; ER — in fruit; FLR — in both flower and fruit. 1959] Gillett — Bartonia and Obolaria 47 considerable portion of their range, intergrades are more or less restricted to the coastal plain area. On the other hand the two species frequently remain quite distinct in the same region, due perhaps to restricting habitat factors. Herbarium sheets bearing a number of plants may contain occasional intermediates. These plants may show B. virginica-type flowers or fruits with B. pani- culata-type leaf arrangement or other character combinations. The following specimens may serve to illustrate: Bissell & Graves 22294, Port Mouton, N. S. — somewhat intermediate between the vir- ginica and the iodandra type; Fernald & Long 24356, Shelburne, N. S.— most are B. virginica but a few have alternate leaves; Torrey Herb. N. Y. — a mixture: #1 and #3 sterile intermediates, #4 B. virginica, #5 B. paniculata; Seymour 1996, Martha's Vineyard — 3 of 4 plants B. vir- ginica, the other intermediate; Tracy s.n., Ocean Springs, Mississippi — some plants of B. paniculata approaching B. virginica; Chapman Herb. Apalachicola, Florida — plant #1 intermediate, #2 B. paniculata, #3 and #4 B. virginica; Ferguson 5208, Speonk, N. Y. — B. virginica, some approaching B. paniculata. However, the following sheets have both species on the same sheet with no intermediates: The mo sheet of the Gray Exsiccatae 390 — B. vir- ginica but one plant of the iodandra group; Lighthipe s.n., Ocean Beach, N. J. — left-hand plant B. paniculata, right hand plant, B. vir- ginica; Chickering s.n., Kennebunk, Maine — 1 plant of B. virginica, 3 of B. paniculata. Only a selected number of examples are given here. With both species having similar chromosome numbers and, according to Rork’s sketches, similar chromosome morphology and with similar flowering times and sympatric distribution, it would seem likely that hybridization is within the realm of possi- bility. Since the actual percentage of such intermediates is low with respect to number of collections and extremely low with respect to the total number of plants examined, I feel that the species are best treated as separate at least until such time as mass sampling can be made and crossing experiments carried out to clarify the situation. Species relationships would be placed on a firmer basis by further cytological studies. Obviously a chromosome count is immediately required for B. verna. Morphologically, the poorly developed corolla tube, decurrent stigmas and variable position of leaf scales would suggest that this species is the more primitive; 48 Rhodora [VOL. 61 the distinct corolla tube, distinct stigmas and definite position of leaf scales would indicate that the other two species are more advanced. Of the two it is difficult to set one above the other. It would appear that the northern subspecies of B. paniculata has evolved by selection from occasional crosses and backcrosses between B. virginica and B. paniculata. ‘The presence of pointed anthers in the northern group rather suggests this. A biometrical study employing anther shape as a character may lend support to this idea. This study has been based on the collections of most of the eastern North American herbaria. ‘The smaller herbaria were added in order to try to fill out the range of the species in more detail. “Thanks are due to the curators of the herbaria listed and to those of the British Muscum of Natural History and the Paris Museum for their assistance in obtaining photographs of type material. I should like to thank Dr. B. Boivin of this Division for reading the manuscript. Material from the following herbaria has been seen: DAO, GH, NY, CAN, TRT, US, TENN, MO, OKLA, FSU, GA, TEX, NCU, FLAS, SMU, BUS, NEBC. Although a formidable list, this does not represent a very large number of specimens. Material of the genus Obolaria is included. Selected representative specimens of each species are cited and a few selected intermediate sheets have been discussed. For each entity a list of the counties per State or Province is given to sup- plement the distribution maps. The publication of an Index of Exsiccatae is impractical. SYSTEMATIC “TREATMENT Bartonia Mühl. ex Willd. Ges. Naturf. Freunde Berlin Neue Schr. 3: 444. 1801, nomen conserv. Agina Neck. Elem. Bot. 2: 153. 1790, nom. rejic. Centaurella Michx. Fl. Bor. Amer. 1: 97, 98 t. 12. f. 1 & 2. 1808. Centaurium Pers. Synops. I: 137. 1805, non Hill, 1756. Andrewsia Spreng. Syst. 1: 368 & 428. 1825; non Andreusia Vent., 1804. Filiform, erect saprophytic annuals. Leaves reduced to minute, oppo- 1959] iullett — Bartonia and Obolaria 49 site or alternate subulate scales. Inflorescence cymose or racemiform, frequently I-flowered. Calyx with a short tube or the lobes nearly free, the outer overlapping the inner. Corolla deeply 4—lobed, marcescent, campanulate, each lobe with a single vascular strand. Stamens 4, alter- nate with the lobes of the corolla, the filaments short, attached at the sinuses of the corolla lobes, the dart-shaped to blunt anthers introrse and frequently deciduous after anthesis. Ovary sessile, unilocular, bi- carpellate, the placenta parietal, the ovules covering the entire inner surface. Fruit an ovoid, thin-walled capsule, dehiscent along the sutures. Seeds very numerous, minute, ellipsoid, smooth to minutely reticulate. —Type species: B. tenella Mühl. ex Willd. KEY TO THE SPECIES A. Aestival to autumnal flowering. Corolla slightly longer to twice as long as the calyx, the lobes 2-4 mm. long, lanceolate to oblong. B. Leaf scales essentially opposite, the numerous nodes progressively more crowded towards the base; corolla lobes oblong, the apex apiculate, erose to entire; capsule dehiscing below the elongate style . m eene ene eee l. B. virginica. BB. Leaf scales essentially alternate, the few nodes but slightly “closer towards the base; corolla lobes lanceolate, the apex inapiculate, entire, acute; capsule dehiscing by terminal separation of the short style . 2. B. paniculata. XX. Vernal flowering. Corolla three times m length of the calyx, the lobes 6-11 mm. long, obovate to spatulate eee. 3. B. verna l. Bartonia virginica (I.) BSP. Prel. Cat. N.Y. Pl. 36. 1888 Sagina virginica L. Sp. Pl. 2: 128. 1753. (Clayton 619 BM type, photo DAO!). Bartonia tenella Mühl. ex Willd. in Ges. Naturf. Freunde Berlin, Neue Schrift 3: 444. 1801. (Muhlenberg s.n. B. Willd. Herb. 2991 type, photo pao!). Centaurella autumnalis Pursh, Fl. Amer. Sept. 1: 100. 1814, pro parte typ. incl. based on Sagina virginica L. Andrewsia autum- nalis (Pursh) Spreng. Syst. Veg. 1: 428. 1825. Centaurella moseri Steudel & Hochstein ex Griseb. Gen. et Sp. Gent. 308. 1839. Centaurella autumnalis Pursh f) brachysepala Griseb. Gen. et Sp. Gent. 308. 1839, ex char. (Drummond s.n. type presumably London). Bartonia moseri (Steud. & Hochst. ex Griseb.) Robins. & Schrenk ex Gilg in Engl. & Prantl Nat. Pflanzef. 4, abt. 2: 76. 1895. Bartonia virginica (L.) BSP. forma abortiva Vict. in Roy. Soc. Can. Proc. & Trans. Ser. III. Sect. V. 13: 113. 1919. (Victorin 19570 MT type!). Annuals 0.4-4.5 (av. 2.2) dm. tall, erect, simple or branched above, the stem slender, wiry or stout, terete or angled, occasionally slightly winged, frequently twisted, the nodes numerous and becoming progres- sively closer together towards the usually purple base. Leaves 0.9—4.7 (av. 2.4) mm. ouo: scale-like, subulate, decussate, usually opposite or subopposite, the lowermost occasionally alternate. Flowers in simple 50 Rhodora [VOL. 61 terminal cymes with solitary flowers or branches in the lower axils. Calyx lobes 0.4-1.1 (av. 0.74) mm. wide, 2.0-3.5 (av. 2.7) mm. long, almost distinct, subulate-lanceolate, carinate. Corolla 2.4—4.2 (av. 3.4) mm. long, greenish-yellow, the petals united in the lower third, the lobes 0.75-1.2 (av. 1.1) mm. wide, oblong, obtuse, mucronate, erose or entire, with a prominent nerve. Stamens 1.5-3.5 (av. 2.55) mm. long (measured from the base of the corolla), flattened, the anthers 0.6-1.2 (av. 0.91) mm. long, oblong, mucronate, frequently purple. Pistil 2.5-4.5 (av. 3.6) mm. long, ovate, the style 1.0-2.0 (av. 1.3) mm. long, the stigmas decurrent. Capsule 4.0-5.5 (av. 4.8) mm. long, ovate, dehis- cence below the persistent style. Seeds averaging 0.14 x 0.08 mm., ex- tremely numerous, light brown with a testa of irregular cells. Corema barrens, sphagnous bogs, margins of swamps, peaty and sandy lake shores, openings in brush and in dry thickets. Flowering from early July until late September; fruiting from early August until November. Common Names: Virginian Bartonia; Virginian Screwstem. Representative specimens: N.S. Queens County, Louis Lake, Port Joli, Gray Exsiccatae 390, Fernald, Linder & Long s.n. (DAO, GA, GH, MO, TENN, NY, TEX, TRT, US). Que. Terrebonne County, Ste. Therese, Victorin, Rolland & Meilleur 45750 (pao, Gu, us). N.Y. Cattarangus County, Steamberg, Alexander & House 12979 (GH, Us). Va. Northampton County, Eastville, Fernald & Long 51/20 (MO, NY, us). Wisc. Brown County, Shuette s.n. (NY, US). Distribution by Counties: CANADA. NOVA SCOTIA: Annapolis, Lunenberg, Queens, Shelburne and Yarmouth, QUEBEC: Chambly, Chateauguay, Hunting- don and ‘Terrebonne. UNITED STATES. ALABAMA: Lee. CONNECTICUT: Fast Haven. Hartford, New Haven, New London, Southington and Stratford. DELAWARE: Sussex. DISTRICT OF COLUMBIA, FLORIDA: Brevard, Gulf, Nassau and Osceola. GEORGIA: Bartow, Chatham, Jenkins and Miller. 1nptana: Lake, Por- ter and Steuben. LOUISIANA: St, Tammany. MAINE: Cumberland, Hancock, Oxford and York. MARYLAND: Anne Arundel, Cecil, Frederick, Garrett, Prince Georges, Queen Annes and Worcester, MASSACHUSETTS: Barnstable, Dukes, Essex, Hampden, Middlesex, Nantucket, Norfolk, Plymouth, Suffolk and Worcester. MICHIGAN: Berrien, Kalamazoo, Oakland, Presque Isle, St. Clair, Van Buren and Washtenaw. Mississippi: Harrison and Jackson. NEW JERSEY: Atlantic, Burlington, Camden, Cape May, Cumberland, Essex, Hudson, Mid- dlesex, Monmouth, Morris, Ocean, Passaic and Sussex. NEW YORK: Catta- raugus, Cayuga, Herkimer, Madison, Nassau, Oneida, Ontario, Oswego, Put- nam, Richmond, Seneca, Suffolk, Tompkins, Warren, Washington, Wayne and Westchester. NORTH CAROLINA: Chowan, Dare, Durham and Macon. onto: Cuyahoga, Jackson, Licking, Lucas, Summit and Wood. PENNSYLVANIA: Cen- ter, Chester, Lackawanna, Lancaster, Monroe, Philadelphia and Pike. RHODE ISLAND: Newport, Providence and Washington. soUTH CAROLINA: Aiken, Darlington and Kershaw. TENNESSEE: Coffee, Cumberland, Marion, Morgan and Sequatchee. VERMONT: Chittenden. vrRGINIA: Accomac, Augusta, Bath, Dinwiddie, Fauquier, James City, Nansemond, Norfolk, Northampton, Prin- 1959] Gillett — Bartonia and Obolaria 51 cess Anne, Southampton and Sussex. WEST VIRGINIA: Preston, Raleigh, Randolph and Tucker. wisconsin: Brown, Jackson, Munroe and Vilas. MAP 1. Distribution of Bartonia virginica. MAP 2. Distribution of Bartonia panicu- lata: solid dots: ssp. paniculata. open circles: ssp. iodandra. crosses: putative intermediates between B. virginica and B. paniculata. 52 Rhodora [VOL. 61 The original description of Centaurella moseri gave “Cymis racemiformibus, ramis alternis", and "foliis non oppositis!" Other characters given seem to fit B. paniculata quite closely. ‘The specimens cited are: "Salzburg ‘Township Pennsylvaniae (Moser!); Covington. (Drummond!). — Fl. Septembri!-”. Grisebach gave the source of the material as "ab Union. Wür- temburg," which may mean that the holotype material is at Stuttgart if it still exists. However, there is a collection at the New York Botanical Garden labelled, “Salzburg Taunship Penn- sylvaniae, Unio itiner C. J. Moser Jul. Aug. 1832." This has both a printed and handwritten label. ‘There is a duplicate of this at the Missouri Botanical Garden. In addition to these sheets ” I have seen the Drummond specimen from Covington, Louisiana. Both collections are undoubtedly B. virginica in spite of the de- scription, but Grisebach stated so clearly the characters of his plant that it is quite possible that he had a mixed collection. Until the type comes to light it is pointless to select a lectotype. * For the present I am placing C. moseri provisionally in the synon- ymy of B. virginica. 2. Bartonia paniculata (Michx.) Muhl. Cat. 16. 1813. Centaurella paniculata Michx. Fl. Bor. Am. 1: 98. f. 1. 1803. (Michaux s.n. P type, the sheet from the Richard Herb. (see explanation below), photo pao!). Annuals, 0.3-3.9 dm. tall, the green or purple angled stems strict to twining, simple or branched above with curved ascending branches. Leaf-scales 0.5-3.0 mm. long, alternate, subulate. Flowers solitary or in simple cymes with or without axillary ones below. Calyx tube evident or the frequently crinkle-keeled, ovate to lanceolate lobes barely united at the base, the lobes 1.5-3.2 mm. long, 0.5-1.1 mm. wide, the outer lobes shorter and wider than the inner. Corolla white to somewhat erubescent, the lobes oblong to tapered oblong, acute to slightly mucro- nate with a single nerve or with two lateral ones sometimes evident. Stamens 2.2-5.0 mm. long (measured from the base of the corolla), the anthers varying from 0.3-0.6 (av. 0.4) long in the typical subspecies to 0.5-1.0 (av. 0.7) mm. long in the northern subspecies, usually with a rounded apex. Pistil ovate, with an evident style, the stigmas decur- rent along the whole length. Capsule ovate, the style persistent, dehis- cence apical by separation of the stigmas. Seeds very small, light brown with a testa of inflated cells, extremely numerous. 1959] Gillett — Bartonia and Obolaria 53 I select the Michaux collection from the Richard herbarium in the main collection at Paris as the holotype of Centaurella paniculata Michx. The first and last lines of the handwritten label are in Richard's handwriting; the central portion is in Michaux's hand. The separate Michaux collection labelled C. paniculata I was surprised to find is perfectly good B. virginica. KEY TO THE SUBSPECIES OF B. PANICULATA A. Plants slender, lax, frequently twining, green or essentially so; anthers usually yellow, about 0.5 mm, long, rounded at the apex. 2a. B. paniculata ssp. paniculata. AA. Plants strict or stout, the stem frequently thickened upwards and pur- ple throughout; anthers usually purple or if yellow, the filaments fre- quently purple, 0.51.0 mm. long, rounded or apiculate. 2b. B. paniculata ssp. iodandra 2a. Bartonica paniculata (Michx.) Muhl. ssp. paniculata Centaurium autumnale Pers. Syn. Pl. 1: 137. 1805, ex char. Bartonia tenella Muhl. ex. Willd. p brachiata Wood, Class-book, ed. 2. 586. 1866, ex char. excl. syn. Bartonia lanceolata Small, Fl. SE. U.S. 932, 1336. 1903. (Chapman s.n. NY, type!) . Barionia paniculata (Michx.) Robinson in RHODORA 10: 35. 1903, superfluous comb. Plants 1.0-3.9 (av. 2.5) dm. tall, the slender usually green but rarely purple stem erect or twining. Flowers in simple cymes and with axillary flowers below. Calyx lobes slightly united at the base, frequently crinkle- keeled, the sinuses acute, the lobes 1.5-2.8 (av. 2.1) mm. long, 0.5-1.0 (av. 0.75) mm. wide, ovate-lanceolate to lanceolate. Corolla 3.0-5.0 (av. 3.8) mm. long, white, the tube 14-14 the length of the corolla, the tapered oblong lobes 1.0-1.5 (av. 1.1) mm. wide, acute. Stamens 2.2-3.8 (av. 2.8) mm. long (measured from the base of the corolla) , the yellow anthers 0.3-0.6 (av. 0.4) mm. long, the tips rounded. Pistil 2.0-5.0 (av. 3.3) mm. long, the style 0.5-1.5 (av. 0.9) mm. long. Capsule 3.5-5.8 (av. 4.2) mm. long. Seeds averaging 0.19 x 0.12 mm., 1000 to 1500 per capsule (actual count). Sandy and sphagnous bogs, grassy swamps, cedar swamps, wet open woods, sandy and peaty shores of ponds. Flowering from late July until late September, fruiting from mid-September until November. Common Name: Screwstem. Representative specimens: Mass. Nantucket Island, Bicknell s.n. (Nv). N. J. Middlesex County, Cheesequake State Park, Alexander & Gilly 428 (Ny). N. Y. Suffolk County, Bridgehampton, Ferguson 6277 (NY). Tenn. Coffee County, Manchester, Svenson 9124 (cH). Va. Greenville County, Dahlia, Fernald & Long 9103 (cu, Ny, US). 54 Rhodora [VOL. 61 Distribution by counties: ALABAMA: Mobile. ARKANSAS: Hempstead and Pulaski. CoNNecricut: Hartford, Litchfield. DELAWARE: Sussex. DISTRICT OF COLUMBIA, FLORIDA: Duval, Okaloosa, Putnam, Santa Rosa and Walton. GEORGIA: Camden, Chatham, Fannin, Randolph, Screven, Sumter, Talbot. KENTUCKY: Laurel. LOUISIANA: Rapides. MARYLAND: Prince Georges. MASSA- CHUSETTS: Barnstable, Duke and Nantucket. Mississippi: Harrison and Simp- son. NEW JERSEY: Atlantic, Burlington, Camden, Cape May, Gloucester, Mid- dlesex, Monmouth, Ocean and Salem. NeW york: Kings, Nassau, Queens, Suffolk. NORTH CAROLINA: Brunswick. OKLAHOMA: Leflore. PENNSYLVANIA: Location unknown, TENNESSEE: Carroll, Coffee, Fentress and Grundy. TEXAS: Houston. VIRGINIA: Caroline, Greenville, Nansemond, Prince George, South- ampton and Sussex. 2b. Bartonia paniculata (Michx.) Muhl. ssp. iodandra (Robins.) stat. nov. Bartonia iodandra Robins. in Bot. Gaz. 26: 47. 1898 (Robinson & Schrenk 5 GH type, isotypes Mo, Ny, us!) . Bartonia paniculata (Michx.) Muhl, var. codandra (Robins.) Fern. in Ruopona 23: 288. 1922. Bartonia todandra Robins. var. sabulonensis Fern. in Proc. Bost. Soc. Nat. Hist. 36: 89. 1921. (St. John 1307 cn type, isotypes Ny, us!). Bartonia pani- culata var. sabulonensis Fern. in Ruopora 23: 288. 1922. Bartonia paniculata var. intermedia Fern. in RnopoRgA (23: 287. 1922. (Fernald & Long 22299 cn type, isotype NY. Plants 0.3-2.5 (av. L2) dm. tall, the usually purple angled stems occasionally stout. Flowers in simple cymes or solitary or with axillary flowers below. Calyx tube frequently well-developed, slightly winged at the base, the sinuses rounded, the lobes 1.5-3.2 (av. 2.4) mm. long, 0.5-1.1 (av. 0.8) mm. wide, ovate to lanceolate. Corolla 3.0-6.2 (av. 4.4) mm. long, white to somewhat erubescent, the tube shorter to about equal to the lobes, the oblong lobes variable, 0.8-2.0 (av. 1.3) mm. wide, acute to slightly mucronate with usually an evident nerve. Stamens 2.4-5.0 (av. 3.6) mm. long, the frequently purple filaments broadened below, the usually purple anthers variable with or without an apiculate tip. Pistil 3.0-5.5 (av. 3.9) mm. long, the style 0.6-1.5 (av. 0.95) mm. long. Capsule 4.0-6.0 (av. 5.0) mm. long. Seeds averaging 0.15 x 0.10 mm. Wet peaty barrens, sphagnous bogs, peaty margins of lakes and boggy depressions. Flowering from about mid-July to mid-September; fruiting from September to late October. Representative specimens: Nfld. Head of Exploits River system, Fernald & Wiegand 6082 (Gu, NY); Avalon Peninsula, Fernald & Long 26987 (GH, NY). N. S. Cape Breton County, Gabarus, Rousseau 35627 (Gu) ; Digby County, Central Grove, Gray Exsiccatae 388 (GH, GA, DAO, NY, TENN, US). 1959] Gillett — Bartonia and Obolaria 55 Distribution by counties and districts: sT. PIERRE & MIQUELON: Island of St. Pierre; Island of Miquelon. CANADA. NEW BRUNSWICK: Charlotte. NEW“ FOUNDLAND: Green Bay, Burgeo and La Poile, Harbour Main-Belle Isle, Hum- ber, Gander Falls, St. Barbe, St. Marys and St. Georges-Port au Port Districts. NOVA scoTIA: Cape Breton, Cumberland, Annapolis, Digby, Guysborough, Halifax, Inverness, Lunenburg, Pictou, Queens, Richmond, Shelburne, Yar- mouth and on Sable Island. MASSACHUSETTS: Barnstable and Nantucket. NEW YORK: Suffolk. RHODE ISLAND: Newport. VIRGINIA: Greensville. 3. Batonia verna (Michx. Muhl. Cat. 16. 1813. Centaurella verna Michx. Fl. Bor. Amer. 1: 98. t. 12. f. 2. 1808. (Michaux s.n. v, type the sheet from the Richard Herb. in the general herbarium, photo pao!). Centaurium vernum (Michx.) Pers. Syn. Pl. I: 137. 1805. Centaurella vernalis Pursh, Fl. Amer. Sept. 1: 99. 1814, ex char. Centaurella vernalis g uniflora Pursh. l.c. 100, ex. char. Andrewsia verna (Michx.) Spreng. Syst. Veg. 1: 428. 1825. Annuals, 0.3-2.3 (av. 1.3) dm. tall, erect, simple, rarely branched above, the stem slender, angled but not winged, with few nodes, fre- quently purple. Leaves 0.6-3.0 (av. 1.9) mm. long, those subtending the upper flower to 3.5 mm. long, scale-like, usually opposite or sub- opposite, lanceolate to subulate. Flowers solitarv, in terminal cymes, or with additional axillary flowers. Calyx lobes united only at the base, the lobes 0.9-2.8 mm. long, up to 1.5 mm. wide, lanceolate. Corolla white, the petals united only at the base, the lobes, 4.8-8.0 (av. 5.8) mm. long, 1.5-3.5 (av. 2.4) mm. wide, widely spreading, spatulate to obovate, obtuse, entire, with a single nerve. Stamens 3.0—4.8 (av. 3.6) mm. long, the filaments about 0.5 mm. wide at the base, flattened, tapering towards the apex, the anthers 0.5-1.1 (av. 0.8) mm. long, oblong, slightly in- curved, dorsifixed, occasionally contorted after dehiscence. Pistil to 7.5 mm. long, ovate, the style about 2 mm. long, stout, the stigmas decur- rent as far as the upper part of the ovary. Capsule ovate, slightly shorter than the marcescent corolla, the style persistent, dehiscence septicidal in the center of the capsule only. Seeds averaging 0.15 x 0.12 mm., light brown with a testa of round inflated cells, extremely numerous, over 1000 per capsule. Moist pine barrens, sandy soil, edge of ponds, ditches and grassy depressions. Flowering from the first week of November until the first week of April. Seldom collected in mature fruit. Common Names: White or Spring Bartonia. Representative specimens: N.C. Brunswick County, Southport, God- frey 48001 (pao, Ny, Us). S.C. Beaufort County, Fluffton, Mellichamp in 1875 (ny). Ala. Damp pine barrens, Mohr in 1898 (us). Fla. Bra- denton, Tracy 7540 (DAO, GH, NY, vs) ; Hibernia, Canby in 1869 (cn, NY, US). 56 Rhodora [VOL. 61 Distribution by counties: ALABAMA: Mobile. FLorma: Alachua, Brevard, Broward, Clay, Columbia, Duval, Franklin, Glades, Gulf, Hardee, Highlands, Hillsborough, Lake, Lee, Levy, Manatee, Orange, Osceola, Pasco, Pinellas, Polk, Putnam, Santa Rosa, Seminole, St. Johns, Volusia, Wakulla and Walton, GEORGIA: Camden and Charleston. Louisiana: Orleans and Tangipahoa. MISSISSIPPI: Harrison. NORTH CAROLINA: Brunswick. SOUTH CAROLINA: Dor- chester and Jasper. 7 ^ TF — iF J ( 5 Ta — x 8 ^ S cj , | | te TU N yo (A LE ) i N \ "n AA MEM um Y N ? | 1 J \ | NL ) a j "UE p^ l ( n \ 1 3 | 4.4.-.- = y | x | id J t »* i | P. \ -! / : * f {| F / ^ n A U / \ N 2 à 1 AAO A * | ` | | f — — T * | SN S tr . isi N wWocl--fR4- 3 \ ) F —— Aer J 4 a * . p "n \ . 5 1 uA ( „5 o rd =A ^ E { > LoT T \ 1 N JS NI S f 7 \ f Vv i C / \ N Y i H Y I an NM [^ . ' Jo J y — r . ! N ) \ | E J f ? 1 | N * i — 9 JA y aH FON \ | I j] | , Y |y s ] \ I KOJ. 7 * I / | \ Y . N / | ) —.—. 4 Ji N — i A e. pL - ` | 4 S c4 ig vt SUNY É \ N J SORS e 9^ | eo? \ | | A | J : V | SCALE i ai 3 — "um eH | 0 MILES 400 * ENSE | | —— ane MAP 3. Distribution of Bartonia verna. There are two sheets of Centaurella verna Michx. in the Paris herbarium. I select the Richard collection in the General Herba- rium as the type rather than the separate Michaux collection be- cause it is more like the figure in the Flora Boreali-Americana (t. 12. f. 2.). Actually it is of little consequence which sheet is 1959] Gillett — Bartonia and Obolaria 07 selected for they are both the correct plant but the selection should be made in order to establish a single sheet as the type. DOUBTFUL AND EXCLUDED SPECIES BARTONIA ALBESCENS Gill. & Arn. in Edinb. Geogr. Jour. 2 974 Vegy =Nuttallia albescens (Gill. & Arn.) Standl. in Jour. Wash. Acad. Sci. 6: 239. 1916. BARTONIA AUREA (Nutt. Lindl, Bot. Reg. 22; t. 1831. 1836. — Mentzelia aurea Nutt. Gen, Am. Pl. BARTONIA DECAPETALA Sims Bot. Mag. 36: t. 1487, 1812 (incorrectly attributed to Pursh by authors). =Mentzelia decapetala (Sims) Urb. & Gilg in Engl. & Prantl, Nat. Pflanzenf. 3. Abt. 6a: 111. 1894. BARTONIA LAEVICAULIS Dougl. ex. Hook. Fl. Bor.-Am, 1: 221, t. 69. 1834. — Mentzelia laevicaulis T. & G., Fl. N. Am. 535. 1840. BARTONIA MICRANTHA (T. & G.) Hook. & Arn. in Bot. Beech. Voy. 343. t. 84. 1841. — Mentzelia micrantha T. & G., Fl. N. Am. 535. 1840. BARTONIA MULTIFLORA Nutt. in Jour. Acad. Sci. Phil. n.s, 1: 180. 1847. — Mentzelia multiflora (Nutt) A. Gray in Mem. Am. Acad, n.s. 4: 48. 1849. BARTONIA NUDA Pursh, Fl. Am. Sept. I: 328. 749. 1811. — Meníizelia nuda (Pursh) T. & G., Fl. N. Am. 535. 1840. BARTONIA ORNATA Nutt. Gen. N. Am. PI. 1: 297. 1818. — Mentzelia ornata (Nutt) T. & G., Fl. N. Am. 534. 1840. BARTONIA PARVIFLORA Dougl. ex Hook. Fl. Bor.-Am, 1: 221. 1845. — Touterea parviflora (Dougl. ex Hook.) Rydb. in Bull. Torr. Bot, Cl. 80: 276. 1903. BARTONIA PUMILA (T. & G.) Nutt. ex Jackson, Ind. Kew. 276. 1895. — Mentzelia pumila T. & G., Fl. N. Am. 1: 535. 1840. BARTONIA SINUATA Presl. Rel, Haenk. (no other known combination in the Loasaceae). BARTONIA WRIGHTH (Grav) Gray ex Jackson, Ind. Kew. 276. 1895. — Mentzelia wrightii ^. Gray, PI. Fendl. 48, 1849. CENTAURELLA AFSTIVALIS Pursh, Fl. Am. Sept. 1: 100. 1814. The type has not been located and the description is inadequate to establish identity. This name is later than any employed in this treatment. IIl. THE GENUS OBOLARIA L. Although this monotypic genus has no specific problem, it seemed desirable to assemble the history and taxonomy because there has existed in the past a conflict over the family position and also because this information may assist in an understanding of generic relationships. In the following study, the range has been established, the type specimen sought out and the descrip- 58 Rhodora [VOL. 61 tion amplified. A separate generic description has been drawn up which will emphasize those characters considered to be com- parable with other gentianaceous genera. No direct relationship with Bartonia is implied by this arrangement. Obolaria was described by Linnacus in his Hortus Cliffortianus in 1737 under the spelling "Obularia" but was later redescribed and respelled Obolaria in the Species Plantarum and in the sixth edition of the Genera Plantarum. According to Article 74 (1) of the International Code, 1956, the latter spelling should be re- tained as the preferred spelling of Linnaeus. Linnaeus considered that he was describing a relative of the genus Orobanche for in his Mantissa he wrote: "folia suprema extus purpurea. Flores pallide rubentes. Orobanches affinis.” The unfortunate association with the Orobanchaceae begun by Linnaeus was continued by later authors, particularly by Grisebach, who ignored the genus completely in his treatment of the Gentianaceae in De Candolle’s Prodromus. In a later volume of the work, however, Reuter included Obolaria in his treatment of the Orobanchaceae. Griscbach also omitted Obolaria from his Species Gentianearum. George Don placed it in the tribe OBOLARIEAE, the "terrestrial, not parasitical" Orobanchaceae. The species was assigned to the genus Shulizia by Rafinesque in 1808. The earlier pre-Linnacan Obolaria of Siegesbeck was taken up for Linnaea L. of the Caprifoliaceae by Otto Kuntze in 1891 in his Revisio Generum Plantarum. In the same volume Kuntze adopted Rafinesque's name for the plant, recognized its position in the Gentianaceae and pointed out the need to dis- tinguish it from Schulzia Spreng. of 1813 (Umbelliferae) named after Professor I. H. Schulz of Halle. Acording to Kuntze, Sprengel's name had been spelled "Schultzia" by many authors. Gilg, in Die Pflanzenfamilien assigned Obolaria to the Gen- tianaceae sublamily Gentianoideae, under the group Gentianeae- Erythraeinae of unspecified rank. Gilg’s grouping is a somewhat heterogeneous one having Enicostemma-like pollen as the chief feature in common. Of Gilg’s list, Sabatia is the only genus that has been recently monographed, Lapithea being merged with it 1959] Gillett — Bartonia and Obolaria 59 (Wilbur, 1955). Beck-Mannagetta in his thorough treatment of the Orobanchaceae in Das Pflanzenreich (1897), omitted the genus Obolaria from the family and made no reference to it. The uncertainty of family position has stemmed from Lin- nacus’s description and from the gross appearance of the plant. Although most of the Orobanchaceae are parasitic, there is no definite evidence that this is true for Obolaria. The following characters will serve to establish a Pennywort as a member of the Gentianaceae: OBOLARIA OROBANCHACEAE Placentae from the sutures with Placentae non-sutural; 2 lateral l lateral lamella per carpel (2) branched lamellae per carpel (2-3) Plants glabrous Plants frequently pubescent Autophytic or saprophytic with Parasitic without chlorophyll evident chlorophyll Actinomorphic corolla Zvgomorphic corolla with label- late lobes Opposite leaves or bracts Alternate leaves or scales Stamens equal, 4 rarely 2 by abor- Stamens didynamous; one anther tion of the opposing pair; anther sac frequently aborting. sacs normally developed. SYSTEMATIC TREATMENT Obolaria L. Sp. Pl. 2: 632. 1753, non Obolaria Siegesb. Prim. Fl. 79. 1736, ex Ktze. Rev. Gen. 275. 1891 = Linnaea L. (Caprifoliaceae) . Shultzia Raf. Med. Repos. II. 3: 422. 1806, nom.; 5: 356. 1808 descr. & Jour. Bot. (Paris) 1: 219. 1809. Perennial herbs with mycorrhizal fleshy corolloid bitter roots. Leaves opposite. decussate, sessile. Flowers in terminal dichasia or simple cymes and axillary. Calyx of two free sepals, usually with two decussate leafy bracts except in inner flowers of congested inflorescences, the sepals similar to the cauline leaves but more slender. Corolla 4—lobed, cam- panulate to urceolate, delicate but marcescent, without internal fimbriae or glands. Stamens 4, rarely 2, inserted at the sinuses of the corolla lobes, frequently appearing didynamous by partial abortion of the opposing pair, anthers 2-celled. introrse. Ovary unilocular, the parietal placentae sutural and with an additional lamellate intrusion on each of the two carpel walls. Stigmas two. Capsule unilocular, thin-walled, rupturing irregularly rather than by a terminal separation of the car- pels. Seeds numerous. 60 Rhodora [VOL. 61 I. Obolaria virginica L. Sp. Pl. 1: 632. 1753. (Clayton 286 BM, TYPE, photo pAo?). Shultzia obolarioides Raf. Med. Repos. II. 5: 356. 1808. Shultzia virginica (L.) O. Ktze. Rev. Gen. 2: 430. 1891. Herbs 0.4—1.7 dm. tall, simple or strictly branched, the stem stout, fleshy and somewhat angled, frequently stouter above than below, occa- sionally fasciated. Lower cauline leaves 0.4-1.5 cm. long, 0.3-1.0 cm. wide, lunate, scale-like, loosely appressed, upper leaves becoming foliace- ous or bract-like, closely subtending the flowers, fleshy, oblong to ob- liquely spatulate or soleaeform, truncate, abruptly acuminate, with numerous dichotomous veins anastomosing towards the tip, yellow to purple-green. Flowers short-pedicellate, the pedicels 0.5-2.0 mm. long. Corolla 0.6-1.2 cm. long (av. 0.9), white, campanulate with the lobes ascending, the tube 2.0-4.5 mm. long (av. 4.0), the lobes 4.5 mm. long, 2 mm. wide, oblong to slightly obovate, acute or strongly erose or ir- regular with entire margins, with three veins dichotomizing towards the apex. Stamens included, the short filaments 0.25 mm. wide, 1.0 mm. long, flattened, slightly tapered, the anthers 0.5 mm. long, dorsifixed, the thecae facing sideways after dehiscence. Pistil sessile, the ovary 3.0-3.5 mm. long, somewhat flattened, the style 1.0 mm. long, straight, the two stigmas 0.25 mm long, oblong and spreading. Capsule 5.5 mm. long, 3.5 mm. wide, ovoid, somewhat angular. Seeds averaging 0.23 mm. long, 0.16 mm. wide, light brown, ovoid, the surface minutely striate under X 60 magnification, approximately 1600 per capsule. Rich woods, bottom lands, river banks, shrubby slopes, but chiefly in deciduous forest. Flowering from early March until late May; fruit- ing in late May and early June. Distribution by Counties: ALABAMA: Blount, De Kalb, Franklin, Jackson, Tuscaloosa. ARKANSAS: Phillip. prLAWARE: New Castle. DISTRICT OF COLUM- BIA. FLORIDA: Jefferson, GEORGIA: Dade, Fulton, Gwinnett, Rabun. ILLINOIS: Pulaski. INDIANA: Brown, Greene, Knox, Posey. KENTUCKY: Bullitt, Edmon- son, Whitley. LOUISIANA: East Feliciana. MARYLAND: Baltimore. NEW JERSEY: Essex, Mercer, Somerset, Sussex, Union. NORTH CAROLINA: Buncombe, Cald- well, Craven, Cumberland, Durham, Henderson, Iredell, Jones, Macon, Madi- son, Orange, Polk, Wilkes. omo: Fairfield, Highlands, Jackson, Orain, Meigs, Scioto. PENNSYLVANIA: Berks, Bucks, Delaware, Chester, Lancaster, Lebanon, Montgomery, Northampton, Westmoreland. somit CAROLINA: Anderson, Greenwood, Laurens, Oconee. rrNNrEssrE: Anderson, Blount, Carter, Cheat- ham, Davidson, Franklin, Grainger, Grundy, Hamilton, Hawkins, Jefferson, Knox, Lake, Lewis, Madison, Marion, Roane, Shelby, Sullivan, Tipton, Union, White. Texas: Location unknown. VIRGINIA: Accomac, Arlington, Bedford, Brunswick, Fairfax, Fauquier, Gloucester, Henrico, Isle of Wight, James City, Madison, Pittsylvania, Prince Edward, Prince William, Princess Anne, South- ampton, Spotsylvania, Surry, Wythe. Wirst VIRGINIA: Barbour, Cabell, Cal- houn, Hardy, Upshur. * 1959] Gillett — Bartonia and Obolaria 61 Linnaeus in his Species Plantarum gave a reference to the Hortus Cliffortianus where he wrote: “Crescit in Virginia, unde translatum communicavit DD. Gronovius." He also cited Grono- vius' Flora Virginica. In this flora, Clayton n. 286" is cited as it was by the younger Gronovius in the second edition of 1762. Clayton 286 then may be considered as the type. l ON Br Hf te € 0 ) NU TERR 4 [ J / P y ) ) "i ^ o e | RÀ *. e r 7 i b ^ ^ f ve n ` j t a \ e i 7 60 1 4 "4 vl Ü E Lllj[A.I-—r£-———-üR-23 UE d — \ ~~ A 8 i S| i Dx ) M / H „ Y \ rs AMa a9 ©, i 70 K sA . , Nes P ) \ NI ( A 9 "d \ \ À Y | i f \ d \ i J — REUS. CCC NEA V ? ^ Ys 6 ¢ SN 2 ay "iia i X M Sd 5 «4 <> x N k A ^ à * — 2 SCALE , 0 MILES 400 T 2 UN p MAP á. Distribution of Obolarta virginica. There is no possibility of confusion with Orobanche virginiana L. Sp. Pl. 2: 633. 1753, because Clayton 604 was cited by both Gronovius and by Linnaeus. ‘This specimen located at the Brit- i sh Museum is the type of Orobanche virginiana L. A total of 236 sheets have been seen from the herbaria listed in 62 Rhodora [VOL. 61 Part I. Since there can be no confusion of species identity, no specimens have been cited. EXCLUDED SPECIES OBOLARIA CAROLINIANA Walt. Fl, Car. 166. 1788. — Monniera caroliniana (Walt) O. Ktze. Rev. Gen. 2: 463. 1891. (Labi- atac). OBOLARIA BOREALIS (I.) Kuntze, Rev. Gen. 1: 275. 1891. — Linnaea borealis L. Sp. Pl. 2: 631. 1753. (Caprifoliaceae) LITERATURE CITED BECK-MANNAGETTA, G. 1897. Orobanchaceae in Die Natürlichen Pflanzen- familien 4: 123—132. BICKNELL, E, P. 1915. The ferns and flowering plants of Nantucket — xiv. Bull. Torr. Bot. Cl. 42: 27-47. Braun, E. L. 1937. Some relationships of the flora of the Cumberland Plateau and Cumberland Mountains in Kentucky. RHODOERA 39: 193—208. Dox, G. 1837. General system of gardening and botany. J. G. and F. Riving- ton. London. 4 vols. (Obolaria 4: 635). FrRNALD, M, L. 1921. The Gray Herbarium expedition to Nova Scotia, 1920. RHODORA 23: 284—300. — — — 19383. Recent discoveries in the Newfoundland flora. RHODORA 35: 265—283. — — 1946. Does Bartonia verna grow in Virginia? RHODORA 48: 327—328. FERNALD, M. L. AND C. X. WEATHERBY. 1932. Bartonia; a comedy of errors. RuoponA 34: 164—167. Gic, E. 1895. Gentianaceae in Engler & Prantl, Die natürlichen Pflanzen- familien 4 (2): 50—108. Grisepacu, H. R. A. 1839. Genera et species gentiancarum. Stuttgart, viii. 364 pp. 1845. Gentianaceae in DeCandolle, Prodromus systematis naturalis regni vegetabilis 9: 38—141. Kunrzr, O. 1891. Revisio generum plantarum 2: 275, 430. Linnarus, C. 1737. Hortus Cliffortianus. Amsterdam. 501 pp. —. 1771. Mantissa Plantarum 2: 422 (fide Richter, C. 1840. Codex botanicus Linnacanus 612). Rarinesgur, C. S. 1808. Medical Repository II. 5: 536. Reurer, G. F. 1847. Orobanchaceae in DC. Prod. 11: 1-45. Ronivsov, B. L. 1898. Notes on the genus Bartonia. Bot. Gaz. 26: 46—48. — — 1908. Notes on the vascular plants of the northeastern United States. RHODORA 10: 29—35. Rork, C, L. 1949 Chromosome numbers in the Gentianaceae. Amer. Jour. Bot. 36: 687—701. Wiper, R. L. 1955. A Revision of the North American genus Sabat (Gen- tianaceae). RHODORA 57: 1-33; 43-71; 78-104. 1959] Harris—Elsholtzia Ciliata 63 ErsHOLTZIA CILIATA IN Essex County, MassAcHUsETTS — In September of 1958 Mr. Francis C. Wade gave me a specimen of a small weed which Mrs. Ralph Foster had found growing in her flower garden on Main Street in Rowley, Massachusetts. While taking out some class material from the Labiatae covers in the herbarium a few days later I happened on to a specimen from Japan of the same plant so that identification was much easier than is often the case with adventives. ELSHOLTZIA CILIATA (Thunb.) Hylander is a native of Asia which has become natural- ized in Temiscouata County, Quebec. In the Gray Herbarium and the Herbarium of the New England Botanical Club there are only five collections from the United States: Devil's Lake, Wisconsin; New Limerick, Maine; Chelsea, Vermont; and Fitch- burg and Revere, Massachusetts. It is an annual with peculiar one-sided spikes of very small flowers which are almost hidden by broad bracts. A specimen has been deposited in the herbarium of the New England Botanical Club. Weed in garden, Main Strcet, Rowley, Essex County, Massachusetts, Francis C. Wade 18774 (22 September 1958); same locality, Francis C. Wade, Stuart K. Harris 18799 (26 September 1958). — STUART K. Harris, BOSTON UNIVERSITY. Volume 61, Number 721, including pages 1-33, was issued 29 March, 1959 QW REFERENCE imn, MAY 11 1959 Dodota JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor-in.Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR, IVAN MACKENZIE LAMB Vol. 61 March, 1959 No. 723 CONTENTS: The Northernmost Station of Magnolia virginiana, Its History and Present Status. %%% V ð⁰˙ w. Re en Ae 65 Allium speculae, A New Species of the Allium canadense Alliance from Alabama. Marion Ownbey and Hannah C. Aase esses 70 An Alternative Explanation of Subspeciation in Asclepias tuberosa. QW: ae 8 72 Notes on the Distribution of Ohio Compositae: II. Eupatorieae, Senecioneae, Cynareae, Cichorieae. Robert W. Long 76 The Balanophoraceae in the Carribbean Flora. r e ß eo rieda 79 New Names within the Section Tridentatae of Artemisia. Mlan A Beetle ees 4 e naed tie S ECT ETE 82 A New Varietal Combination in Oxybaphus. Wetan AT DOT. . ronds oe e E e: 85 The New England Botanical Club, Inc. Botanical Museum, Oxford St., Cambridge 38, Mass. RHODORA.—A monthly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid, in funds payable at par in United States currency in Boston; single copies (if available) 60 cents. Back vol- umes 1—58, with a few incomplete, can be supplied at $5.00 per volume. Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes, relating directly or indirectly to the plants of North America, will be considered by the editorial com- mittee for publication. Articles concerned with systematic botany and cytotaxonomy in their broader implications are equally accept- able. All manuscripts should be double-spaced throughout. Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance, will be furnished at cost. Address manuscripts and proofs to Reed C. Rollins, Gray Herbarium, 22 Divinity Avenue, Cambridge 38, Mass. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill, Botanical Museum, Oxford Street, Cambridge 38, Mass. Second Class Postage Paid at Boston, Mass. EATON PRESS, Inc. Specialists in Learned Publications 17 SPRING STREET WATERTOWN 72, MASS. Rhodora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 61 March, 1959 No. 723 THE NORTHERNMOST STATION OF MAGNOLIA VIRGINIANA, ITS HISTORY AND PRESENT STATUS STUART K. HARRIS The most northern station of Magnolia virginiana L. is located in the town of Magnolia situated on the mainland of Cape Ann and in the township of Gloucester, Essex County, Massachusetts. The nearest known station to the South is on Long Island, New York. The presence of the plant in Massachusetts has elicited a great deal of interest since the time of its discovery and a number of articles have been written concerning the history and condi- tion of the plants at this station. The best historical account of the species was published in 1916 by Dr. George G. Kennedy in RHODORA vol. 18. In 1928 Richard J. Eaton reported on the status of the species in RHODORA vol. 30. Because over thirty years have elapsed since the appearance of Mr. Eaton's article it secms time to review once more the past history of the Sweet Bay in Massachusetts and report on its present condition. Much of the historical data has been taken either directly from Dr. Kennedy's paper or from the references given in that paper. In the past there was confusion as to who discovered the Sweet Bay and when it was discovered. John Robinson in his Flora of Essex County, Massachusetts (1880) states, "First brought to notice by Rev. Manassah Cutler during the last century." Prof. John G. Jack (1889) says, "Here it has been known for over a hundred years, having been first brought to notice by the Rev. Manasseh Cutler." Mr. T. Otis Fuller (1890) quotes a marginal 65 66 Rhodora (VOL. 61 note made by Judge John Davis of Boston in his copy of the first edition of Bigelow’s Florula Bostoniensis (1814) to the effect that the first specimen was obtained by Chief Justice Parsons in the summer of 1805. None of these is correct as to the date of the discovery. In volume two of Cutler and Cutler's Life Journals and Cor- respondence of Rev. Manassch Cutler, LL.D. (1888) is informa- tion which leaves no doubt as to the discovery. The magnolia was first noticed on Tuesday 22 July 1806 by Chief Justice The- ophilus Parsons while riding through the woods in Gloucester during a shower of rain! I regret I have been unable to ascertain whether it was in the morning or the afternoon. Specimens were collected by him on the following Friday. He promptly wrote a letter to the. Rev. Manasseh Cutler, the outstanding botanist in Essex County at the time, who set out in scarch of the plant on Monday, July 28, the day after he received the letter. While hav- ing dinner with Captain Ingolson at Kettle Cove a Mr. Goldsmith brought in specimens of the magnolia without being aware that Rev. Cutler was looking for it. In the afternoon Cutler found that the plant was abundant in two swamps close to the road from Manchester to Gloucester. Being more familiar with Magnolia stellata and M. soulan- geana which blossom in the spring, the last half of July seemed a late date to me to find magnolia in flower. However, in check- ing all the dated specimens I have seen, all collections bearing blossoms have been taken in July and Jack says that the first week in July is the best time to go and see the shrubs. A note by an unsigned correspondent appears on page 612 of Garden and Forest vol. 2 and again as a quotation in Kennedy’s article suggesting that Magnolia virginiana is not really native to Gloucester but was introduced there from a more southern state by the early settlers. As far as I know there is not a shred of evi- dence to support this interesting theory. ‘This note is the major portion of a letter dated 21 November 1889 and sent to John Robinson at the Peabody Academy of Science at Salem by M. A. Walton, “Hermit.” ‘This letter is attached to one of the sheets 19:9] Harris — Magnolia virginiana 67 of Magnolia in the herbarium of the Peabody Museum of Salem. Mason A. Walton, the Hermit of Gloucester, was suffering from very poor health and about 1884 decided that continued life in the city would soon prove fatal, so he moved to the woods of Cape Ann and lived first in a tent and later in a cabin and nursed himself back to health. He was an untrained nature lover with a keen eye, as is evidenced by the contents of this letter and by the book he published in 1903, A Hermit's Wild Friends, which is full of observations he made of the plants, birds and mammals of the region. The note in Garden and Forest terminates as follows, “It must be evident to any careful observer that Magnolia glauca is here strugeling in an unnatural climate. The primary roots grow straight down into the muck and in the fall are thickly covered with succulent rootlets, snowy white in color. In the spring these rootlets are mostly dead, and the greater part of the young shoots die down to the moss, and a certain per cent of the old plants winter kill, which goes to show that there is no harmony between shrub and climate." There is more pertinent information in the unpublished portion of the letter. "Magnolia glauca, does not extend into Essex, so far as I know. I have traveled through many of the swamps of that town without discovering it, and persistent enquiry of Essex people, long ago convinced me that it did not extend beyond West Gloucester. I believe the shrub is confined wholly to Ward 8, City of Gloucester. Below I give the names of some of the swamps where it grows: ‘Magnolia Swamp’; Barrel Swamp’; 'Rust's Swamp’; ‘Cedar Swamp’; 'Bray's Swamp’ and several other minor swamps. During the time I have lived here, five years, Magnolia glauca has increased in Magnolia Swamp. I do not think it has noticeably increased in other swamps, but certainly, it has not decreased." This brings us to a consideration of the range of Magnolia virginiana in New England. In the ‘Trees and Shrubs of Massa- chetts, ed. 2 (1875), George B. Emerson states, "It is said to have been found, in a single spot, in the county of York, Maine." Since no specimen is known to support this claim and no botanist 68 Rhodora [VOL. 61 has reported finding the Sweet Bay in Maine since 1875, this report can be eliminated. ‘Thus the range can be confined to Cape Ann, Massachusetts. Emerson (1846) reports it from a swamp in deep woods in Essex but there are no specimens to substantiate the claim. Walton, as noted above, believed that the species grew only in West Gloucester and I think this is essen- tially correct. However, there are two collections in the Gray Herbarium, one made by William Oakes and the other by Charles E. Faxon, which give the adjoining town of Manchester as the locality. The location of the Magnolia Swamp became well-known soon after its discovery and great numbers of the magnolia were dug and moved to private gardens. During the season when the shrub was in flower large numbers of the blossoms were picked, with little or no regard for the welfare of the plants, and sold on the streets of Salem and Boston. In the herbarium of the Peabody Museum are two sheets of blossoms purchased from small boys in 1878 and 1879. As early as 1846, Emerson expressed the fear that the station would soon become extirpated but nothing was done to remedy the situation. Kennedy quotes a letter written in 1916 by Charles E. Faxon to Walter Deane which states that forty-five years before he had found plenty of good specimens fif- teen feet tall or more and that it was casy to find them because the boys who sold the flowers on the Boston trains had made trails from one plant to another all over the swamp. However, when he visited the place two years previously (July 1913) in the company of Dr. Kennedy and the local ‘Tree Warden, they could find only two little plants a few feet high. This must have been the low point of the stand. It is possible that they hap- pened to make their visit at a time when most of the plants had been killed back to the ground by a severe winter and the young shoots had not yet appeared. Eaton in 1928 notes that the Magnolia Swamp had been made a part of Ravenswood Park in the early 1920's and that the mag- nolia was at last protected. Along the paths constructed across the swamp he saw about a dozen species. In talking with him 1959] Harris — Magnolia virginiana 69 recently he told me that he has counted over twenty-five plants close to the paths. All my visits to Magnolia Swamp have been made in the winter which is not as illogical as it first seems because the ground is then frozen and progress is easy in the swamp and the magnolias being evergreens are easy to spot because they are about the only woody plants there holding their leaves. Besides I would rather fight chilblains than mosquitoes. As Walton states the swamp contains several hundred acres and is long and relatively narrow, ten to over one hundred rods wide. In late February of this year I made a rough census of the distinct plants seen during twenty or twenty-five zigs and zags across the swamp. I was able to count eighty clones and I am sure that I missed some so that it is prob- ably safe to say that at least one hundred still exist. The length of the stems varied from less than a foot to about fifteen feet. I saw no evidence of fruit on any of the plants and this was also true on two previous visits made during the past seven ycars. This leads me to wonder if the magnolias are now being killed by kindness through being shaded out by the red maples and other taller trees in the swamp which are also being protected. In the fall of 1957, Miss Frances L. Burnett of Manchester told me of another apparently natural stand of Magnolia virginiana in a swamp near the Manchester line and at least a mile and one half from the Magnolia Swamp. I have visited this twice. The swamp itself is small and contains only five or six plants but these represent the only wild magnolias outside of Magnolia Swamp of which I am aware. In 1957, I found several fruit on this group but I saw none this February. — DEPARTMENT OF BIOL- OGY, BOSTON UNIVERSITY, BOSTON, MASS. LITERATURE CITED CurLER, W. P. & J.P. CurrER. 1888. Life Journals and Correspond- ence of Rev. Manasseh Cutler, LL.D. 2:310, 311. EATON, R. J. 1928. The Present Status of Magnolia virginiana in Massachusetts. RHODORA 30:207, 208. Emerson, G. B. 1846. Trees and Shrubs of Massachusetts, ed. 1. 527. Emerson, G. B. 1875. Trees and Shrubs of Massachusetts, ed. 2. 603. 70 Rhodora [VOL. 61 FurrtreR, T. O. 1890. Magnolia glauca in Massachusetts. Garden and Forest 3:23. Jack, J. G. 1889. Magnolia glauca in Its Northern Home. Garden and Forest 2:363, 364. krNNEDY, G. G. 1916. Some Historical Data Regarding the Sweet Bay and Its Station on Cape Ann. Ruopora 18:205-212. Watton, M. X. 1889. Note on Magnolia glauca. Garden and Forest 2:612. ALLIUM SPECULAE, A NEW SPECIES OF THE ALLIUM CANADENSE ALLIANCE FROM ALABAMA! MARION OwwNnbBEY AND HANNAH C. AASE The Allium canadense alliance comprising ten North Ameri- can species has recently been revised by the authors.? Scarcely was this monograph off the press than there appeared in our living collection an undoubted eleventh member of this group. We are indebted to Dr. Carroll E. Wood, Jr., for supplying bulbs and later herbarium specimens of this novelty which we describe below. Allium speculae, sp. nov. Bulbus ovoideus non bulbuliterens saepe unus ex pugno, tunicis interioribus albidis, cellulis cuticulae indistinctis recte elongatis regularibus, tunicis exterioribus fuscis persistentibus anguste fibroso-recticulatis, maculis vacuis; foliis aliquot canaliculatis in sectione transversa concavo-convexis 1-2 mm. latis integris scapo brevi- oribus in flore viridibus; capo uno tereti 2-3 dm. alto; spatha mem- branacea caudata, bracteis plerumque tribus lanceolatis attenuatis plus- minusve connatis plerumque uninervatis; umbella pauci (10-15—) flora erecta, pedicellis tenuibus demum subaequilongis, perianthio plerumque 2-3-pio longioribus; perianthii segmentis 5-6 mm. longis ellipticis obtusis ad apicem involutis pallide roseis late patentibus non valde reflexis in fructu marcescentibus super ovarium conniventibus; stamini- bus perianthii segmentis paulo brevioribus ascendentibus, filamentis subulatis basi dilatatis coalitisque, antheris oblongis obtusis versatilibus; ovario turbinato trilobato 6-caniculato distincte cristato, cristae processis 1 This investigation was supported in part by funds provided for biological and medical research by State of Washington Initiative Measure No. 171. "ON NDBY, M., AND H. C. AASE. 1956. Cytotaxonomic Studies in Allium. I. The Allium canadense alliance. Research Studies of the State College of Washington. Monographic Sup- plement, No. 1. 106 pp. 1959] Ownbey and Aase — Allium speculae 71 6 binis complanatis horizontalibus, stylo lineari filamentis subaequi- longo, stigmate capitato; seminibus ignotis; 2n = 14. Bulb ovoid, without basal bulbets, often one of a cluster, inner coats whitish, with the epidermal cells indistinct, vertically elongate and regu- lar or nearly so, outer coats persisting as a series of grayish or brownish very fine meshed open recticula, usually enclosing only a single bulb; leaves several per bulb, channeled, concave-convex in cross section, 1—2 mm. broad, entire, shorter than the scape, green at anthesis; scape 2-3 dm. tall, terete, solitary (or sometimes a second one appearing in culti- vated plants) ; spathe membranaceous, caudate, breaking at anthesis into usually 3, lanceolate, attenuate, partially united, mostly 1-nerved bracts; umbel comparatively few (10-15) — flowered, erect, pedicels slender, becoming subequal in length, mostly 2-3 times that of the perianth; perianth segments 5-6 mm. long, elliptic, obtuse, involute at apex, pinkish, widely spreading, but not strongly reflexed, remaining thin and withering over the ovary; stamens a little shorter than the perianth, ascending, filaments subulate, dilated and united into a ring at the base, anthers oblong, obtuse, versatile; ovary turbinate, 3-lobed, but 6-grooved, each lobe with a pair of flattened horizontal processes which together form the distinct crest; style linear, about as long as the filaments, stigma capitate; seeds unknown; 2n = 14. The type specimen (WS) was collected in bud April 25, 1955, in black sandy soil (wet at this season) with Schoenolirion cro- ceum on an open expanse of flat sandstone surrounded by Pinus- Quercus-Carya woods at the northwest rim of Little River Can- yon, about 1.3 miles from the northeast end of Little River Canyon Parkway, Lookout Mountain, southeast of Fort Payne, De Kalb County, Alabama, by Carroll E. Wood, Jr. (No. 8695). The plants were allowed to flower, and specimens prepared May 4. A selected isotype is in the Gray Herbarium. Additional data were obtained from plants of this collection (WS) grown in the greenhouse at Pullman, Washington, in 1956. The species is abundant at the type locality, and several isotypes obtained remain to be distributed by the collector. In cultivation, the plants did not survive beyond the first. year. Allium speculae habitally resembles the widespread A. cana- dense var. mobilense of the Gulf states, and if it has been col- lected before the specimens will probably be found under this name or one of its synonyms. It differs from var. mobilense, how- ever, in its prominently crested ovaries, its mostly ]-nerved bracts, 72 Rhodora VOL. 61 and its more widely spreading perianth segments. Its nearest relationship is probably not with that variety and, if this is true, it has no close relatives. The only other species of this alliance with crested ovaries in eastern North America is A. Cuthbertii, which is at once so conspicuously distinct from A. speculae that they could not be confused. A. Cuthbertii has only two leaves per scape, the processes of the conspicuous crest are contorted, the perianth segments reflexed, and the bracts mostly 5-nerved. Yet, A. Cuthbertii seems to be the closest relative of A. speculae. Among the western species of this alliance, only A. Geyeri seems to be a possible relative. This species, however, has urceolate- campanulate flowers, and the procesess of the crest are little more than inconspicuous knobs. ‘The relationship with A. Geyeri can- not be close. It seems, therefore, that A. speculae represents an eleventh distinct evolutionary line in the A. canadense alliance, or that it stands ancestral to A. Cuthbertii. The later hypothesis is particularly appealing. Morphologically, A. speculae is inter- mediate between A. Cuthbertii and the less specialized western species, such as A. Geyer’. Furthermore, its present distribution fits this hypothesis, inasmuch as it is apparent that the A. cana- dense alliance as a whole radiated from the Southwest. One can- not overlook, however, some resemblance between A. Cuthbertit and A. Plummerae and the possibility that the latter, although tetraploid, might be the most primative surviving member of the alliance. This might imply an early separation of the lines which gave rise to A. speculae and A. Cuthbertii, respectively, so that the former could not stand as ancestral to the latter. — STATE COLLEGE OF WASHINGTON, PULLMAN. AN ALTERNATIVE EXPLANATION OF SUBSPECIATION IN ASCLEPIAS TUBEROSA C. W. JAMEs Asclepias tuberosa L. (Woodson, 1954) is represented in the southeastern United States by a relatively extensive Appalachian subspecies and a somewhat more restricted Coastal Plain repre- 1959] James — Subspeciation in Asclepias 73 sentative, a relationship which has been observed repeatedly in many species and genera for some time. It is easily conceived how many such Coastal Plain forms could have originated from Appalachian species some time after the close of the Cretaceous and differentiated in response to ecogeographical factors present in the newly emerging Coastal Plain. However, in such in- stances, unless isolating mechanisms other than those of ecogeo- graphical character have since become established or unless the ecological barrier is sharply delimited, intergradation is expected between such forms. Yet, Woodson's (1947b) detailed statistical analyses of leaf variation in Asclepias tuberosa revealed that this intergrading zone presently evident between the Coastal Plain subsp. Rolfsii and the Appalachian subsp. tuberosa is due to hybridization between the two subspecies and not to an ecogeo- graphical differentiation into a cline. This led Woodson (1947a) to conclude that Rolfsii must have originated independently in more or less complete isolation from tuberosa. To account for this, Woodson then postulated that Rolfsii evolved on Orange Island, a hypothetical island or archipelago in north Florida during Oligocene times. Since Woodson's account of the sup- posed origin of Rolfsi! has been considered biological evidence substantiating the existence. of a functioning Orange Island Refugium by Woodson, Thorne (1949), and others, it seems highly desirable to explore other ways in which this subspecies could have originated lest we find ourselves relying too freely and perhaps unjustifiably at times on this Island as a refugium. Granted the occurrence of hybridization between the sub- species of Asclepias tuberosa, how then can one account for this apparently independent origin of Rolfsii by means of the known processes of ecogeographical subspeciation? As Woodson pointed out, Rolſsii, surely, could not have maintained a separate exist- ence with tuberosa upon the Appalachian upland, later mi- grating to Florida only to return in panmixy with its sister subspecies.” The Coastal Plain today is a geographical province differing considerably from the adjacent and generally more mesic Pied- 74 Rhodora [VOL. 61 mont and Appalachian provinces. Since the habitats of the Coastal Plain presumably developed gradually through a succes- sional series over a period of time, it seems reasonable to assume that these differences (notably the edaphic ones) could only have been more striking when the Coastal Plain was in its initial phase of exposure. This would mean that the genetic system of the pioneering Coastal Plain clement of tuberosa would have been selected under much more rigorous and quite different condi- tions from those prevailing today. It possibly required. consid- erable time for a genetic system to evolve from the parental species which was sufficiently adapted to be aggressive in this newly available environmental complex. But once such biotypes had evolved, they could migrate southward as rapidly as the successional stages and dispersal would permit since environ- ments characteristic of the succeeding portions of the Coastal Plain would be very similar to the first to which the invading element must necessarily have been adapted. The rapidity with which this genetic clement migrated away from the parental stock would result in a progressively more effective geographical isolation, thus accounting for the apparently independent evo- lution of Rolfsii in more or less complete isolation. (Although not the case in Ztolfsii, this could provide conditions facilitating the evolution of other isolating mechanisms). ‘The basis of the reasoning employed here is dependent upon the following tenets: The less favorable a newly available area is for occupancy and invasion by organisms of an adjacent area, l. the less the probability of the presence of existing biotypes which can immediately invade the new area, 2. the greater the difference there will be in the genosystem of a derived race which can invade the new area, 3. the greater the probability of a longer period of time required for the evolution of this genosystem, 4. the greater the differential in rate of migration between the best adapted and the least adapted biotypes of this genosystem, 5. the greater the degree of morphological and physiological differ- entiation and/or specialization of the invading race, 6. the greater the effectiveness of ecogeographical isolation (if such isolating mechanisms are involved.) 1959] James — Subspeciation in Asclepias 75 7. the longer the newly evolved race can maintain a separate identity Irom the parental species. The genetic system of Rolfsii is presumably a specialized one derived from only a part of the broader genetic system of tuber- osa, and one which, perhaps, has become even further specialized. There would then be little or no pressure on a northward move- ment of Atolfsii genes into tuberosa. There would be, however, continued forces operating at the juncture of subspecific differ- entiation on a flow of tuberosa genes southward just as there had been since the time of exposure of the Coastal Plain. Fur- thermore, as the Coastal Plain became more mature it offered more variety and less severity in habitat; consequently, many tuberosa genes and gene complexes which previously were insuf- ficiently adapted to that environment could then flow southward in addition to some of the previous ones. This would result in an invasion of the Coastal Plain Rolfsii by increasingly less dif- ferentiated biotypes of tuberosa which could then hybridize with the remaining Rolfsit element in that area of the Coastal Plain where subspeciation was first initiated. ‘This progression of sec- ondary invasion elements of tuberosa would then tend to absorb Rolfsii and could then account for the present hybridization occurring between the two subspecies. If Rolfsii were ever present in the Carolinas and. northward it has apparently since been absorbed by tuberosa. It is of in- terest to note in this connection that geological evidence suggests that a considerable portion of the Carolina Coastal Plain has been exposed and available to plants since the end of the Cre- taceous. At the present time, Rolfsii appears to be losing its identity throughout the remainder of the Coastal Plain with the exception of peninsular Florida which is farthest from the pre- sumed point of origin. This analysis, of course, does not disprove the possible exist- ence of an Orange Island Refugium. It is merely an attempt to offer an alternative explanation of subspeciation in Asclepias tuberosa which could account for the hybridization presently occurring between the subspecies without having to rely on an Orange Island Refugium. 16 Rhodora [VOL. 61 The author expresses his appreciation to Drs. R. E. Woodson, R. C. Rollins, J. J. Westfall, and C. E. Wood for reading and evaluating the manuscript. — DEPARTMENT OF BOTANY, UNIVERSITY OF GEORGIA. LITERATURE CITED Tuorne, R. F. 1949. Inland plants on the Gulf Coastal Plain of Georgia. Castanea 14: 88-97. Woopson, R. E. 1947a. Notes on the “Historical Factor" in plant geography. Contr. Gray Herb. 165: 12-25. 1947b. Some dynamics of leaf variation in Asclepias tu- berosa. Ann. Mo. Bot. Gard. 34: 353-432. 1954. The North American species of Asclepias L. Ann. Mo. Bot. Gard. 41: 1-211. NOTES ON THE DISTRIBUTION OF OHIO COMPOS- ILIAE: II. EUPATORIEAE, SENECIONEAE, CYNAREAE, CICHORIEAE ROBERT W. LONG This is the second paper of a series of three that presents some results of a recent study of Ohio Compositae. In part I! it was noted that plants discussed in these reports are ones whose occurrence in Ohio is questionable, judging from information given in Gray's Manual (1950) and The New Britton and Brown Illustrated Flora (1952). For the present, the nomenclature is derived chiefly from Gray's Manual, but this does not imply it is necessarily the best treatment for the taxa listed. All specimens and county records cited here are deposited in the Herbarium of The Ohio State University, and the identifi- cations have been verified by the writer. EUPATORIEAE Eupatorium album L. var. glandulosum (Michx.) DC. This variety is easily separated from the typical one by the occurrence of minute, dark glands on the phyllaries; thus, the variety is quite distinct. Its presence in southern Ohio represents a northward extension of the range given by Fernald. COLLECTION DATA: Jackson Co., Liberty Twp., ROBERT W. LONG. Notes on the distribution of Ohio Compositae: I. Heliantheae, Anthemidae. RHODORA 60:125-128. 1958. 1959] Long — Ohio Compositae 77 Big Rock, Leslie Pontius and Floyd Bartley, August 27, 1933. Other collections from Scioto County. Eupatorium rotundifolium L. var. rotundifolium. The distribution of this plant is apparently limited to the southeastern quarter of the state on the Allegheny Plateau. ‘These collections, however, are evi- dently from the northwestern edge of the range for the species, judging from accounts in the manuals. COLLECTION DATA: Scioto Co., Nile Twp., Conrad Roth, August 26, 1956. Other collections examined from Fair- field and Hocking counties. Eupatorium rotundifolium L. var. ovatum (Bigel.) Torr. That there are two recognizable varieties of this species in Ohio is clearly apparent from the specimens examined. Both manuals agree that two can be dis- tinguished, with E. pubescens Muhl. being a synonym for var. ovatum. Recent collections place both varieties in southcentral counties and for var. ovatum this record will be a range extension. COLLECTION DATA: Hocking Co., "Neotoma", Gareth Gilbert, August 12, 1955. Also, col- lections from Jackson County. Liatris borealis Nutt. The collections of this plant are very near L. scariosa (L.) Willd. If they are correctly identified, then Ohio records constitute a westward range extension. The following are provisionally placed in this species. COLLECTION DATA: Adams Co., Andrews School Prairie, H. R. DeSelm, September 18, 1952. Other county records for Erie, Henry, Pike, Ross, and Wood counties. Liatris punctata Hook. The single collection suggests that the species occurs only rarely. This is considerably east of its chief distribution, that being in the dry prairies from Canada to Texas. The specimen does not appear to be of var. nebraskana Gaiser. COLLECTION DATA: Franklin Co., Clintonville, John H. Schaffner, October 3, 1903. Liatris scariosa (L.) Willd. Ohio plants appear to be relatively abundant, but predominantly in the southern counties which extends the range of the species westward. Intergrades to L. borealis, and L. aspera Michx. were found. The following collections have been provi- sionally traced to L. scariosa as described in Gray's Manual. COLLECTION DATA: Vinton Co., Vinton Twp., Sec. 22, roadside on ridgetop, Janice Beatley, August 30, 1952. Other collections from Adams, Athens, Erie, Fairfield, Franklin, Fulton, Jackson, Meigs, Perry, Pike, Ross, and Scioto counties. SENECIONEAE Senecio antennariifolius Britt. The presence of this species in Ohio is a significant westward range extension from the general area of dis- tribution, as given in both manuals. Schaffner does not give it in either the Revised Catalogue or in The Field Manual of the Flora of Ohio. All collections examined had been made since 1950. Three collections 78 Rhodor: [VOL. 61 intergrade to S. plattensis Nutt. COLLECTION DATA: Hocking Co., Cedar Falls, Floyd Bartley and Lawrence E. Hicks, June, 1956. Senecio Smallii Britt. Although this is a southern species, ranging into Kentucky and Pennsylvania, collections have been seen from sev- eral parts of the state. Possibly the species has migrated only recenlty into Ohio, or has been introduced as a weed. Schaffner does not list it and all collections examined were made since 1949. One specimen (Geauga Co.) intergrades to S. pauperculus Michx. in stem and inflo- rescence characteristics. COLLECTION DATA: Lawrence Co., in old field 2 mi. west of Oak Ridge Furnace, Lawrence Hicks and Floyd Bartley, May 30, 1954. Also, specimens were examined from Adams, Cuyahoga, Geauga and Summit counties. CYNAREAE Centaurea repens I. This is a distinctive star-thistle, with bushy branches bearing small, linear leaves abundant to the numerous heads that terminate the branches. It is represented in Ohio by at least a single collection and this is an important eastward range extension. COLLECTION DATA: Clinton Co., collected along railroad tracks near New Vienna, a large patch, Katie M. Roads, June 27, 1939. Cirsium arvense (I.) Scop. var integrifolium Wimm. and Grab. Two collections have been made in widely-separated parts of the state. Ac- cording to published accounts, these locales are considerably west and south of the usual range for the variety. COLLECTION DATA: Cuyahoga Co., North Olmstead, Freda Deimers, August 4, 1925. Also, a collection was seen from Logan county. Cirsium carolinianum (Walt) Fern. and Schub. The omission of this species from Schaffner's published accounts of the flora was owing to his identification of Ohio collections as C. virginianum (L.) Michx. The specimens seen, however, clearly belong to C. carolinianum judg- ing from the size of the involucre and number of leaves. Two county records extend the range of the species into central Ohio. COLLECTION DATA: Pike Co., Floyd Bartley and Leslie Pontius, June 7, 1945. Other collections examined from Franklin, Madison, and Scioto counties. Cirsium Hillii (Canby) Fern. This species is distinguished from the common Ohio thistle C. pumilum (Nutt.) Spreng. by the presence of short prickles on the outer phyllaries that also have a dark band present running the length of the phyllary. Schaffner includes it in his Field Manual of the Flora of Ohio, but not in the Revised Catalogue. The single head of the specimen examined was unusually large. COLLECTION DATA: Coshocton Co., in dry open fields, leaves green beneath, location N. A. E. W., Harold N. Moldenke 13232, July 15, 1942. CICHORIEAE Krigia Dandelion (I.) Nutt. Ohio records result in a northward and eastward range extension. The single collection, however, comes from 1959] Long — Ohio Compositae 79 an area that several authors have commented on as being a peculiar mosaic of prairie and Allegheny vegetation which is not typical for the state. The specimen appeared not unusual, morphologically, judging from the description. COLLECTION DATA: Adams Co., sw. corner Oliver Twp., post oak-white oak woods, E. Lucy Braun, May 18, 1954. Leontodon autumnalis L. var. autumnalis. The plant has become established as a weed of northern counties, and its occurrence repre- sents an eastward and southward range extension. COLLECTION DATA: Summit Co., lawn weed, Cannon road 34 mi. e. of Twinsburg, Ervin M. Herrick, August 28, 1955. Other specimens examined from Ashta- bula, Franklin, Lake, and Medina counties. Leontodon autumnalis var. pratensis (Link) Koch. A single collec- tion was seen of this variety, although one might expect to find it to be more abundant, especially in the northern part of the state. COLLEC- TION DATA: Ashtabula Co., Trumbull, IL. E. Hicks, June 10, 1931. — DEPARTMENT OF BOTANY, OHIO WESLEYAN UNIVERSITY, DELAWARE, OHIO. THE BALANOPHORACEAE IN THE CARIBBEAN FLORA! RICHARD A. HOWARD This small family of root parasites is represented in the Antilles by two genera, Scybalium and Helosis. One species of Scybalium, S. jamaicense, has been found in Cuba, Jamaica, Hispaniola and Puerto Rico. The second genus, Helosis, consists of three species known from South America and Central America. Sandwith (Kew Bull. 1931:59. 1931.) and Harms in his monograph (Pflanzenfam. 2nd ed. 16b:321. 1935.) suggest that one of them, H. cayennensis, may possibly occur in Guatemala and Cuba as well as in northern South America. Standley and Stevermark (Flora of Guatemala, Fieldiana, Bot. 24:93. 1946.) refer the Guatemalan specimens to Helosis mexicana, but state that "three species have been de- scribed, all of them perhaps to be reduced to H. cayennensis (Swartz) Spreng. of northern South America. Only the follow- ing [H. mexicana] is known from Central America.” The specific differences suggested by Harms for the three species of Helosis do not appear to be substantial taxonomic characters. However, 1 Work on the flora of the Lesser Antilles is supported by a grant from the National Science Foundation. 80 Rhodora [VOL. 61 as a group, the plants have been poorly collected and have not received careful comparative study for taxonomic classification. The earliest record of Helosis from the West Indies appears in Grisebach, Flora British West Indies (509. 1860.), when he reports Helosis guianensis from "Trinidad. In his Catalogue Plantarum Cubensium (118. 1866.) Grisebach lists “Helosis mexicana Liebm.", citing the Charles Wright collection 2636 (cu). This plant has not been recollected. Leon and Alain (Flora de Cuba 2:84. 1951.) change this identification to Helosis guianensis L. C. Rich. Two additional collections of Helosis from the Lesser Antilles are now on hand, and allow further consideration of this peculiar root parasite. Dr. Walter Hodge, in the course of his field work in Dominica, received a report that Helosis was “quite abundant on Morne Anglais”. He did not personally encounter the plant, but re- ceived a specimen from Mrs. Alastair Forbes collected on the banks of the Laurent River near the western base of Morne Negre Maron (Hodge 1206. Gu). This specimen has been un- identified since 1940. More recently, Mr. George Proctor made an excellent collection (Proctor 17764, Gu), including material preserved in alcohol, from the mossy montane forest between 2500 and 3000 feet on the northwest spur of Morne Gimie in St. Lucia. The three collections from the West Indies can now be exam- ined together and compared with material from South and Cen- tral America. All specimens are small, with nearly globular heads. None exceeds 5 cm. in height in mature condition. ‘The heads range from 1-2 cm. in length and I-1.5 cm. in thickness. ‘The deciduous hexagonal scales of the anthophore are only 1.5-2 mm. in diameter, and the apex is extended in a setaceous tip. Both male and female flowers are borne in the head. The pistil- late flowers remain smaller than the abundant paleae, with only the styles protruding. The bilabiate perianth is represented by two triangular appendages, 0.1-0.2 mm. long. The perianth of the staminate flowers is well developed by contrast with a tube 2.5 min. long and three ovate lobes 1.5 mm. in length. The peri- anth lobes are valvate and strongly concave. The three filaments 1959] Howard — Caribbean Balanophoraceae 81 are stout but cylindrical, and in specimens preserved in alcohol these are free throughout their length. The short anthers, how- ever, are completely connate. This description agrees in the main with the excellent diag- nosis given by Sandwith for Helosis cayennensis (Kew Bull. 1931: 58-9. 1931.). It differs in the smaller size of all parts and in the description of the free filaments. In a key by Leon and Alain Scybalium is distinguished from Helosis by the former genus having three stamens and the latter two. Standley and Steyermark also describe the staminate flower as having two stamens. Both of these pairs of workers refer to fused filaments. Apparently their descriptions have been based on dried herbarium material where the filaments seem to adhere tightly. The report of only two stamens is apparently an error, for a reéxamination of the Wright collection from Cuba and of several specimens from Central America reveals three stamens in all specimens. 'The West Indian specimens are referred to Helosis cayennensis (Sw.) Spreng. Helosis guianensis L. C. Rich. is regarded by Harms and Sandwith as a synonym. All the West Indian material is of smaller size, has longer sctaceous tips to the hexagonal scales of the head, and a more obvious bilabiate development of the perianth of the pistillate flowers. Until additional collections are available from the Lesser Antilles and from Cuba it is not worth while to use these minute characters in an obviously reduced and specialized plant as distinctions for a new species. While Harms uses ovule characteristics to distinguish between Scybalium and Helosis, the single species of each in the West Indies can be separated on the basis of the numerous overlapping scales of the peduncle of Scybalium and the naked peduncle of Helosis which may have a single annulus of short but broad scales. The bracts of the flower head of Seybalium are broadly triangular, flat and imbricated, while those of Helosis are peltate, hexagonal in outline, valvate, and extended to a setaceous tip. — ARNOLD ARBORETUM, HARVARD UNIVERSITY. 82 Rhodora VOL. 61 NEW NAMES WITHIN THE SECTION TRIDENTATAE OF ARTEMISIA! ALAN X. BEETLE A study of the ecotypes of big sagebrush and its relatives was made over a three-year period in the eleven western states as well as in adjacent portions of Mexico and Canada. Comparison of morphological. characteristics, ecological. development, and dis- tribution of species has resulted in the discovery of one new sub- species and two new species within Section Tridentatae Rydb. of Artemisia. All of the types will be treated in detail in a forthcom- ing Wyoming Agricultural Experiment Station bulletin. While the use of Seriphidium for a subgenus grouping involv- ing both Old World and New World types (having homogamous heads) may vet have to be proved natural, there is much evidence that the use of Section Tridentatae Rydberg embodies a closely knit group of species endemic to the North American continent. These species may be outlined as follows: l. A. bigelovii Gray (4. petrophila Wooton & Standley) 2. A. nova Nelson 3. A. pygmaea Gray 4. A. rigida Gray 5. A. tripartita Gray (A. trifida Nutt.) 5(a). A tripartita subsp. rupicola Beetle, subsp. nov. Affinis A. tri- partita subsp. ¿ripartita sed nana, ad 1-1.5 dm. alta, foliis ad 3 cm. longis, singuli parte 1 mm. lata. Type collection: Wyoming, Albany County, Medicine Bow National Forest, Pole Mt., Sept. 7. 1958, 4. A. Beetle 13185. Type in the Rocky Mountain Herbarium: duplicates in Gray Herbarium, Chicago Natural History Museum, U. S. National Herbarium, and the herbarium of the University of California, Berkeley. A. tripartita subsp. rupicola is a dwarf plant rarely over 1.5 dm. o tall. Its leaves are often 3 cm. long, with both the basal portion and cach lobe at least | mm. wide. In contrast 4. tripartita subsp. tripartita is an erect plant up to two meters tall, with leaves seldom over 2 cm. long. both the basal portion and each of the three lobes about 0.50 to 0.75 mm. wide. 4. tripartita subsp. rupicola occupies rocky knolls from 8,000 to 9,000 feet elevation ip ublished with approval of the director, Wyoming Agricultural Experiment Station, as Journal paper no. 126. 1959] Beetle New Names of Artemisia 83 from the Owl Creek Mountains and from South Pass in Central Wyoming to the Laramie Range in southeastern Wyoming. 4. tri- partita subsp. tripartita occupies the deeper soils at the base of foothills from 5,000 to 7,000 feet elevation and occurs from south- ern British Columbia, Canada, southward through Washington, Idaho, and Montana to western Wyoming and northern Utah. 6. A. tridentata Nutt. 6(a). A. tridentata subsp. vaseyana (Rydb.) Beetle, comb. nov. Based on A. vaseyana Rydb. North American Flora 34 (3) :283. 1916. A. tridentata subsp. vaseyana has a range geographically dis- tinct from that of A. tridentata subsp. tridentata. It differs mainly in the characters emphasized by Rydberg, namely broader invol- ucre, more flowers per head, and broader, and more truncate or cuneate leaves. 6(b). A. tridentata subsp. vasyana (Rydb.) Beetle, f. spiciformis (Osterhout) Beetle, comb. nov. Based on 4. spiciformis Osterhout, Bull, Torrey Club 27:507. 1900. This plant is an extreme form, always occurring at the upper elevational limits of the subspecies and always in close proximity of A. cana subsp. viscidula, leading to the speculation that it may partially represent crossing between 4. tridentata subsp. vaseyana and 4. cana subsp. viscidula. 6(c). A. tridentata subsp. tridentata, f. parishii (Gray) Beetle, comb. nov. Based on 4. parishii Gray, Proc. Am. Acad. 17:220. 1882. While A. tridentata subsp. tridentata f. parishii 1s an eye-catch- ing form in the field because of its strikingly reflexed branches of the inflorescence (in addition to having somewhat hairy achenes), it occurs sporadically throughout the range of A. triden- tata subsp. tridentata in many more localities than have been reported previously. It has not been reported occurring within the range of A. tridentata subsp. vaseyana. 7. A. arbuscula Nutt. 7(a) A. arbuscula subsp. thermopola Beetle subsp. nov. Afhnis A. arbuscula, arbuscula sed planta gracile, cum foliis filiformibus, tri- partitis profunde. Type collection: Wyoming. ‘Teton County, along banks of Snake River near south entrance to Yellowstone National Park, August 10, 1957, A. A. Beetle 12631. rype in the Rocky Mountain Herbarium. This is the variation described by Ward (Contrib. from the 84 Rhodora [VOL. 61 Dudley Herbarium 4 (8):180, 1953) as having "deeply trifid leaves" with the exception that its distribution seems to be con- fined to the area from Yellowstone National Park, Wyoming, south to Salt Lake City, Utah. 8. A. longiloba (Osterhout) Beetle, comb. nov. Based on 4. spiciformis longiloba Osterhout, Muhlenbergia 4:69. 1908. While this plant has usually been treated as a part of A. arbus- cula, it occurs on different sites, preferring the most strongly alkaline and highly impermeable soils. It also blooms approxi- mately a month earlier and morphologically is distinguished by its larger, many-flowered heads. 9. A. rothrockii Gray 10. A. cana Pursh 10 (a). A. cana subsp. bolanderi (Gray) Ward I0(b). A. cana subsp. viscidula (Osterhout) Beetle, comb. nov. Based on A. cana var. viscidula Osterhout, Bull. Torrey Club 27:507. 1900. A. viscidula (Osterhout) Rydberg. Bull. Torrey Club 33:157. 1906. This subspecies has a distinct geographical range from 4. cana subsp. cana. It occurs in the high mountain valleys of the Rocky Mountains and is distinguished by the smaller, dark-green leaves, which are frequently asymmetrically lobed. I. A. argilosa Beetle sp. nov. Planta inter A. cana subsp. viscidula et A. longiloba intermedia: rami erecti, ad 1 m. alti; foliis ad 4 cm. longis, tripartitis profunde singuli parte 2-3 mm. lati; florum 5-10, 3.0-3.5 mm. longum; achenium 1.8 longum. mm. longum; achenium 1.8 mm. longum. Type collection: Colorado, Jackson County, Coalmont, July 31, 1957, A. A. Beetle 12872, TYPE in the Rocky Mountain Herbarium. This plant has deeply threc-Iobed leaves, very much of the gen- eral appearance of those of 4. tripartita. ‘They are, however, commonly up to 4 cm. long, and not only the basal portion but each lobe is 2-3 mm. broad. The plants occur on strongly alka- line soil associated with greasewood (Sarcobatus vermiculatus) and saltsages (Atriplex spp.). The plants are erect in habit and approximately | m. tall. All the collections of this plant are from a very limited area in the vicinity of Coalmont. A detailed study of the morphological characters of this plant indicates that they are intermediate between those of 4. cana subsp. viscidula and those of A. longiloba, both of which occur in the vicinity. This 1959] Beetle — New Names of Artemisia 85 new species supposedly of hybrid origin does not now occur mixed with its theoretical parents, and intergradation of the kind so common in this group of species of Artemisia does not occur in the area. The work reported here was sponsored by the regional research program, W-25, entitled Ecology and Improvement of Brush In- fested Range Lands. — WYOMING AGRIC. EXPER. STA., LARAMIE, WYOMING A New VARIETAL COMBINATION IN OXYBAPHUS. — In a pre- publication review of the writer's mss. on west-American range forbs the eagle eye of Dr. S. F. Blake has detected an improperly published new combination in Oxybaphus, a defect which this note assays to correct: Oxybaphus linearis var. subhispida (Heimerl) Dayt., comb. nov. Mirabilis linearis subhispida Heimerl, Ann. Conserv. & Jard. Bot. Genéve 5: 186. 1901. Allionia linearis subhispida Standl., Contrib. U. S. Nat. Herb. 12: 342. 1909. A. gausapoides Standl., Contrib. U. S. Nat. Herb. 13: 406. 1911. A. subhispida (Heimerl) Standl., Contrib. U. S. Nat. Herb. 16: 120. 1913. This plant's hairiness seems to rate hardly more than varietal relationship to typical Oxybaphus linearis (Pursh) Robins. (syn. Allionia linearis Pursh), as Heimerl indicated when he first de- scribed this entity. The generic status of Oxybaphus seems now to be widely recognized; its fruit and floral characters suggest a closer relationship to Mirabilis than to Allionia. O. linearis var. subhispida occupies rather dry to medium moist sandy or grav- elly soils but sometimes also heavy clays and moist rich loams, from "desert" areas to the ponderosa pine type, often partly pro- tected such as under mesquite bushes or canyon cottonwoods. Its range, not too well known, is from extreme southwestern Colo- rado, New Mexico and western ‘Texas south into Mexico. As a rule it is eaten little, if at all, by domestic livestock. However, there appear to be some exceptions. It is reported as common on shale banks about 9,000 feet on the Montezuma National Forest (southwestern Colorado) and there eaten with some relish by cattle. — WILLIAM A. DAYTON, ARLINGTON, VA. Volume 61, Number 722, including pages 35-63, was issued 27 April, 1959 JUL 2 3 1253 Dodota JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor-in-Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR, IVAN MACKENZIE LAMB Vol. 61.. April, 1959 No. 724 CONTENTS: The Genus Piricauda (Deuteromycetes). Roya AEM OOne ESO ene eiie ET OO TS 87 Naturalized Mugo Pine in New Hampshire. Radcliffe B. Pike and A. R. H. Hodgdon . 121 Discovery of Arenaria marcescens in the United States. ,,,, r D SE 123 New Forms of Trillium recurvatum. erm “r 124 A Modern Style in Floristic Studies. John H. Beaman (Review)... (ſͥꝓlſIjiIi . 125 Botanical Museum, Oxford St., Cambridge 38, Mass. RHODORA.—A monthly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid, in funds payable at par in United States currency in Boston; single copies (if available) 60 cents. Back vol- umes 1—58, with a few incomplete, can be supplied at $5.00 per volume. Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes, relating directly or indirectly to the plants of North America, will be considered by the editorial com- mittee for publication. Articles concerned with systematic botany and cytotaxonomy in their broader implications are equally accept- able. All manuscripts should be double-spaced throughout. Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance, will be furnished at cost. Address manuscripts and proofs to Reed C. Rollins, Gray Herbarium, 22 Divinity Avenue, Cambridge 38, Mass. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill, Botanical Museum, Oxford Street, Cambridge 38, Mass. Second Class Postage Paid at Boston, Mass. EATON PRESS, Inc. Specialists in Learned Publications 17 SPRING STREET WATERTOWN 72, MASS. Nhodora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 61 April, 1959 No. 724 THE GENUS PIRICAUDA (DEUTEROMYCETES) Rovarr I. Moore In a previous paper (9) the genus Piricauda was emended and set in sharp contradistinction to other genera that had been immixed under the names, sensu Saccardo, Sporidesmium and Stigmella. This current report is limited to a monographing of those species that are considered to fall within this reestab- lished concept of Piricauda. The nearly 250 fungi that have been given the name of Sporidesmium range from ?insect eggs or feces (Sp. epicoccoides) to a Phycomycete (Sp. aurantiacum) to a lichen (Sp. scutellare) to a very heterogeneous mixture of Deuteromycetes. Of this last named group the species of Sport- desmium and allied Phragmosporae have been monographed by Ellis (1). ‘These and a number of other species can readily be excluded from Piricauda, but there are certain forms that re- quire close scrutiny in order to be distinguished. ‘These include the similitudes of the closely appressed conidia of Cheiromyces and Dictyosporium, the sorosporoid "conidia" and bulbils of such Mycelia Sterilia as Papulospora, disrupted, intercalary chlamydospores of old somatic hyphae, certain alternarioid fungi, and the genus and subgenus Stemphylium and Pseudos- temphylium! respectively, but which are unique in their manner 1'5. . . The subgenus Pseudostempbylium of Stempbylium . .. is not a true Stempbyl- ium in the original sense. Continued growth of the conidiophore after production of the first spore is mot up through the first scar (as in Stemphylium) but laterally so that the older conidiophores are geniculate with a spore-scar more or less at the side of each bend." (E. G. Simmons, personal communication. ) 87 88 i Rhodora VOL. 61] of conidiation (e.g. cf. Fig. 14). Further, colony morphology is paramount and serves to distinguish Piricauda from Berkleas- mium and Steganosporium. However, these considerations have already been extensively dealt with and here we are concerned only with the genus Piricauda. Of the about 160 species of clavisporum for P. stigia and Hysterium karstenii for P. nitens; ined, only 32 are considered to properly belong within Piricauda, and to this number are added six new species. Perfect stages are known for only two of the species presented here: Glonium clavisporum for P. stigia and Hysterium kastenii for P. nitens; Lohman (6) also figures a Piricauda sp. stage for Hysterium hyalinum. Several additional, general, remarks should be made about the genus. Unlike Stemphylium, Alternaria, Sporidesmium and Papulospora there appear to be no real parasites in the genus. In the few instances where parasitism is suggested it is very mild and the fungus may very well be acting only as an epimycete. Another distinguishing feature is that in most of the species the conidia are formed as modified hyphal termini and in contra- distinction to the above and other similitudinous genera an abscission septum is not formed, but rather the conidia break off irregularly and not infrequently are subtended by a remnant of fertile hypha. A notable exception to this is observed in P. serendipita in which the side walls of the penultimate cell of the conidiophore break down and the discharged spore bears with it the ultimate conidiophore cell. While Piricauda is of little ter- restrial economic importance, it is "very abundant in the north- ern marine area"? and may be another of those Deuteromycetes that are believed to be primary in the initial softening of woods prior to marine borer invasion, (but also cf. Ray and Stuntz (11)). This monograph has been divided into three main sections: I. A key to species that provides in its flow a complete descrip- tion of each species and to which is appended a glossary provid- ing the interpretation of certain terms as used in the key. ?S. P. Meyers, personal communication, 1959] Moore — Genus Piricauda 89 II. In this part is to be found the formal taxonomy, figure references and any additional notes or comments. In brackets on the line below each species name is a formula setting forth the key steps that encompass the description of the species so that for any given species the description may be read through with ease. For further convenience the species have been com- piled alphabetically. It is worth noting here that three of the new and several of the other species are in culture and grow and sporulate readily on a number of media. One can not but feel a sense of loss that other species are not also in culture, particularly such exquisite forms as P. ulmicola, P. paraguayensis and P. curvata. Since most are saprobes and the few, at most, weak parasites, it is believed that most could be established in culture. Further, living material of P. fusus would provide a most interesting study to determine whether the crystals are indigo. There are only, to my knowledge, four other ‘instances of fungi producing similar crystals: Helicoma asperothecum, Helicoma recurvum, Helicosporium elinorae and a mutant strain each of Schizophyllum commune and Sch. umbrinum, belonging to the collections of Prof. J. R. Raper. Only in the last two species is it known for a fact that the metabolite is indigo, the other three species being represented solely by herbarium speci- mens. In an appendix to this section, Species Inquirendae, are listed those species of Sporidesmium and Stigmella which have not been available for study and whose final taxonomic placement is not, therefore, possible at this time. III. This part is limited to the plates of figures. Attention is drawn to the fact that these are all to the same scale, serving to emphasize the wide divergency in specific morphology, partic- ularly size, and, further, it stresses the previously made observa- tion (9) that the overall taxonomy of this group is best served by a unitary approach. Most of these photomicrographs were taken with an Exa camera using a Bellowscope attachment and Kodak Panatomic-X film with several different filters. Further, a number of the slides were stained with phloxine or lacto-fuch- sin to emphasize hyaline structures. Therefore, one is cautioned 90 Rhodora [VOL. 61 that the apparent degree of darkness observed in the plates is not necessarily a true index of the degree of pigmentation. I. KEY TO SPECIES l. Mature conidia, by transmitted light, translucent, melleous to fuscous s 8 19 J. Mature conidia, by transmitted light, opaque to subopaque.... 2 2. Conidia up to about IScelledP l.. 12 2. Conidia multi-cellular, sessilUUUP ... 3 3. Conidial profile crenulate LL aaa aaaea 11 3. Conidial profile regulala e.. 4 4. Conidia subglobose to elliptical to obovate to ovate to oblong to oblong-ellipsoid, glabrous to asperate, never striate 8 41. Conidia spherical to globose, laevigate, or, if obovate to oval, striate 5... eiie 5 5. Conidia with hyaline basal cells or with translucent subtending cells ies eens e adn tatsed meses a X TN 5. Conidial base consistent with the rest of the spore........... 6 6. Conidia laevigate, spherical, opaque, 44.558 (C63) u diam- eter AMEN IM P. globifera 6. Conidia closely covered with raised longitudinal wavy and dark ridges, obovate to oval, subopaque, 23.5-35 x 44.5-52.5u. . . P. striata 7. Conidia with hyaline basal cells, globose to subglobose to pyri- form, subopaque, 16-20 x 20-260W20n. P. suffulta 7. Conidia subtended by a few to several translucent, supernu- merary cells; primary portion spherical to globose, opaque, (15.5-) 18.5-23.54 diameter P. melanopus 8. Conidia scabrous, cellulation imperceptible, regular in form though the bilateral halves may be subequal, umbilicus stout, centric ee bt ee cen ENE VE oe ey i 10 8. Conidia laevigate or asperate, conspicuously cellular, tending to be irregular of form, subglobose to elliptical to oval to obovate, umbilicus stout, may be obcentric............... 9 9. Conidia glabrous, oblong to elliptical to obovate to subglobose, umbilicus frequently obcentric, 15.5-36.5 x (25.5—) 29-39.5 (—42-52.5) Wo ...d P. putredinis 9. Conidia and umbilicus papillose to subtuberculate, subglobose to oval (to obovate to pyriform) , (23.5—) 29-43 (-47.5) x (26-) 33.5-58 (—71) ky— eee eee P. aspera 10. Conidia 52-83 x 104-155 (C192) u, (X71.2 x 135.2), (obo- vate to) oblong to oblong-ellipsoid (to ovate) ...... P. stygia 10. Conidia (23.5-)31.5-36.5 (-53) x 31.5-79u, (X 35.1 x 50.50%, 1959] Moore — Genus Piricauda 91 (globose to) ovate to ov all. P. bogoriensis ll. Basal and distal ends truncate, conidia oval, deep fuscous to opaque, cells very consistent, 23.5-35 x (34-) 44.5-50u P. elliptica 11. Basal and distal ends rounded, conidia globose to oval, cells quite variable in size, 23.5-68.5 x 26.5-79 (-108) u. (X 45.5 J ͤ K ee P. composita 12. Conidia borne on prominent conidiophores, glabrous........ 18 12. Conidia sessile or nearly $0......... e f e n 13 13. Conidia not at all constricted at the septa, tending to be poly- mor phie 17 73. Conidia in some degree constricted at the septa, regular and consistent in hape: 14 14. Mature conidia differentially pigmented, pyriform, the upper, spherical portion opaque to subopaque, the lower or basal cells fuscous; colonies at maturity with both sessile conidia and those which are borne on conidiophores of several cells; le nh p REN 16 14. Mature conidia uniformly pigmented, depressed or subglobose, not pyriform; terrestrialll +. n tte: 15 15. Conidia depressed, composed generally of 4-6 radially arranged cells, early dark-translucent and 18.5-26.5u diameter, at maturity opaque, up to 40u diameter and not infrequently somewhat laterally compressed, borne acrogenously upon short laterals of the netlike somatic mycelium which also produces peltate hyphopodia ......... P. manilensis 15. Conidia subglobose, not depressed, composed of about 10-15 cells, dark-translucent, 18.5-26.5 K (21-) 24-34... P. vernoniae 16. Somatic hyphae thin walled, subhyaline with a greenish-gray cast, producing abundant intercalary chlamydospores, phrag- mous to dictyous; conidia 5-10 celled, opaque to subopaque in the globose portion, fuscous below, 12.5-20 x 21-24.5 (27.5) ů nn „„ P. arcticoceanorum 16. Somatic hyphae thick walled, fuscous, without chlamydospores; conidia 5-10 celled, opaque except for the basal, fuscous cells, (17-) 19-22 (-31) x (34-) 36-41 (744) u ... P. pelagica 17. Somatic mycelium fuscous, monilioid with cells spherical or bacilliform to regular, walls smooth to scrupose to, in places, up to 2u thick and scrobiculate; conidia opaque to dark- translucent, reddish-brown by strong transmitted light, sub- globose to suboval, 10.5-18 x 13.5-19u, (X 14.4 X 16.7) ; growing saprobically on old coniferous Woo! P. nodosa 17. Somatic mycelium hyaline, regular, thin walled (in the hair cells) or (on the leaf surface) forming rough, thick walled, intercalary chlamydospores that are a translucent gray-green 18. bo ws Rhodora VOL. 61 (Storm Gray of Ridgway); conidia opaque to a dark-translu- cent, dull green (Deep Slate-Green of Ridgway) by strong transmitted light, subglobose to obovate to oblong (to clav- ate), frequently subtended by a single proliferating cell, borne on a short conidiophore of a couple of cells, or sessile, 9.5-13.5 (-17) Xx (10.5-) 12-25 (-30) u (X 12.2 X 1770; hollowing out, and conidiating profusely on, the leaf hairs of Anona cherimolia, on which it is a ?casual parasite P. trichophila Conidiophores long-flexuous, 50-90; high, tapering from about Su thick distally to about 2.54 basally; conidia globose to pyriform, paucicellular, 18.5-24 x 24-31.5u.. P. paraguayensis Conidiophores 21-31.5u high, about 5u broad throughout their length; conidia somewhat obovate, multicellular, the basal one or two tiers of cells tending to remain translucent until maturity, 13-18.5 x 21-26.51, disjuncting bv the dissolution of the penultimate cell of the conidiophore.... P. serendipita Conidia multicellular, glabrous . MEME 33 Conidia paucicellular, up to 15 cells.. 20 Conidia lacvigatdtgagagdlJnla sou. 29 Conidia asperate 00. . 2l Conidia staphyloid, each cell partially spherical s 28 Conidia regular, with very little or no constriction . 22 Conidia with single distal and basal cells, walls thick. may be produced on conidiophores 9.00.00. 0 — MM 27 Conidia without distinct terminal cells, sessile... . 223 Conidia generally bearing distally 3 prominent hyaline papil- lae, the rest of the spore glabrous, up to about 15 cells, globose to broadly oval, 10.5-14 x 11.5-16(-19)u, width may exceed length |... 8 — P. trigonella Conidia uniformly asperate 9000. 2344 Conidia fuscous, composed generally of up to 5 or 6 quite reg- ular cells, walls thin and aculeolate or thick and papillate, umbilicus centric |... 0 0 ooo MEN .. 26 Conidia melleous, composed of up to 10 or so somewhat irreg- ular cells, walls SCTUPOSE ..... 5. iss ss ssh 25 walls EN ä P. quadrata Conidia subglobose to obovate to suboval, 8.512 x (10.5-) 12.5-18u, linear, umbilicus seldom present, walls thin. . mE PD. tumulosa Conidial walls thin, aculeolate; conidia obovate to subovate 1959] n2 -I 907 af. 28. 33. B. Moore — Genus Piricauda 93 to oval to subglobose, 11.5-17 X 16-24.5u, initially aseptate, becoming phragmous and late paucidictyous P. sarkara Conidial walls becoming very thick, papillate; conidia pyriform, early dictyous, 21-32 * 26-40 353 ae P. damonis Conidiophores absent; conidia globose to oval, subfuscous, scrupose, tending to become glabrous with age, 6-8 cells, may be undulant at the septa, walls up to 5.5u thick and may in places be perimetrically cracked, 17-27.5 x 22.5-30u Mere P. sacchari Conidiophores present, hyaline, up to 4 cells long, 2-4u broad, stoutly attached to the tending-to-be-somewhat-squarish basal cell; conidia oval, submelleous, aculeolate, 5-8 cells, tending to become notched at the septa, particularly the mesial, walls up to ou thick, 10.5-15 x 21-28.5u i P. chartarum Conidia strongly tuberculate, subglobose to pyriform, about 5 cells, tapering into the fertile hyphae, 9.5-16 x 6.5-20u, width may exceed length: hh P. exasperata Conidia spiculate to papillose, subglobose, 5-10 cells, not bas- ally tapered but rounded, umbilicus often obcentric, 12.5-19 x 12.5-24.5u, width may exceed length.........-. P. funerea Conidia broadly oval, or fusiform, seldom constricted at the c] ↄ ets ow ke eg he E 32 Conidia subglobose, or elliptical to oval to clavate, generally constricted at the sept. 8 0 Conidia subglobose, about 3-5 cells, (7.5-) 10.5-14 x 12.5-17 D A ys Cos} st Y P. apheles Conidia oval to elliptical to clavate, 6-lorcells T a 31 Conidia basally and distally didymous, centrally dictyously quadri-, rarely hexa-, partite, elliptical to slightly clavate, (13.5-) 18.5-21 X (34.5—) 47.5 (-52.5) 1 P. scorobylos Conidia elliptical to oval to clavate, may be partially curved, 6-8 cells, 8.5-16 x (13.5-) 21-24.5 (32) Bee P. viticola Conidia fusiform; terminal cells difficult to distinguish, conic, hyaline; 15.5-18.5 X 60.5—76.5u. 10-15 cells. Pe. pulchella Conidia broadly oval, median septum most prominent, 3-5 cells, 6-8.5 (-9.5) x (7.5-) 9.5-1 IE uni- and bi-cellular spores common .........- eee eee ees Pe. subcuticularis Length up to half again the width, generally less; conidia gla- brous, globose to oval-ellipsoid, melleous to fuscous, peri- metric cells tending to be integumentoid and somewhat con- stricted at the septa, 18.5-37 X 26.545. (X 25.2 X 35.6) P. nitens Conidia elongate, length no less than twice the width..... . 34 94 Rhodora [VOL. 34. Conidia conspicuously constricted at one or more of the prom- inent horizontal septa, cylindrical-ellipsoid to oblong-ellip- soid, frequently with a slight distal tapering, may be slightly curved, strongly undulate to constricted at the up to 6 pri- mary horizontal septa, 18.5-26.5 x (44.5) 58-68.5u 61 ä P. heteromera 34. Conidia not conspicuously constricted at the septa, though profile may be slightly undulate 35. Conidia fusiſowi ii. 35. Conidia lageniform or oblong, sessile 36. Conidia tending to be lageniform, the wider, muriform base developing first, followed by the growth of the slenderer ros- trum that may become paucidictyous and strongly bent, up to 80-110 x 10.5-18.5u at the base, 7.5-10.5(-13) u distally. Se P. curvata 36. Conidia oblong, multicellular, 18.5-21 x 5560.50. De. itochna 37. Conidia sessile, covered by a thin separable layer of purplish iridescent crystals, 29-34.5 x (73.5—) 8497.50 37. Conidia borne on slender conidiophores, 15-304 high, that en- large continuously into the spore base, without crystals, ta- pering terminally into a short, hyaline rostrum, initially hya- line, phragmous, upon maturation expanding from the width of the fertile hyphae to broadened, fuscous, dictyous, 13.5-16 BEEN P. fusus x 44.5-76u ee. P. ulmicola GLOSSARY asperate — bearing projections or points; (n.b: all terms applying to the conidial surface are applied as that surface is seen by oil-immer- sion magnification, X 1500) . aculeolate — having somewhat spine-like processes. papillose — having minute nipple-shaped projections, scrupose — covered with very small points. spiculate — bearing minute slender pointed projections. tuberculate — bearing wart-like processes. -ellipsoid — sides parallel and ends almost hemispherical. lageniform — (florence) flask-shaped. scrobiculate — pitted, furrowed. staphyloid — resembling a compact bunch of grapes. umbilicus — the attachment or remnant of the fertile hyphae of sessile conidia. X — the statistical mean. II. TAXONOMY PiRICAUDA Bubák. Stigmella sensu Saccardo, Michelia 1: 264. 1878. 1959] Moore — Genus Piricauda 95 Sporidesmium sensu Saccardo, Michelia 2: 23. 1882. Monodictys Hughes, Can. Jour. Bot. 36: 785. 1958. Piricauda apheles sp. nov. Fig. 28 I, 79, 20, 29, 30] Conidia in mycelio sessilia, cellulis paucis (circ. 3-5) composita, sub- globosa, mellea, ad septa saepe nonnihil constricta, (7.5-) 10.5-14.0 x 12.5-17.0 (-23.5) f. In ligno putrido, Herb. K. Holotypus; preparatio microscopica RTM I: 19541, Isotypus. (Etym. adeAns — even, smooth, simple, in reference to the conidia.) Herb. K has a number of specimens labeled Sporidesmium Lepratia, none of which is labeled or recognizable as the type. These collections are composed of a great variety of fungi—Dem- atiaceous, acervulate with monacrogenous conidia, and sporo- dochial with catenate conidia. However one collection, prob- ably a Cooke specimen, is on a small card and bears two pieces of wood. It carries as its only notation “Sporidesmium Lepraria,” and, between the two pieces, sketches of a few spores. Of these two pieces the lower bears only what appears to be old hyphae, but the upper bears a good Piricauda. This is given a separate name because 1) the figure of Sp. Lepraria var. nigerrima (which is stated to differ only in its darker color) is multicellular, 2) Sp. Lepraria Berkeley is a nomen confusum and 3) this material is not part of the typification of 5f. Lepraria. Piricauda arcticoceanorum sp. nov. Fig. 47 UI. 2, 12, 13, 14, 16] Coloniae nigrae; hyphae praesertim subterficiales, parietibus tenuibus, sed chlamydosporeas intercalares copiose proferentes quae primum parietibus incrassatis indutae sunt; mycelium fuscescens, e cellulis sub- globosis deinque phragmoideis, dictyoideis et subopacis; conidia pyrifor- mia, primum fuscidula, e cellulis 5-10 composita, maturitate parte ter- minali sphaerica opaca et parte basali fuscescenti, saepe e cellula singula inflata oriunda; 12.5-20.0 x 21.0-24.5 (27.5) uu. sessilia aut ad coni- diophoros paucicellulares enata. In ligno putrido in mare immerso, Argentia, Terra Nova. Cultura dessicata et preparationes microscopicae in Herb. FH, Holotypus; cul- tura viva, S. P. Meyers F-30 et preparatio microscopica RTM 1:259, Isotypi. (Etym. arcticus + oceanorum — of the northern seas, in reference to its habitat.) 96 Rhodora [VOL. 61 This species was communicated by Dr. S. P. Meyers and it is considered. to. be conspecific with three other cultures com- municated from him: F-23 from Nanaimo, British Columbia, F-65 from Kodiak, Alaska and F-73 from Halifax, Nova Scotia. All were isolated from submerged wood taken from the sea. Piricauda aspera (Corda) comb. nov. Fig. 24 / 2. 3. J. 8, 9] Sporidesmium asperum Corda, Icones Fung. 2: 6. 1838. Glasterosporium asperum (Corda) Saccardo, Syll. Fung. 4: 383. 1886. Stemphylium phaeosporum de Notaris. Comm. Soc. Crit. Ital. 2 (1): 81. 1864. [Ex descript.: non vidi.] Sporidesmium phaeosporum (de Not.) Saccardo, Syll. Fung. 4: 197. 1886. Material examined: PR. 155653. (Corda coll.) , Type, on fallen twigs of Fagus silvatica, Brezina, Czechoslovakia. (slide RTM 1:144). PAD (Saccardo coll. 1448), on defunct wood of Populus tremula, Riva, Val- dobbiadene, Italy, (slide RTM 1:216). Piricauda bogoriensis (Penz, & Sacc.) comb. nov. Figs. 30, 31 /. 2, 3, J. 8. 10] Sporidesmium bogoriense Penzig & Saccardo, Malpighia 15: 248. 1901. Monodictys bogoriensis (Penz. & Sacc.) Hughes, Can. Jour. Bot. 36: 785. 1958. Also present, but rare, on this material are clavate phrag- mospores (Fig. 31); from the evidence of this mount the hypo- thetical relationship. that suggests itself is that the abundant spores described and figured by Saccardo are actually pedicelled sclerotia of a Sporidesmium type fungus. Material examined: PAD (Saccardo coll.) Type, Bogor, Java, on defunct petioles and blades of palm (slide RTM 1:211). Piricauda chartarum. (B. & C.) comb. nov. Figs. 19, 20 1. 79%. 20, 21, 22. 27] Sporidesmium chartarum Berkeley & Curtis, apud Berkeley in Grev- illea 3: 50. 1874. Sporidesmium bakeri [var. bakeri] H. & P. Sydow, Ann. Myc. 12: 204. 1914. The combination Sporidesmium bakeri var. maydicum [Clas- terosporium maydicum Sacc.] was made by Hughes (3) and he cites C, F. Baker's Fungi Malayana 217 as the type. However, 1959) Moore — Genus Piricauda 97 the type cited by Saccardo is material collected by Baker from Los Baños, Philippines, and numbered 3733. Hughes does not mention this collection and whether 217 is a part of it is not known. Hughes's drawings, though, of conidia of 217 in his Fig. 38A are virtually identical with those sketched by Saccardo on his packet. Both are uniformly biseptate, but Saccardo records his conidia as 14-15 X 6u while Hughes records those of 217 as being 15-20 X 6-10u. It is very disconcerting, then, that numerous mounts of 3733 have totally failed to elicit any conidia of this type. The material does have a fair quantity of amero- spores of the type Hughes depicts in his Fig. 38B & C and very rarely these are didymous. These spores are about half the size noted by Saccardo, but they could be immature. Also present on this material are 1) long, dark, mycelial processes, ?conidio- phores, emerging from the stomata, 2) an immersed, fuscous, pycnidial fungus that produces abundant, melleous, acerous phragmospores, and 3) small patches of a Curvularia. If 217 is part of 3733 these latter fungi should also be present. While 217 as reported by Hughes bears only tricellular conidia, Hughes considered it a Sporidesmium (sensu Saccardo) because other collections that bore comparable conidia upon aging become dictyous by a single, medial, vertical septum. Further, Hughes based the conspecificity of var. maydicum with var. bakeri on the characteristics of two 1949 collections identified as var. bakert on leaves of Bridelia ferriginea (Euphorbiaceae) and Zea mays fron. Hohoe and Bisba respectively. But the tvpe collection, made in 1913, is on leaves of Musa sapientum from Los Banos, Philippines. Examination of the Sydows’ type material presents a fungus quite different from that figured by Hughes and Sac- cardo, both in its larger size and different shape. Further, it is not satisfactorily distinct from the earlier Sporidesmium charta- rum with which it is here synonymized. While a disposition of Cl. maydicum is not presently possible, it can not be maintained with Piricauda chartarum. Hughes (4) in his recent paper trans- fers this species as Scheleobrachea maydica. Material examined: FH (Curtis coll, Car. Sup. 6419), Type, on de- cayed paper, Hillsborough, N. C. (slide RTM I:11). S (Sydow coll.) 98 Rhodora [VOL. 61 on defunct leaves of Musa sapientum, Los Banos, Philippines, C. F. Baker 1728, type of Sp. bakeri (slide RTM 1:105). PAD (Saccardo coll. 3733) , on defunct leaves of Zea mays, Los Banos, Philippines, type of Cl. maydicum (slide RTM 1:217). QM numbers QM7051, QM7102 and QM7140 are identified as P. chartarum. Piricauda composita (Berk. & Rav.) comb. nov. Fig. 44 J. 2, 3, 11] Sporidesmium compositum Berkeley & Ravenel, apud Berkeley in Grevillea 3: 17. 1874. Sirodesmium compositum. (Berk. & Rav.) Saccardo, Syll. Fung. 4: 517. 1886. Material examined: FH (Curtis coll.) , on defunct wood of Catalpa cordifolia, Santee Canal, S. C. (Rav. 1801) ; on defunct wood of oak, Cotoos Springs, Hendersonville, N. C. (Rav. Car. Sup. 4441) : Syntypes (slides RTM I:13a/b resp.) Piricauda curvata (B. & C.) comb. nov. Figs. 32, 33, 84 J. 19, 33, 34, 35, 36] Sporidesmium curvatum Berkeley & Curtis, apud Berkeley in Grev- illea 3: 50. 1874. Clasterosporium curvatum (B. & C.) Saccardo, Syll. Fung. 4: 385. 1886. Hughes (2) has designated material in Herb. K. labeled by Currey "Sporidesmium curvatum B. & C. — on Crataegus — Ex herb. Berkeley" as the specific lectotype. His examination of this material showed it to produce 2-armed conidia representative of Hirudinaria macrospora. However, examination of holotypic material evidences lageniform conidia basally attached which clearly places the species in Piricauda. Material examined: FH (Curtis coll.) , car. Sup. 2561, Holotype, on Crataegus leaves, mountains of North Carolina (slide RTM 1:14). Piricauda Damonis sp. nov. Fig. 25 [J. 79, 20, 27, 22, 23, 94, 26] Conidia fuscidula, papillata, ovoidea, obovoidea vel pyriformia, e cel- lulis circ. 6 composita, parietibus crassis (usque ad 3u), sessilia et e mycelio prostrato copiose prolata, 21-32 x 26-40. In cultura (agar-agar) contaminata in Herb. QM. Cultura dessicata et preparatio microscopica in FH, Holotypus; cultura viva OM 646 et preparatio microscopica RTM 1:255, Isotypi. (Etym. In honor of Samuel C. Damon, Deuteromycetologist.) 1959] Moore — Genus Piricauda 99 Piricauda elliptica (Cke.) comb. nov. Fig. 21 po» » 19 Sporidesmium ellipticum Cooke, Grevillea 12: 28. 1883. Sirodesmium ellipticum (Cke.) Saccardo, Syll. Fung. 4: 517. 1886. As noted by Cooke, the production of conidia is concentrated around the eruptions of Diatrype disciformis, perhaps because the wood is more degradated in these areas. I would not say that the one is the imperfect stage of the other. Material examined: NY, on defunct wood of Magnolia glauca, Pinop- olis, S. C., Rav. Fung. Amer. exs. 562, Isotype (slide RTM 1:32). Piricauda exasperata (Ellis & Barth.) comb. nov. Fig. 42 [I. 19, 20, 21, 28] Sporidesmium exasperatum Ellis & Bartholomew, Erythea 4:29. 1896. Material examined: BPI, on wood from an oak barrel-bottom in a cellar, Rockport, Kansas, Barth. 1461, Type (slide RTM 1:63). Piricauda funerea (Ellis & Langl.) comb. nov. Fig. 9 [I, 19, 20, 21, 28] Sporidesmium funereum Ellis & Langlois, apud Ellis and Everhart in Jour. Myc. 4: 124. 1888. Material examined: BPI, on rotten pieces of an old coffin taken from a brick tomb, Pointe a la Hache, Louisiana, Langl. 1456, Type (slide RTM I:64). Piricauda fusus (B. & C.) comb. nov. Fig. 35 [E-19, 92; 34, 85, 9H Sporidesmium fusus Berkeley & Curtis, apud Berkeley in Grevillea 3: 50. 1874. Material examined: FH (Curtis coll. 3322), Type, on defunct wood of Magnolia acuminata, Virginia mountains (slide RTM 1:16). Piricauda globifera (B. & C.) comb. nov. Fig. 10 IJ. 2, 3, 4, 5, 6] Sporidesmium globiferum Berkeley & Curtis, apud Berkeley in Jour. Linn. Soc., London 10: 354. 1869. Material examined: FH (Curtis coll), on rotten logs, Cuba, C. Wright 566 (=B. & C. Fung. Cuba 579), Type (slide RTM 1:21). Piricauda heteromera (Kirsch.) comb. nov. Figs. 17, 18 II. 19, 33, 84] Sporidesmium heteromerum Kirschstein, Hedwigia 81: 202. 1944. Material examined: B, on Juncus filiformis, upper valley of the Eder near Lützel, Siegen County, Germany, Type, (slide RTM 1:84). 100 Rhodora B [VOL. 61 Piricauda itochna sp. nov. Fig. 27 (J. 19, 53, 34, 35, 36] Conidia in mycelio sessilia, oblonga, cellulosis numerosis composita, laevigata, mellea, 18.5-21.0 & 55.0-60.5u. In ligno putrido una cum FHelicoma acrophalerium Moore, Porto Rico, 24 January ad 5 April 1923, Fred J. Seaver et Carlos E. Chardon 580, Typus, in Herb. NY; preparatio microscopica RTM 1:128, Iso- typus. (Etym. rov + xvos — fungal porous substance, in reference to the conidia.) Piricauda manilensis (Sacc.) comb. nov. Fig. 29 [7, 2, 12, 13, 14, 15] Stigmella manilensis Saccardo, Ann. Myc. 11: 320. 1913. Stigmella palawanensis H. & P. Sydow, Philip. Jour. Sci. Bot. 9: 189. 1914. The Sydows state that palawanensis differs from manilensis "by the thinner and longer hyphae provided with numerous hypho- podia, and net-like mycelium on black, patch-like colonies that translucent it is probable that it is not as mature as manilensis which has opaque conidia. As further evidence of this the conidia of palawanensis are not observed to have more than 5 lobes per spore while there are conidia on manilensis with 6 lobes. Both collections produce morphologically comparable conidia, hypho- podia, and net-like mycelium o nblack, patch-like colonies that are readily separated [rom the substrate as small plaques. Material examined: PAD (Saccardo coll. 256), FH (Bartholomew coll., Sydow — Fung. Exotici exs. 198) , Syntypes, on dead pods of Cassia tora, Manila, Luzon (slides RTM 1:214, E215 resp). S. on living leaves of Celastrus paniculatus, Taytay, Palawan, Merrill 8832, Type of St. palawanensis (slide RTM 1:103). Piricauda melanopus (B. & Br.) comb. nov. Fig. 36 (J. 2, 3, 4, 5, 7] Sporidesmium melanopum Ach. ex Berkeley & Broome, Ann. & Mag. Nat. Hist. 5(2 ser.) : 459. 1850. Spiloma melanopa Acharius, Methodus qua omnes detectos Lich- enes, p. 10. 1803. The Herb. K folder has two sheets bearing seven collections each that in all consist of one acervulate and several different Dematiaceous fungi. The type, though, is limited to one identi- 1959] Moore Genus Piricauda 101 fiable collection on apple bark on the label of which appears the name of W. Borrer, corresponding, thus, to the type descrip- tion's statement that the fungus is "Common on the bark of Apple-trees" and further to the mention that "We are indebted to Mr. Borrer for authentic specimens." Finally, the several sketches accompanying this material agree with the characteris- tics set forth in the type description. The sheet bearing the type has two additional collections that are comparable, both on apple bark and identifiable by their collection localities — "Kings Cliff" and "Essex" respectively. The second sheet like- wise has two comparable collections, also on apple bark, and identifiable by their respective collector and locality — "Bloxam" and "Orchard Gopsal." Material examined: K, on apple bark from Sussex, leg. W. Borrer, Type: on apple bark, Kings Cliff, England; on apple bark, Essex; "fungus 23" from hills above Port Louis, Mauritius; specimen labeled “Lepraria nigra E. B.“; Car. Sup. 4448 identified as Halysium atrum Corda; specimen numbered 5829, on apple wood, New England, (slides RTM L19lag resp.). K, on apple bark, leg. Bloxam, England; on apple bark, leg. Bloxam, England: on apple bark, Gopsal, Leicestershire; “fungus 23" from hills above Port Louis, Mauritius; specimen labeled "Sporidesmium melanopum M, B”; two collections of Ravenel's Fungi — North America 3051, on rotting pine logs, Aiken, S. C.; specimen labeled “69 Victoria," (slides RTM I:192a-g resp.) . Piricauda nitens (Schw.) comb. nov. Figs. 11, 13 1. 19, 33] Sporidesmium nitens Schweinitz, Amer. Phil. Soc. Trans. II 4: 306. 1332. Clasterosporium nitens (Schw.) Saccardo, Syll. Fung. 4: 392. 1886. Monodictys nitens (Schw.) Hughes, Can. Jour. Bot. 36: 786. 1958. Sporidesmium paradoxum Corda, Icones Fung. 2: 6. 1838. Stemphylium paradoxum (Corda) Fuckel, Fung. Rhen. 1515, 1865. Coniosporium paradoxum (Corda) Mason & Hughes, CMI Myc. Pap. 37: 16. 1951. Monodictys paradoxa (Corda) Hughes, Can. Jour, bot. 36: 7806. 1958. Stigmella nemopanthis Dearness, Mycologia 16: 174. 1924. The three species placed together here present the following ranges in spore sizes: 102 l Rhodora [VOL. 61 nitens 18.5-21.5 x 29.0-39.5u, X 20.4 x 34.41; paradoxum 18.5-25.0 x 26.0-34.0u, X 21.8 x 29.8; nemopanthis 21.5-37.0 x 26.5 —45.0u, X 28.8 x 38.31. Considering that the collections are from different substrates and that they were inevitably subjected to different environ- mental conditions, e.g., moisture and temperature, it is felt that the combined measurements of 18.5-37 X 26.5-45u, X 25.2 X 35.64 are more meaningful than trying to establish dubious criteria for maintaining them separately. Hughes (4) has not examined the type of Sp. paradoxum but bases his judgment on Fuckel’s material and lists St. nemopanthis among its syno- nyms. He maintains nitens and paradoxum separate in Sub- genus 2 of his Monodictys for reasons unstated, though on the type material of the former species, in Herb. PH, he has written the following comment: ‘This is congeneric with Sporidesmium paradoxum Corda and only critically distinct from it." Lohman (8) in his elaboration of Hysterium karstenit: men- tions and figures a conidial stage, then assigned as Sporidesmium species. Using the measurements and characteristics apparent in his Fig. 2A to run the species through the key we readily arrive at P. nitens; a particular confirmatory character that shows up in these drawings is the integumentoid nature of the perimetric cells. Material examined: PH (Schweinitz Syn. Fung. 3082), Type, on the denuded stems and branches of Spirea opulifolia, Bethlehem, Penn., (slide RTM 1:72). PR, 515145, (Corda coll.), type of Sp. paradoxum, on old birch bark, Brezina, Czechoslovakia, (slide RTM 1:138). DAOM (Dearness coll. 3825), type of St. nemopanthis, on bark of defunct branches of Nemopanthes mucronata, (slide RTM 1:82 (=slide E. G. Simmons IX-15)). Piricauda nodosa (Preuss) comb. nov. Fig. 43 J. 2, 12, 13, 17] Sporidesmium nodosum Preuss, Linnaea 24: 103. 1851. Material examined: B, 4294, Preuss 1475, Type on wood of Abies from near Hoyerswerda, Silesia. FH, Klotzsch Herb. Viv. Myc., (au- thentic), from near Hoyerswerda, Silesia. (Slides RTM 1:89, 1:204 resp.) Piricauda paraguayensis (Speg.) Moore (9). J. 2, 12, 18] 1959] Moore — Genus Piricauda 103 Piricauda pelagica Johnson (5). tie, 12, 13, 14 26] Piricauda pulchella (Sacc.) comb. nov. Fig. 22 [l, 19, 20, 29, 32] Sporidesmium pulchellum Saccardo, Atti, Accad. Sci. Veneto-Trent- Istr. 10: 87. 1919. Material examined: PAD (Saccardo coll. 4445), on defunct branches of Sapindus saponaria, Los Banos, Philippines, (slide RTM 1:208). Piricauda putredinis (Wallr.) comb. nov. Fig. 37 D» 3 4 87 Melanconium putredinis Wallroth, Fl. Crypt. German. 2: 181. 1833. Monodictys putredinis (Wallr.) Hughes, Can. Jour. Bot. 36: 785. 1958. Sporidesmium polymorphum Corda, Icones Fung. 1: 7. 1837. Stemphylium polymorphum (Corda) Bonorden, Handb., p. 83. 1851. Sporidesmium fumagineum Saccardo, Nuovo Giorn. Bot. Ital. 24: 42. 1917. Material examined: STR, on worked wood on Abies, Type, (slide RTM 1:256). PR, 155661, (Corda coll), type of Sp. polymorphum, on defunct wood of Betula alba, Reichenberg, Czechoslovakia, (slide RTM I:140). PAD (Saccardo coll.), type of Sp. fumagineum, on dying branches of Populus tremula, Piccolo, S. Bernardo, Italy, (slide RTM I:221). Piricauda quadrata (Atk.) comb. nov. Figs. 45, 46 [1, 19, 20, 21, 22, 23, 24, 95] Sporidesmium quadratum Atkinson, Cornell Univ. Bull. 3: 40. June, 1897. Scheleobrachea quadrata (Atk.) Hughes, Can. Jour. Bot. 36: 802. 1958. Stigmella crataegt Ellis & Everhart, Torrey Bot. Club Bull. 24: 475. October, 1897. Stemphylium crataegi (E. & E.) Höhnel, Ber. deutsch. Bot. Ges. 2525515. -: 1918. Material examined: CUP (Atkinson coll), two collections on Cra- taegus leaves, Highland Park, Montgomery, Ala., Holotype (scr.), and Isotype (typ.) respectively, (slides RTM I:234b/a resp.. NY (Ellis coll), on leaves on Crataegus parvifolia, Newfield, N. J. type of St. crataegi, (slide RTM 1:61). Piricauda sacchari (Speg.) comb. nov. Fig. 15 [I. 19, 20, 21, 22, 27] Stigmella sacchari Spegazzini, Rev. Facultad Agron, y Veterin., La Plata 2: 251. 1896. 104 Rhodora [VOL. 61 Sporidesmium bakeri Syd. var. sacchari (Speg.) Hughes, CMI Myc. Pap. 50: 69. 1953. Material examined: LPS, 13054, (Spegazzini coll. Type, on leaves of Saccharum officinarum, Tucumán, Argentina, (slide RTM 1:98). Piricauda sarkara nom. nov. Figs. 6, 7 [J. 19, 20, 21, 22, 23, 24, 26] Sporidesmium sacchari Spegazzini, Anal. Museo Nacion. Buenos Aires 20: 443. 1910. Scheleobrachea sacchari (Speg.) Hughes, Can. Jour. Bot. 36: 802. 1958. (Etym. sarkara — Sanskrit paronym of sacchari.) Material examined: LPS, 13006, (Spegazzini coll), Type, on old culms of Saccharum officinarum, in fields near Ledesma, Argentina, (slide RTM 1:97). Piricauda scorobylos Moore (10). 1. 79, 20, 29, 30, 31] Piricauda serendipita sp. nov. Figs 3, 4 J. 2. 12, 18] Conidia cellulis numerosis composita, nonnihil obovata, laevigata, maturitate opaca vel subopaca (praeter series basales 1-2, quarum cellulae hyalinae manent), 13.0-18.5 x 21.0-26.5y. singulatim ad extremitates conidiophorum longit. 20-3156 prolata, et solutione cellulae penultimae conidiophori disjuncta. In cauli putrido Zea mays, in vasculo humido, Iowa City, Iowa, G. W. Martin 6454a, Holotypus, in Herb. LA; preparatio microscopica RTM 1.257 et cultura viva OM 7165, Isotypi. (Etym. Serendipity — the finding of valuable things not specifically sought for, in recognition that the desire for a specific determination led to, among other things, the ordering of its congeners.) Piricauda striata (Petch) comb. nov. big. 26 J. 2, 3, 4, 5, 6] Sporidesmium striatum Petch, Ann. Royal Bot. Gard. Peradeniya 6: 249. 1917. Material examined: K, on Hevea brasiliensis [rom Peradeniya, Ceylon, Cotype, (slide RTM 1:246). Piricauda stygia (B. & €.) comb. nov. Figs. 2, 5 J. 2, 3, 4, 8. 10] Sporidesmium siygium Berkeley & Curtis apud Berkeley in Gre- villea 3: 17. 1874. Perfect stage: Glonium clavisporum Seaver, (Lohman (7)). 1959] Moore — Genus Piricauda 105 Material examined: FH (Curtis coll. 3972) , from a maple log, Penn- sylvania, Michener 1243, Type, (slide RTM 1:28). Piricauda subcuticularis (McAlp.) comb. nov. Fig. 38 [15 -19, 20,29, 92) Sporidesmium subcuticulare McAlpine, Fungus diseases of stone- fruit trees in Australia, and their treatment. Agric. Dept. Vic- toria, p. 116. 1902. Material examined: Dept. Agric. Victoria, on defunct twigs of apricot, Armadale, Victoria, Australia. Type, (slide RTM 1:107). Piricauda suffulta (Pound & Clem.) comb. nov. Fig. 1 [hum 1 455] Sporidesmium suffultum Pound & Clements, Bot. Surv. Univ. of Nebraska, p. 6. 1896. Material examined: NEB, on decorticated cottonwood, Memphis, Neb. Type, (slide RTM 1:254). Piricauda trichophila (H. Syd.) comb. nov. Fig. 12 II. 2, 12, 13, 17] Sporidesmium trichophilum H. Sydow, Ann. Myc. 23: 498. 1925. Material examined: CUP, FH (Bartholomew coll), Sydow, Fung. exotici exs. 716, on leaves of Anona cherimolia, La Caja near San José, Costa Rica, Lectosyntypes, (slides RTM I: I, 1:205 resp.) Piricauda trigonella (Sacc.) comb. nov. Fig. 16 [I. 19, 20, 21, 22,9] Sporidesmium trigonellum Saccardo, Michelia 2: 641. 1882. Material examined: PAD (Saccardo coll), on defunct bark of Ailanthus, Libert 432, Type, (slide RTM 1:212). Piricauda tumulosa (Sacc.) comb. nov. Fig. 23 [I. 19, 20, 21, 22, 23, 24, 25] > Sporidesmium sculellare B. & Br. subsp. tumulosum Saccardo, Michelia 2: 289. 188l. Material examined: PAD (Saccardo coll), Type, on defunct wood of Fagus sylvatica, woods, Cansiglio, Italy, (slide RTM 1:213). Piricauda ulmicola (Sacc.) comb. nov. Figs. 39, 40 [1.419,25 34, 95, 209 Sporidesmium ulmicolum Saccardo, Syll. Fung. 4: 501. 1886. Saccardo states that this is supposed to be the “Cucurbitariae ulmicola stat. conid. Fkl. Symb. [Myc., p.] 172” that Fuckel refers to in his diagnosis. However, F. rh. 2170, which Fuckel cites as the type, has been examined (Herb. FH) and this particular 106 Rhodora [VOL. 61 exsiccatum shows neither an Ascomycete nor P. ulmicola, though at least two other Deuteromycetes are present. Though Saccardo's fungus seems to fit the “fungus conidiophorous" described by Fuckel, there remains considerable doubt if they are related. Material examined: PAD (Saccardo coll), Type, on dry elm twigs, Rhenogovia, Italy, (slide RTM 1:220). FH, F. rh. 2170, syntype of Cucurbitaria ulmicola, on dry branches of elm, Reichartshausen, (slide RTM L:227). Piricauda vernoniae (Dearn. & Barth.) comb. nov. Fig. 8 J. 2, 12, 13, 14, 15] Stigmella vernoniae Dearness & Bartholomew apud Dearness in Mycologia 21: 330. 1929. Material examined: DAOM (Dearness coll. 5384, (Barth. 8474) ), Type, on leaves of Vernonia gigantea, Williamsville, Missouri, (slide RTM 1:78). Piricauda viticola (Sacc.) comb. nov. Fig. 41 [l, 19, 20, 29, 30, 31] Sporidesmium viticolum Saccardo, Michelia 2: 289. 1881. Material examined: PAD (Saccardo coll.) , on defunct ?grape stems, Selva, Italy, (slide RTM 1:223). SPECIES INQUIRENDAE There are 47 species of Sporidesmium and one species of Stig- mella which have not been definitively examined, either in this present study or by other recent workers. These can be assigned to three categories: [7] those which are not to be found in the expected and known herbaria housing collections of the respective authors, or belong to collections that were destroyed by war; [7] those described by authors the location of whose collections is unknown; [!] species of the first two categories but which from the pre- sumptive evidence of the published figures most probably belong to other genera. SPORIDESMIUM -agapanthi Thuem. ?cavernarum Laub. ?alytospori. Richon -celastri Thuem. ?bulbophilum West. 'celatum Welw. & Curr. -carpineum Schulzer -cellulosum Fries -caulincola Fries -ciliatum Fries 3 Abbreviations of author names are as listed by Wright and Lois (12). 1959] Moore — Genus Piricauda -clavaeforme Preuss !clavatum Lév. -congestum Preuss ?cucumis Niessl -dolichopus Pass. ?effusum P. Henn -elegans Corda lepiphyllum Lév 'eremita Corda -fasciculare Preuss -fuscum Bon. -fusiforme Fries -griseum McAlp. -hyalopus Pat. -hydrangeae Thuem. -ignobile Karst. ?lambottet Roum. -melongenae Thuem. -microscopicum Bon. -phytolaccae Thuem. -populi Crouan -pulvinatum Fries -punctatum Lév. ?punctatum Woron. ?scleroticola P. Henn. !scorzonerae Aderh. !sparsum Fres. -sporotrichi Corda -sterculiae Tassi ?syntrichiae Racov. -tenellum Penz. & Sacc. !tripartitum Bagnis -triseptatum McAlp. !'vermiforme Riess: 107 STIGMELLA ?rubicola Bres. -linguaeforme Preuss ?lycii. Niessl ACKNOWLEDGMENTS The directors and personnel of the world’s herbaria are grate- fully acknowledged for their kindness and assistance in providing on loan the many collections cited herein, for without the avail- ability of such material Taxonomy would hardly be possible. In particular I would like to thank Drs. Albert Pilát (PR), D. P. Rogers (ex NY) and G. Taylor (K) for the loan of materials from the Corda collection, the Cooke and Ellis collections, and the Berkeley and Cooke collections respectively, and Dr. C. Cappel- letti (PAD) for the repeated loans from the Saccardo collection. A personal debt of gratitude is acknowledged to Prof. G. W. Martin for providing foundation and direction to my initial my- cological researches and to Dr. I. M. Lamb for making available the incomparable facilities of the Farlow Herbarium and Li- brary and for being mentor in these researches, to him also credit is gratefully given for the preparation of the Latin diagnoses. — DEPARTMENT OF BIOLOGY, HARVARD UNIVERSITY. 108 Rhodora [VOL. 61 LITERATURE CITED Il. Erris, M. B. 1958. Clasterosporium and some allied Dematiaceae — Phragmosporae. I. CMI Myc. Pap. 70: 1-89. 2. Hvcurs, S. J. 1951. Studies on micro-fungi. VI. Ceratosporium, Hirudinaria, and. Hippociepidium. CMI Myc. 39: 17. 3. —— 1953. Fungi from the Gold Coast. II. CMI Myc. Pap. 50: 66-70. 4, ————., 1958. Revisiones Hyphomycetum aliquot cum appen- dice de nominibus rejiciendis. Can. Jour. Bot. 36: 727-836. 5. JOHNSON, I. W. 1958. Some lignicolous marine fungi from the North Carolina coast. Elisha. Mitchell Sci. Soc. Jour. 74: 42-44. 6. LonwAN, Marion L. 1933. & cultural and taxonomic study of Hysterium hyalinum. Mich. Acad. Sci. Paps. 19: 133-140. 7. — 1937. Studies in the genus Glonium as represented in the Southeast. Torrey Bot. Club. Bull. 64: 57-73. 8. ————. 1939. Karsten’s type specimens of Hysteriaceae on conifers. Mycologia 31: 354-365. 9. Moonr. Rovarr T. 1958. Deuteromycetes I: The Sporidesmium complex. Mycologia 50: 681-692, 10. ————. 1959. Scheleobrachea Hughes. Mycologia [in the press]. II. Ray, D. L. AND D. E. Sruntz. 1958. Possible relation between marine fungi and Limnoria attack on submerged wood. Science 129: 93-94. 12. Wmiicur, J. E. anp R. J. Loss. 1949. Lista de siglas de autores empleadas en micologia y fitopatología. Lilloa 21: 225-269. 1959] Moore — Genus Piricauda 109 l A 2 4 Puate 1237. PIRICAUDA. Fig 1. P. suffulta, conidia and habit. Fig. 2. P. stygia, conidium. Figs. 3, 4. P. serendipita: 3. Discharged conidia bearing attached ultimate conidiophore cells. 4. Habit, note how the wall of the penultimate conidiophore cell is partially dissolved. Scale in micra. 110 Rhodora [VOL. 61 E. p 4 — 4 4 — " 4 4 PLATE 1238. Piricaupa. Fig. 5. P. irkara, conidia. Scale In micra. 1959] Moore — Genus Piricauda 111 Linn FE. Eu Puare 1239. Prricaupa. Fig. 8. P. vernoniae, conidia. Fig. 9. P. funerea, conidia; note the asperate nature of the cell wall. Fig. 10. P. globifera, conidium. Fig. 11. P. mitens, conidia ; note the integumentoid nature of the perimetric cells, (RTM 1:138). Scale in micra. T 3 ib 4 La LA 15 Prate 1240. PIRICAUDA AND STIGMELLA. Fig, 12. P. trichophila: Left, habit on leaf haii of Anona cherimolia; note how the lower hair cell has been attacked. Right, two conidia, lower on a leaf hair. Fig. 13. P. nitens, conidium, (RTM 1:138). Fig. 14. Stigmella mar- fagonis Oud., conidia typical of the genus Stemphylium. Note the prominent constriction at the primary horizontal septa, the approximately oval to sub-angular shape, and the continuous base, dimpled at the point of attachment of the protoplasmic thread, In the lowermest conidium the characteristically prominent basal sear is discernible. (Herb. GRO, Type. on Lilium Martagon leaves, slide RTM 1:94). Fig. 15. P. sacchari, conidia: note the extensive thiekening of the walls. Seale im muera. 1959; Moore — Genus Piricauda 113 PLATE 1241. PIRICAUDA. Fig. 16. P. trigonella, conidia; note the hyaline appendages. Figs. 17, 18. P. heteromera, conidia. Figs. 19, 20. P. chartarum: 19. Conidia, (RTM 1:105). 20. Habit, (RTM 1:11). Scale in micra. 114 Rhodora [ VOL. 61 22 A Ñ 5 „ — — C alii il t E 24 PLATE 1242. Piricaupa. Fig. 21. P. elliptica, conidia. Pig. 22. P. pulchella, conidia; note the light colored terminal cells. Fig. 23. P. tumulosa, conidia asperate. Fig. 24. P. aspera, conidia; note the conspicuous asperate condition, Seale in micra, 1959] Moore Genus Piricauda 115 26 PLATE 1243. Piricaupa. Fig. 25. P. damonis, conidia. Fig. 26. P. striata, conidia. Fig. 27. P. itochna, conidia. Scale in micra 116 Rhodora [VOL. 61 4 Prae 1244. Piricaupa. Fig. 28. P. apheles, conidia. Fig. 29. P. manilensis, habit; note the peltate hyphopodia, (RTM 1:103). Figs. 30, 31. P. bogoriens s, habit; note in Fig. 31 the Sporidesmium-like ?eonidium. Seale in micra, 1959] Moore — Genus Piricauda 117 Puate 1245. Piricaupa. Figs. 32, 33, 34. P. curvata, conidia; note, all are oriented base downwards. Fig. 35. P. fusus, conidium; note the remnants of the crystalline sheath, Fig. 36. P. melanopa, conidia with subtending cells, (second from the left RTM I:192b, rest RTM I:191a). Seale in micra. 118 Rhodor: [VOL. 61 C * t i é 39 40 PLATE 1246. Pirtcaupa. Fig. 37. P. putredinis, conidia, (RTM 1:256). Fig. 38. P. sub- cuticular.s, conidia. Figs. 39, 40. P. ulmicola: 39, left to right, conidial ontogeny. 40. Ma- ture conidia. (RTM 1:220). Scale in micra. 1959] Moore — Genus Piricauda 119 PLATE 1247. Prricaupa. Fig. 41. P. viticola, conidia. Fig. 42. P. exasperata, conidia; note the asperate condition. Fig. 43. P. nodosa, conidia, (RTM 1:89). Fig. 44. P. com- posita, conidia, (RTM I:13a). Scale in micra. 120 Rhodora u All hiss iste laa: a Piare 1248. Piricaupa. Figs. 45, 46. P. quadrata, conidia, (RTM 1:61, 1:234b respec- tively. Fig. 47. P. arcticoceanorum, conidia, (bottom and right margin). Scale in micra. e ~a " Numen 1959] Pike & Hodgdon — Mugo Pine in N. H. 121 NATURALIZED MUGO PINE IN NEW HAMPSHIRE! RADCLIFFE B. PIKE & A. R. HopGDoN The senior author has been aware for a number of years of a pasture in which the mugo pine (Pinus mugo Turra.) has be- come naturalized. So far as we are aware this is the first report of this species becoming adventive in North America. Sponta- neous naturalization of any non-native species is always of in- terest as a matter of record. It would seem in this case to be of special interest to foresters and farmers of Northern New Eng- land as a possible new weed tree of forest lands and pastures. This pasture is located in the town of Shelburne, Coos County, New Hampshire, approximately 44° 20’ N. It is just south of the Androscoggin River and U. S. Highway #2 a few miles from the Maine border. The property adjoins the grounds of the Shelburne public school and is directly across the highway from the former Aston estate, now the Shelburne Motel and Inn. It was on the grounds of this former estate that mugo pines were planted as ornamentals fifty years ago. A number of the original plants are still in existence and are now very large speci- mans yet still retain the typical mugo form of multiple trunks and bushy growth. Whether any of these pines were planted across the highway is not known but there are certainly none at the present time on the pasture site that approach the size or apparent age of those of the original planting. This is a sloping pasture with uneven drainage conditions and is 5-10 acres in extent. It is well occupied by mugo pines of various sizes and ages from seedlings to plants 12 ft. or more high and as much in breadth. All the plants observed had mul- tiple trunks and the typical mugo form. None was found with a single central upright main trunk (of the montana type). Seedling mugo pines were observed in the highway ditches and on graded banks on both sides of the highway as well as on a steep raw gravel bank next to the school grounds. ‘This ten- dency to colonize a variety of sites with different drainage bears published with the approval of the Director of the New Hampshire Agricultural Experi- ment Station as Scientific Contribution No. 236. 122 Rhodora [VOL. 61 out Shaw's? statement that "it grows indifferently in bogs and on the rocky slopes." Associated with the mugo pines were two native pines, Pinus strobus L. and Pinus resinosa Ait. as well as scattered poplars (Populus tremuloides Michx.) red maples (Acer rubrum L.) and gray birch (Betula populifolia Marsh), the usual invaders of abandoned pasture land in this area. Considerable variation. was displayed in the general outline and proportions of these shrub pines from flattish ovoid to up- right ovoid. The foliage also showed variation in depth of green and general density. Needle length on the collected specimens varies from 214 cms. to 614 cms. Cones showed the great varia- tion in size and shape that is characteristic of this polymorphic species. Shaw? emphasizes the variations in the cones by calling attention to the “elaborations that may be seen in Tharand Jahrbüch of 1861 and" — also “Hartigs’ Specifications of 60 forms of this species each dignified with a Latin name.” Cone specimens were collected which agreed with the three types illus- trated in Shaw?, p 57. Specimens of cones and foliage are de- posited in the University of New Hampshire herbarium. One other site of spontaneous colonization of mugo pine has been casually observed in the State of Maine. This is at the east end of the Waldo-Hancock bridge on Verona Island in the town of Verona in Hancock County. About 30 years ago mugo pines were used, as landscape material on the approaches to this bridge and are the obvious source of these spontaneous seedlings. However no detailed examination has been made or material collected. A recent report indicates another probable area in northern New Hampshire where this pine has become naturalized; how- ever we have not yet made personal observations to confirm this. These observations indicate that some caution should possibly be exercised in the use of the mugo pine for landscape purposes in areas where it can easily escape to wild or unused lands because under some conditions it is capable of spontancous nat- ? SHAW, G. R. The Genus Pinus. Publ. Arn. Arb. No. 5. 1914. 1959] Pike & Hodgdon — Mugo Pine in N. H. 123 uralization. Such caution is probably only necessary in Northern New England where the mugo pine is generally free of insect infestation that further south could be a deterrent to its success- ful spread. — DEPT. OF HORTICULTURE AND DEPT. OF BOTANY, UNIV. OF NEW HAMPSHIRE, DURHAM. DISCOVERY OF ARENARIA MARCESCENS IN THE UNITED STATES. — While botanizing in northern Vermont this past summer the writer collected a species of Arenaria which he at first thought to be A. groenlandica, known from alpine areas in the state. How- ever, upon subsequent examination the specimens were found to be distinctly different from A. groenlandica, closely resembling the description of A. marcescens given in the 8th edition of Gray's Manual. Additional material was later collected and sent to Dr. Marcel Raymond of the Montreal Botanical Garden who kindly confirmed the identification of the plant as Arenaria marcescens Fern. The station for the plant is at an altitude of about 3200 ft. on Haystack Mountain in the town of Lowell, Orleans County. Here a series of north-facing serpentine ledges is moistened by seepage from the top of the mountain and provides a habitat suitable for plants of arctic-alpine affinities. A. marcescens is found grow- ing abundantly in crevices and shelves of the ledges, often trailing down thus giving a hanging basket effect. Other species associated with it are Lycopodium Selago, Agrostis borealis, Scirpus caespi- tosus var. callosus, Empetrum nigrum, Vaccinium uliginosum var. alpinum, Houstonia caerulea, and Campanula rotundifolia. Of these, Agrostis borealis and Empetrum nigrum were previously reported in the state only from the summits of Mt. Mansfield and Camels Hump. The distribution of Arenaria marcescens as given by Fernald in Gray’s Manual of Botany, 8th edition, is western Newfoundland and Mt. Albert, Gaspé Co., Quebec. Its occurrence in northern Vermont represents a notable extension south of its previously known range as well as being an interesting addition to the flora of the United States. Specimens have been deposited in the writ- 124 Rhodora [VOL. 61 er’s own herbarium (P. W. Cook, 556, 584) and in the Pringle Herbarium at the University of Vermont (L. A. Charette, 2231). —PuiriP W. Cook, DEPARTMENT OF BOTANY, UNIVERSITY OF VERMONT. New Forms oF TRILLIUM RECURVATUM. — As in Trillium grandiflorum (Michx.) Salisb. and T. erectum L., variations in color and number of floral parts, with various modifications, occur in T. recurvatum Beck and other species of the genus. The recent examination of the herbarium of the Southeast Missouri State College at Cape Girardeau, Missouri, has revealed two additional variations in T. recurvatum which merit recogni- tion. Both of them are apparently not uncommon in Houck's Woods, an area located about three and one-half miles west- southwest of Cape Girardeau, Cape Girardeau County, Missouri. Trillium recurvatum Beck, forma petaloideum Steyermark, f. nov., a f. recurvatum sepalis submembranaceis petaloideis vinaceo- purpureis recedit. — Houck Woods, Bloomfield Road, Cape Gi- rardeau Co., Missouri, April 6, 1949, Wayne Adams, HOLOTYPE, in Herb. Southeast Missouri State College. In this form the three sepals have become petaloid in texture, shape, and color, their color being entirely maroon-colored or nearly so. They have the same thin texture as the petals. An extra sepal may be present, and, if so, is foliar in appearance. Trillium recurvatum Beck, forma foliosum Steyermark, f. nov., a f. recurvatum staminibus carpellis sepalis petalisque viridibus foliosis recedit. — Houck Woods, Bloomfield Road, Cape Girar- deau Co., Missouri, April 6, 1949, Wayne Adams, HOLOTYPE, in Herb. Southeast Missouri State College. In this form the sepals and petals, as well as the stamens and carpels, have become foliar in appearance, producing an entirely leafy, completely sterile plant. Mr. Adams notes on the label that both variations are “not rare." "They were found with normal populations of the species. — JULIAN A. STEYERMARK. 1959] Beaman — Floristic Studies 125 A MopERN STYLE IN Froristic Stupres.! — Recent generic monographs and revisions are frequently based on gross morphology and newer systematic techniques. From time to time floristic studies report chromo- some numbers or other bits of biological information, but the ordinary flora still gives little more than descriptive morphological treatments of species. This recent publication by Olov Hedberg is an outstanding example of the application in a floristic study of some newly developed taxonomic methods. Although the author calls his publication a taxo- nomic revision, it is floristic in treating the alpine plants of seven equa- torial East African mountains. Its format, however, is more like a collection of good generic monographs. For each species, subspecies, and variety treated there is synonymy, type data, taxonomy (often including chromosome numbers), ecology (including altitudinal records), and distribution (with extensive specimen citations) . The taxonomist making a floristic study often is faced with a frus- tratingly difficult task because he must cover large numbers of species which, along with their relatives, could be better understood if studied over their whole range rather than by geographic segments. Since the African alpine region is a relatively homogeneous physiographic unit where it is "summer every day and winter every night", it offers fewer such difficulties than would be encountered in floristic studies in areas delimited by political boundaries. Many of the afroalpine species are most closely related to geographically distant species that are not in- cluded in the treatment. The lack of an overall evolutionary relation- ship among the components of the flora prevents full comparison be- tween this investigation and a generic monograph. Afroalpine Vascular Plants is introduced by a short historical sketch, ecological notes, taxonomic concepts and methods, and notes on the nomenclature and material used. The preliminary ecological discussion is brief, but this aspect was treated in an earlier publication? and addi- tional notes are given under each taxon in the revision. Other related papers, including a pollen-analytical study, a cytological investigation of afroalpine grasses, and generic revisions, also preceded this major con- tribution. A chapter on vicarious taxa provides a review of the perti- nent literature, cites examples of vicariads in the flora, and discusses the significance of vicariads in the understanding of evolution, especially evolutionary rates. The largest portion of the work is in two sections of one chapter “The Flora". The first part (233 pages) of this chapter treats 44 fam- ilies, 125 genera, and 298 species (some with subspecies or varieties) ; 1 Afroalpine Vascular Plants: A taxonomic revision. Symbolae Botanicae Upsalienses 15 (1), 411 pp. 12 pl. 1957. 2 HEDBERG, OLOV. Vegetation belts of the East African mountains. Svensk Bot. Tidsk. 45: 140-208. 1951. 126 Rhodora VOL. 61] 42 new taxa, names, and combinations are included. Nine species hybrids are reported but not named. The second part (125 pages) is an appendix of taxonomic comments. Here the monotony of ordinary floristic presentation is relieved by nomenclature revisions, miscellaneous comments, and population analyses illustrated by tables, histograms, and scatter diagrams. Keys to families and genera are not provided. This omission restricts the floristic usefulness of the paper to persons with considerable taxonomic training. In future years the African mountains will be visited by investigators who could profitably use keys to fam- ilies and genera, and it seems that the additional cost of printing keys would have been justified by extending the utility of the flora. Afroalpine Vascular Plants is illustrated by 22 figures of line draw- ings, 30 figures of graphs, 15 tables, and 12 very clear half-tone plates of afroalpine plants. The line drawings adequately serve their purpose but are not artistic masterpieces. The tables, histograms, and scatter diagrams give detailed analyses of population variation. A map could have been helpful, although one is provided in the author's previous ecological paper.? Field studies on the seven mountains were made in 1948. Most pre— vious investigators had visited only one or two mountains. A minimum amount of time was spent by the author in field work, but it is evident that he made efficient use of his field opportunities. A majority of the herbarium studies were made when the author had access to Kew, the British Museum, and the Linnean Herbarium. Specimens in 15 other herbaria were also examined. ‘The study was supported by several grants and obviously required large investments in money and time. In investigating the afroalpine flora, Dr. Hedberg's ultimate goal is to develop a thorough phytogeographic understanding of these plants. He has now provided the taxonomic treatment upon which phytogeo- graphic analyses can be based. He also has produced an exemplary model for floras of the future. A great array of taxonomic projects lying dormant in the high mountains of America could profitably be studied in the style of Alfroalpine Vascular Plants. — JOHN H. BEAMAN, DEPART- MENT OF BOTANY AND PLANT PATHOLOGY, MICHIGAN STATE UNIVERSITY, EAST LANSING. Volume 61, No. 723, including pages 65-85, was issued 27 April, 1959 ierenewcr Liarary JUL 2 9 1959 Hodova JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor-in-Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR, IVAN MACKENZIE LAMB Vol. 61 May, 1959 No. 725 CONTENTS: Interesting Fungi from Massachusetts. zr aaaea a 127 Separothėca, A New Genus (Commelinaceae) from Mexico. U. T. Waterfall |... . rentur ontario 136 Natural Hybridization of Helianthus longifolius with H. atrorubens and H. occidentalis. Dale M. Smith and William C. Martin . . 140 Wallpaper Cleaner in the Herbarium. John Ebinger sse 147 Rhododendron maximum in New Hampshire. A. R. Hodgdon and R. Pike eee 148 Flora and Vegetation of the Alps. Askell Love (Review). .. . . .. iaia aE . . . a . 150 Potentilla anglica in New York. Joseph Monachino |... ene 153 Hymenopappus filifolius var. alpestris (Maguire) Shinners, comb. nov. Lloyd H. Spinners sss eem 155 The New Ingland Botanical Club, Ene. Botanical Museum, Oxford St., Cambridge 38, Mass. — RHODORA.—A monthly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid, in funds payable at par in United States currency in Boston; single copies (if available) 60 cents. Back vol. umes 1-58, with a few incomplete, can be supplied at $5.00 per volume. Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes, relating directly or indirectly to the plants of North America, will be considered by the editorial com- mittee for publication. Articles concerned with systematic botany and cytotaxonomy in their broader implications are equally accept- able. All manuscripts should be double-spaced throughout. Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance, will be furnished at cost. Address manuscripts and proofs to Reed C. Rollins, Gray Herbarium, 22 Divinity Avenue, Cambridge 38, Mass. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill, Botanical Museum, Oxford Street, Cambridge 38, Mass. Second Class Postage Paid at Boston, Mass. EATON PRESS, Inc. Specialists in Learned Publications 17 SPRING STREET WATERTOWN 72, MASS. Rhodora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 61 May, 1959 No. 725 INTERESTING FUNGI FROM MASSACHUSETTS ! HOWARD E. BIGELOW During the summer and fall of 1958, climatic conditions were especially favorable for the growth of fleshy fungi in western Massachusetts and a number of rare or unusual species appeared. Among these were several agarics about which there is little information in the literature since the original descriptions. Am- anita cinereoconia, Clitocybe fellea, C. socialis, Hygrophorus basidiosus and H. flavodiscus, are described here in order to complete the data omitted by the authors. Two gasteromycetes, Pisolithus tinctorum and Pseudocolus schellenbergiae, are of in- terest because of their occurrence in this region. The colors cited in quotation marks are from Ridgway (1912). HYMENOMYCETES Amanita cinereoconia Atkinson, Ann. Mycol. 7: 366. 1909. Plate 1249 Pileus 2.5-7 (II) cm. broad, convex to broadly convex at first, soon expanding to plane, at times with a broad obtuse umbo, margin appen- diculate with whitish, cottony fibrils from a partial veil, becoming appressed, surface dry, with numerous pyramidal warts (up to 2 mm. high) + in concentric rings, texture soft and rather cottony, points finally eroded and leaving only irregular floccose patches, appressed near margin, color “fuscous,” “hair brown," "smoke gray" on warts and patches, whitish to “pale smoke gray" between warts and patches; flesh thick on the disc, tinted pale fuscous near the pileus surface, whitish below, odor and taste not distinctive. 1 Acknowledgment is made to the Faculty Research Council, University of Massachusetts, for financial support of my 1958 field program, and for bearing the cost of photographic plates in this paper. 127 128 Rhodora [VOL. 61 Lamellae free but closely appressed to stipe, broad (up to 7 mm.), rounded at cap margin, narrowed at stipe, close, white, edges straight, fimbriate under a lens. Stipe 3.5-17 cm. long, apex and medial portions 0.6-2.5 cm. in diam- eter, base napiform, bulbous portion 1-3 cm. broad, becoming + equal in age, radicate portion 3-6 cm. long, central, solid (interior whitish) , pale "fuscous" to "fuscous" at bulb, gradually paler to the whitish apex, apex and medial portions fibrillose to somewhat scabrous; volval re- mains evident as "fuscous" warts in three to five + concentric rings on the bulb surface, finally appressed; annulus apical, thick, soft and cot- tony, white, usually adhering to stipe apex and cap margin, becoming appressed. Spores 8.5-10 x 4.5-6.5 u. usually elliptical (polymorphic at times and then oblong, pyriform, obovate, lacrymoid, 6.5-19 x 56.511) , smooth, amyloid; basidia 27.5-42(-60) x (6.5—)9—11 (-12) u. four-spored; cystidia not distinct from the numerous basidioles: pileus tissue: surface with numerous globose cells, 19-60 U in diameter, contents brownish in KOH, wall smooth, slightly thickened, originating as cystidioid end cells, clavate at first, but finally inflated to globose and becoming free, cuti- cular hyphae cylindrical, 2—4.5 u in diameter, hyaline in KOH, inter- woven, appearing slightly gelatinous in KOH, tramal hyphae cylindrical to inflated, 5-17 u in diameter, clamp connections present; gill trama divergent, hyphae cylindrical to inflated, 6-20 u in diameter. Solitary or gregarious, on soil in mixed woods or under conifers. Leverett, Amherst. August and September. Bigelow 7360, 7543, 7650, 7756. This distinctive Amanita is well-known in more southerly re- gions. It was originally described from Chapel Hill, North Caro- lina. More recently, Hesler (1937) has collected the species in Great Smoky Mountains National Park. In the field, the distinctive features of expanded specimens are: the gray warts and patches on the pileus and bulb of the stipe, the lack of membranous annulus or volva, the napiform and radicating stipe. Dr. A. H. Smith of the University of Michigan has kindly con- firmed the identity of my material. Clitocybe fellea Peck, Ann. Rep't N. Y. State Mus. 51: 284. 1898. Pileus 6-21 mm. broad, hemispherical at first, becoming convex, finally subplane and somewhat depressed, dry and somewhat shining, radiate- fibrillose to matted-fibrillose under a lens, finally minutely diffracted- scaly, a yellowish-tan to putty color (dingy "cream buff" to "chamois") ; flesh thin, whitish, odor farinaceous, taste bitter but soon fading. 129 Bigelow — Fungi from Massachusetts 1959] EJ 76 * ‘ D1u030042u19 ju y 6 ALVId 130 Rhodora [VOL. 61 Lamellae broadly adnate for some time, finally short decurrent, close to nearly subdistant, broad (3-5 mm.), not forked or intervenose, whit- ish, edges even and straight. Stipe 1.5-3.5 cm. long, 1-2 mm. thick at apex, equal or the base slightly enlarged, often curved and somewhat flexuous, solid, central, surface fibrillose-striate for some time, finally appressed, concolorous with the pileus. Spores 6-8(-9) x 4.5-5.5 u, broadly elliptical, smooth, hyaline in KOH, not amyloid, white in mass; basidia 25—38.5 x 6.5-8 LL four- spored; cystidia not differentiated; pileus tissue: cuticle yellowish in KOH, pigment present as very fine encrustations or in slightly thick- ened but smooth walls, hyphae cylindrical to somewhat inflated, 4.5-8 (-11) u in diameter, surface often with projecting end cells but not organized into a distinct turf, trama hyaline, hyphae cylindrical to slightly inflated, (2-) 4.5-8 u in diameter, clamp connections present; gill trama regular to subparallel, hyaline, hyphae cylindrical to slightly inflated, 3-13 u in diameter. Solitary, scattered to gregarious. On bare soil, Polytrichum, or in grass. In the open near conifers. Amherst, Sunderland, Leverett. June to September. Bigelow 6514, 6515, 6605, 6711, 7191, 7188, 7576, 7630, 7731, 7804. Peck's (1898a) colored plate of C. fellea is an accurate repre- sentation. My material is similar in all respects with the illustra- tion and description. "Through the courtesy of Mr. Stanley J. Smith, I have also had the opportunity to make an examination of the type collection from Gansevoort, N. Y. It is unusual that there are so few records of this species. The abundance of it throughout the collecting season of 1958 would certainly seem to indicate that this agaric is not rare. Possibly, the distribution is restricted to the northeast. Smith's (1944) re- port of C. fellea from Oregon is of another species, at present undetermined. Clitocybe socialis (Fr.) Gillet, Lest Hyménomycétes, p. 159. 1874. Pileus (0.5—) 1.5—4 cm. broad, convex to broadly convex at first with the margin incurved and slightly inrolled, edge white pubescent, ex- panding to plane with the margin remaining decurved for some time, in age becoming subrepand to elevated, rarely undulate, disc flat to somewhat umbonate or broadly depressed, surface dry and opaque, moistappearing in wet weather but not hygrophanous, dull, innately subsilky-pulverulent, finally glabrous, color dark red at first ("dragon's- blood") , fading with age to dull pinkish (dull "coral pink," dull "flesh 1959] Bigelow — Fungi from Massachusetts 131 pink," sordid "onion pink," "Japan rose"); flesh white, rather thin, odor and taste mild to rather disagreeable or subfarinaceous. Lamellae short decurrent at first, finally moderately decurrent, close, thin, narrow, not easily separable from the pileus trama, rather brittle, at times forked and anastomosing, not intervenose, color whitish ("car- tridge buff") at first, becoming buff ("cream color," "ivory yellow") in age, edges even and straight at first, becoming torn and undulate in age. Stipe 1.5-4 cm. long, 2.5-5 (-11) mm. broad at the apex, tapering downward from a slightly enlarged apex, bases often connate by a tomentose-mycelioid covering and strigosity which is intergrown with grass and humus, usually slightly curved, solid (white and spongy within) , central, surface with a faint bloom at first, becoming glabrous or innately fibrillose-striate, scurfy-lacerate in age, pale buff at base, concolorous with cap at apex. Spores 4.5-5.5 x 2.5-3 fu. elliptical, smooth, not amyloid, white in mass; basidia 13-26 X 4-6 TE four-spored; cystidia not differentiated; pileus tissue: surface of young specimens with a few cystidioid end cells, clavate-pedicellate, + 15 x 7 ju pigment evident in cell contents, be- coming appressed in expanded specimens, cuticular hyphae cylindrical, 2-3 (5) u in diameter, tramal hyphae cylindrical to inflated, 3-12 (-18) in diameter, clamp connections present, a few scattered laticiferous hyphae present; gill trama regular to subparallel, hyphae cylindrical to inflated, 4-16.5 u in diameter. Gregarious to subcespitose on lawn. August and September. Amherst. Bigelow 7591, 7760, 7229: The only other authentic specimens of C. socialis known previously in this country are those collected at Ann Arbor, Michigan by A. H. Smith and C. H. Kauffman. While a graduate student at the University of Michigan, I had the opportunity to examine this material. The Amherst collections cited above are identical in all respects except habitat with those from Michi- gan. Smith's and Kauffman's collections were made under pine or in a mixed stand of black locust and pine. Hygrophorus basidiosus (Peck) Peck, Bull. N. Y. State Mus. 116: 57. 1907. Pileus 1.54 cm. broad, convex to plane, subumbonate at times, gla- brous, hygrophanous, grayish-brown when moist, fading to pale gray (no comparable colors in Ridgway), radiate-streaked in fading; flesh whitish, odor and taste not distinct. Lamellae adnate to short decurrent, subdistant, broad, arched, thick, pale gray, edges even. 132 Rhodora [VOL. 61 Stipe 3.5-5 cm. long, 4-7 mm. thick at apex, tapering downward to a slender base, solid becoming hollow, surface glabrous, white. Spores 4-5.5 (-6) x 3-4.5 u, subglobose, smooth, hyaline in KOH, not amyloid; basidia 31-45 x 5-6.5 u, four-spored, sterigmata 5-9 long, curved; cystidia not differentiated; pileus tissue: surface not gel- atinous in KOH, cuticular hyphae cylindrical, 1-3 u in diameter, tramal hyphae mostly cylindrical, 2-6 u in diameter, clamp connections pres- ent; gill trama interwoven, hyphae cylindrical, 2-4.5 y in diameter. Gregarious in sphagnum in bog. Leverett. August. Bigelow 7503. Peck's (1887) orginal description of Clitocybe basidiosa states the lamellae are “whitish with a violaceous tint." When fresh, the lamellae of no. 7503 were pale gray without a violaceous tint. Peck further describes C. basidiosa as possessing a depressed pileus adding in his notes following that the pileus is "rarely slightly umbonate." The pilei of my collection were plane and subumbonate at times. As dried, my specimens are pale grayish not the very pale buff of Peck's type. In other respects, there is agreement. I found the spores of the type specimens to be slightly larger, 4.5-5.5(-6) X 4.5 „, than the measurements reported by Smith and Hesler (1942). Hygrophorus basidiosus is most closely related to H. albipes Peck. As described by Peck (1898b), H. albipes is a smaller agaric, with pileus 1.2 cm. broad, stipe 2.5-3.5 cm. x 3-5 mm. The gills are narrow, not broad as in H. basidiosus. Smith and Hesler found that the type specimens of H. albipes possessed a thin gelatinous pellicle. The surface of H. basidiosus is merely filamentous. Other features of the two species are nearly iden- tical. Hygrophorus flavodiscus Frost, in Peck, Ann. Rep't N. Y. State Mus. 35: 134. 1884. Plate 1250. Pileus 2-6 cm. broad, convex at first with an incurved margin, ex- panding to broadly convex, finally plane and shallowly depressed, mar- gin with white fibrils of partial veil, not striate, surface glutinous from the universal veil, gluten finally drying in radiating streaks, “orange buff" on the disc, paler toward margin and "pale orange buff" to "cream buff," edge whitish; flesh white, firm, odor and taste not distinctive. Lamellae adnate to short decurrent at first, becoming moderately or long decurrent, subdistant to distant, moderately broad (2-6 mm.), occasionally forked and anastomosing, usually intervenose and the sides 19" 9] Bigelow — Fungi from Massachusetts 133 Hygrophorus flavodiscus, X 1. PLATE 1250. 134 Rhodora [VOL. 61 of the lamellae venose, a distinct pinkish flush present at first. ("pale pinkish cinnamon") fading to whitish, edges even and straight. Stipe 3-7.5 cm. long, 6-14 mm. thick at the apex, equal or tapering downward, solid (whitish and firm inside) , eccentric at times, sheathed nearly to apex with glutinous universal veil and ending in an obscure annulus, whitish with pale orange-yellow tint, fibrillose to somewhat scabrous above the gluten, innately fibrillose beneath the gluten. Spores 6-8 x 3-4(-4.5) u. elliptical to elliptic-oblong, smooth, hya- line in KOH, not amyloid; basidia 32-52 x 6-8 u, four-spored; cystidia not differentiated; pileus tissue: pellicle gelatinous, thick, yellow in KOH, pigment intercellular and intracellular, hyphae cylindrical, 2-4 p in diameter, trama hyaline, hyphae cylindrical to inflated, 5.5-18.5 u in diameter, clamp connections present throughout; gill trama divergent, hyaline, hyphae cylindrical to slightly inflated, 2.5-8(-13) u in diam- eter. Gregarious to cespitose under white pine. Leverett. November. Big- elow 7957, 7958, 7974. The study of abundant material of this species both in the fresh and dried conditions confirms the observations of Smith and Hesler (1939) , which were based upon a study of the type. I have also had the opportunity to examine Peck's type at Albany. Smith and Hesler also indicate that H. flavodiscus should be compared with H. melizeus Fr. I have not had the opportunity to examine authentic material of H. melizeus. However, H. flav- odiscus does not become pallid ochraceous throughout as Fries (1836-1838) emphasizes for H. melizeus. Only the pileus of H. flavodiscus becomes ochraceous in drying. The lamellae fade from pinkish to white, and remain so in drying. The stipe is tinted with pale orangish-yellow when fresh. At times this color remains when the specimens are dried, but often the stipe fades to whit- ish like the lamellae. Kühner and Romagnesi (1953) describe H. melizeus from unpublished notes of J. Favre. They state that this agaric has a pileus which is pale alutaceous, beige or beige- chamois, recalling a small Hebeloma; gills and stipe tinted with the same colors; stipe dry; flesh becoming sordid brown in KOH. These characters do not apply to H. flavodiscus. Kühner and Romagnesi suggest that H. chrysaspis Métrod is close to H. flavo- discus. ‘The two species are perhaps related, but certainly not identical. H. flavodiscus does not become reddish-brown as the other. Habitat and the spore width do not agree either. 1959] Bigelow — Fungi from Massachusetts 135 H. flavodiscus seems most closely related to H. gliocyclus Fr. for both species show indications of a partial veil. I believe the two can be separated most easily on characters of the gills and spores. H. gliocyclus has yellowish, narrow gills and spores 7-9 (-11) x 5-6 u, as described by Smith and Hesler (1939). GASTEROMYCETES Pisolithus tinctorum (Pers) Coker & Couch, Gasteromycetes of the eastern United States and Canada, p. 170. 1928. Three fruiting bodies (Bigelow 7774) were found near one another on an old sawdust pile at Mt. Toby, the University res- ervation in Leverett and Sunderland. This unusual fungus is fairly common in the Pacific northwest and in the southeastern United States, but is apparently rare in the northeast. Specimens found on Cape Cod are deposited at the Farlow Herbarium, but there are no previous records of this fungus occurring in western Massachusetts. Pseudocolus schellenbergiae (Sumstine) Johnson, Bull. Ohio Biol. Survey 4: 338. 1929. On two occasions single fruiting bodies were discovered on the same site near the University campus. These were growing in mixed woods on wet soil and humus. Bigelow 6950 was collected July 19; Bigelow 7723 on September 11. The field characteristics and microscopic features are typical of the species. As far as I can determine, the Amherst collections represent the most northern occurrence of the species reported thus far. According to Snell and Dick (1956), Mrs. Sybil Curtis found P. schellenbergiae near Worcester. This fungus is well discussed in the literature, and has been reported previously from Pennsylvania, New York, New Jersey, and Rhode Island. Some investigators believe P. schellenbergiae to be identical with P. javanicus (Penzig) Lloyd. If this is true, the latter name has priority. — DEPT. OF BOTANY, UNIV. OF MASSA- CHUSETTS, AMHERST. LITERATURE CITED Fries, E. M. 1836-1838. Epicrisis Systematis Mycologici, seu Synop- sis Hymenomycetum. pp. 610. Upsaliae. 136 Rhodora [VOL. 61 HEsLER, L. R. 1937. Notes on southern Appalachian fungi, II. Jour. Tenn. Acad. Sci. 12: 239-254. KünwNER, R. and H. Rowacwrsr. 1953. Flore Analytique des Cham- pignons Supérieurs. pp. 557. Paris, France. Peck, C. H. 1887. Contributions to the botany of the State of New York. N. Y. State Mus. Bull. 1 (2) : 5-66. MEE . . 1898a. Report of the State Botanist. N. Y. State Mus. Rep't, 51:267-321. Pl. A, B. .......... 1898b. New species of fungi. Torrey Bot. Club Bull. 25: 321-328. l RipcwaAv, R, 1912. Color standards and color nomenclature. Wash- ington, D. C. SMT. A. H. 1944. Unusual North American Agarics. Amer. Midl. Nat. 32: 669—698. SmitH, A. H. and L. R. Hester. 1939. Studies in North American species of Hygrophorus, I. Lloydia 2: 1-62. T" 1942. Studies in North American species of Hygrophorus II. Lloydia 5: 1-94. SNEIU, W. H. and E. A. Dick. 1956. An unusual phalloid from Massa- chusetts. Mycologia 48:327. SEPAROTHECA, A NEW GENUS (COMMELINACEAE) FROM MEXICO U. 'T. WATERFALL While collecting in Mexico in August 1956, the author found in pine woods in the Sierra Madre southwest of El Salto, Durango, a dwarf member of the Commelinaceae only 3-7 cm. high, grow- ing from small, succulent, elongate, tuber-like roots. It has few flowers, with the inflorescence subtended and enfolded by oppo- site, or subopposite, ovate-falcate to ovate-attenuate foliaceous bracts which are longer than the few cauline leaves. Examination showed it to have separate sepals which are nearly hyaline, with a few long hairs at their apices and along the outside of the upper parts of their midribs. The corolla is sympetalous with a tube 6-8 mm. long, and lobes 5-8 mm. long, ovate to ovate-lanceolate. The 6 stamens are fertile, with fila- ments 1—3 mm. long, filiform to flattened, sometimes somewhat spiralled or twisted, inserted on the lower part of the corolla 1959] Waterfall — Separotheca from Mexico 137 lobes. The conuectives are 1-6 mm. long, flattened, wider than the filaments, sometimes longer. In Woodson's key to the genera of the Commelinaceae (1942) the combination of sympetalous corollas, leafy-bracted cymes and separate sepals could indicate only Setcreasea. How- ever, it could not be any of the species of that genus as treated by E. Matuda (1955). It seemed possible that it might be an anom- alous new species of that genus, but, in a family in which generic circumscriptions have been drawn as diversely as in the Commelinaceae, it was thought advisable to check related genera before describing a new species. In so doing, reference was found to Zebrina (?) pumila described by Greene (1888) from the Sierra Madre west of the city of Durango. This description fits fairly well the species under consideration. An examination of the type, A. Forrer, s. n., from the Sierra Madre west of Durango, altitude 8100 ft., Sept. and Oct. 1881 (Np), shows that it is the same taxon as the one under consideration. The genus Treleasea was erected by Rose (1899) to include three species, previously referred to Tradescantia, which have “petals tapering into a claw, forming a tube... ", epipetalous stamens and stipitate fruits (actually they seem to be pedicellate, not stipitate). He believed its affinities to be “perhaps nearest Zebrina ... Greene (1888) in describing Zebrina pumila stated, This curious little plant must surely be a congener of . . . Tradescantia leiandra, Torr., which Mr. C. B. Clarke (DC. Monogr. III:318) has referred, with a doubt, to the Central American genus, Zebrina.” Following the erection of Treleasea by Rose, Greene (1900) transferred Zebrina pumila to Treleasea, stating This plant, which I, twelve years since, was strongly inclined to make the type of a new genus, certainly falls into Mr. Rose's genus Treleasea newly established. It may even be identical with one or the other of the two species recognized by Mr. Rose. But my specimen is not now to be found, unless at the University of California." Rose later (1903) substituted the name Neotreleasea, stating that Treleasea had already been used by Spegazzini for another 138 Rhodora VOL. 61] genus. He included the same species as he had done previously. Then he appended a note concerning Treleasea pumila (Greene) Greene (Zebrina pumila Greene). He excluded it from his genus Neotreleasea, stating that he had seen the type and that “ . . . its relationships are more nearly with true Zebrina. The two have in common a terminal cluster of flowers subtended by a two- leaved spathe, a narrow tubular corolla, and widely separated anther-cells, etc." Thus we find Greene transferring his species to Rose's genus, and Rose subsequently excluding it. Woodson places Neotrelea- sea in Setcreasea while Matuda (1955) does not mention this species in his treatment of Setcreasea, but he does exclude it from Zebrina. From Matuda's treatment it would seem that he does not know the species through actual study, but only through reference to the literature. It seems probable that it has not been recollected until found by the author in 1956. No material has been found at the Gray Herbarium, the U. S. National Herbar- ium or the Chicago Natural History Museum according to their respective curators to whom the author is grateful for aid. This taxon has separate sepals as does Setcreasea, but has a long, strongly-united corolla tube and long connectives similar to Zebrina, and glabrous filaments, sometimes shorter than the connectives, which is characteristic of neither. Four possibilities seem to occur for the disposition of this species: (1) to place it in Setcreasea (sen. lat.) as a distinct species; (2) to place it in Zebrina as a distinct species; (3) to consider it a connecting link between the two genera and unite them; and (4) to create for it a new genus. The author has concluded that the latter alternative seems to be a logical one. Separotheca, gen. nov. Commelinacearum. Herbae erectae: caulibus annuis ex radicibus succulentis perennibus; foliis ovato-attenuatis vel linearibus; sepalis 3, distinctis, hyalinis vel subhyalinis; corollae tubo et lobis aequalibus vel subaequalibus; staminibus 6, ad basim corollae lobis affixis; filamentis glabris; connectivis quam filamentis latioribus, nunc brevioribus, nunc longioribus; fructibus siccis, stylis filiformibus, — stigmatibus capitatis; seminibus 5, angulatis. Separotheca pumila (Greene) Waterfall, comb. nov. based on Zebrina 1959] Waterfall — Separotheca from Mexico 139 (?) pumila Greene, Pittonia 1: 157-158. 1888: Treleasea pumila (Greene) Greene, Pittonia 4: 225. 1900. Stems short, 3-7 cm. tall, annual from elongate tuberous roots; stems and leaves glabrous; stem simple, or branching from the first or second node; first leaf 8-15 mm. long, ovate to ovate-lanceolate; second leaf 2-4 cm. long, ovate-caudate to linear-lanceolate; sometimes a longer third leaf present; bracteal leaves 1.5-5 cm. long, usually longer than the leaf below it, ovate-attenuate to ovate-caudate; flowers few; sepals distinct, subhyaline to hyaline, especially on their margins, pilose along midribs and at apices; corolla pink, 11-15 mm. long, basal half united into a tube, lobes ovate to ovate-lanceolate; stamens 6, fertile; filaments 1-3 mm. long, filiform to flattened, inserted on lower part of corolla lobes; connective 1-6 mm. long, flattened, wider than the filaments, inverted V-shaped, sometimes longer than the filaments; styles filiform; stigmas capitate; fruit a capsule; seeds 5, ca. 2 mm. long and 1.5 mm. wide, more or less angulate, dark brown. In addition to the type, the following collections have been seen, all from the Sierra Madre; Waterfall 12673, shallow soil on rock strata in pine forest, 5 miles west of El Salto, Durango, Aug. 11, 1956; Waterfall 12704, open pine woods, wet flats, 24 miles west of El Salto, Durango, Aug. 12, 1956. — DEPARTMENT OF BOTANY AND PLANT PATHOLOGY AND THE RESEARCH FOUNDATION, OKLAHOMA STATE UNIVERSITY, STILLWATER, OKLA- HOMA. LITERATURE CITED GREENE, E. L. 1888. New Species from Mexico. Pittonia 1: 157-158. 1900. Corrections in Nomenclature. Pittonia 4: 225. Matupa, Eizi. 1955. Las Commelináceas Mexicanas. An. Inst. Biol. Univ. Mex. 26 (2) : 303-432. Rose, J. N. 1899. Treleasea, a New Genus of Commelinaceae. Contr. U. S. Natl. Herb. 5 (4): 207—208. 1903. Neotreleasea. Contr. U. S. Natl. Herb. 8: 5-8. WoopsoN, ROBERT E. JR. 1942. Commentary on the North Ameri- can Genera of the Commelinaceae. Ann. Mo. Bot. Gard. 29: 141- 154. 1 10 Rhodora (VOL. 61 NATURAL HYBRIDIZATION OF HELIANTHUS LONGIFOLIUS WITH H. ATRORUBENS AND H. OCCIDEN TALIS! Dare M. SMITH AND WILLIAM C. MARTIN Several studies of natural hybridization between species of Helianthus have been made in recent years. Many natural hy- brids have been reported and most of them have been substanti- ated by controlled crossing. In general, whenever two diploid species grow in close proximity under disturbed conditions, local hybrid swarms result. The gene flow may be restricted as in the case of Helianthus divaricatus X H. microcephalus (Smith and Guard, 1958), or extensive introgression may occur as in H. annus X H. bolanderi (Heiser, 1949). Regardless of the extent of hybridization, all of the previously described examples have dealt with pairs of species. In September of 1957, a mixed population consisting of four species of Helianthus was found one mile northeast of Albert- ville, Alabama. These species were H. microcephalus, H. longi- folius, H. occidentalis, and H. atrorubens. Putative hybrids of H. longifolius X H. atrorubens, and H. longifolius X H. occi- dentalis were discovered, but H. microcephalus apparently had not hybridized with the other species. The area in which the mixed population was found was a relatively new highway em- bankment bordering a small scrub-oak woodlot. The four species were more or less aggregated at specific places within the area, and hybrids were found only at the zones where two species overlapped. At the west end of the population there were a few scattered individuals of H. occidentalis, and approximately 200 yards distant at the east end of the arca was a large population of H. atrorubens with scattered individuals of H. longifolius be- tween these first two species. Helianthus microcephalus was found at the edge of the oak woodlot adjacent to the plants of H. atrorubens and H. longifolius. While mixed populations of H. atrorubens and H. microcephalus are quite commonly en- 1 Supported by the University of Kentucky Faculty Research Fund and NSF grant #G-1836. The authors express their appreciation to Dr. C. B. Heiser, Department of Botany, Indiana University, for valuable suggestions following a critical appraisal of the manuscript. 19:9j Smith & Martin—Helianthus longifolius 141 countered in many places, the other two species are rarely ever seen in the southeast much less being found in mixed popu- lations. Helianthus occidentalis is a common plant of the prairies and sand hills of the midwest from Michigan to Texas, and is occasionally encountered in the southeast. Helianthus longi- folius is one of the rarest species of Helianthus, and its range covers only a small area in the northern parts of Alabama and Georgia, where it typically grows on or near rock outcroppings. Helianthus atrorubens and H. microcephalus are widely distrib- uted throughout the southeast in a variety of habitats. The rea- sons for this unusual occurrence of four species in such a limited area are somewhat obscure at the moment, but it seems possible that the plants could have been introduced into the area along with fill material which was used in constructing the embank- ment, since the plants are perennials and their underground parts could. be transported along with soil very easily. PROCEDURES Herbarium specimens and living plants, as well as flower buds for cytological examination were taken from the population. Selected individuals and a random mass collection were pressed for later study. Whole plants were pressed, since the basal ma- terial as well as the upper parts of the plants possessed import- ant morphological features for comparison. Herbarium speci- mens have been deposited at Indiana University and the Univer- sity of Kentucky. The cytological material was fixed in a mix- ture of 3 parts ethyl alcohol to I part acetic acid for 24 hours, and transferred to 70% ethyl alcohol for storage. Anthers of the appropriate age were squashed in acetocarmine for study. MORPHOLOGY 'The species occurring in this population are quite distinct and are not easily confused, although H. atrorubens, H. occidentalis, and H. longifolius all have the upper cauline leaves reduced, giving the plants a decidedly scapose aspect. The leaves of H. microcephalus are essentially similar along the length of the stem except in the inflorescence where they are reduced. ‘The leaves and stems of H. longifolius are completely glabrous, so 142 Rhodora VOL. 61] that it too is quite distinct. Actually, only Helianthus occiden- talis and H. atrorubens present a somewhat similar aspect, but the purple disk corollas and obtuse phyllaries of H. atrorubens serve to distinguish it quite adequately. The outstanding mor- phological features of the four species are presented in Table l. The morphology of the putative hybrids will be considered individually. HELIANTHUS LONGIFOLIUS X H. OCCIDENTALIS One putative hybrid of this combination was discovered and appeared to be essentially intermediate between the two parents. Especially evident were the modifications of leaf shape and pubescence, the hybrid having narrowly lanceolate leaves which were setaceous on the margin. HELIANTHUS LONGIFOLIUS X H. ATRORUBENS The many usually well defined differences between these two species are, in this population, bridged by hybrid individuals. Considerable variation was encountered in leaf width, phyllary shape, head diameter, disk color and leaf pubescence, which were analyzed in this study, and are summarized in Table 2. The lower stature and more prominent basal rosettes of H. longifolius are usually sufficient to separate it from H. atrorubens but were not used here because of the great environmental modification of these two characteristics in this population. LEAF SHAPE. The basal leaves in H. longifolius are generally numerous, elongate and linear or narrowly obovate, rarely exceed- ing one centimeter in width. In H. atrorubens, the leaves show considerable variability in shape, but are always many times broader than one centimeter. In this population the hybrids bridge the gap between these two extremes. LEAF PUBESCENCE. Trichomes are completely lacking in H. long- ifolius except for a few at the tips of the receptacular bracts. In contrast to this, H. atrorubens is usually rather harshly pubescent over most of the plant. The leaves of the putative hybrids are characterized by the presence of a few scattered trichomes. PHYLLARY SHAPE. One of the most outstanding features of H. atrorubens is the presence of broad, obtuse phyllaries. Those of Å e 14 Smith & Martin—Helianthus longifolius 1959] MI s2qo'] “Wd O I-90 3]EurUln5E£ IO ayenua ye AÁ[MOLIEN "uD 09-9 yyeauaq oso[njuaui0] daO SNOIGLIS yuNstp o[onad ‘dy enue unn 31v[o22ue[-3]EA() wia}s suoye panqysig snolqv[4o snjpud22042114 * FY snjpud2320421u ‘H pup 'snijofi2uo] ^H *sijpjuap1220 Koll o “Wd 9 I-60 3jeuruinov 40 dienusne Á[MOIIEN "up 9 [-6:0 SnoIlqe[o aseq 0o) Sueden Ayyenpeis *o1e[oooue[ vou 0) 1e2aurT [eseq ApoS1v'T sno1qev[45 snijoji2uo] ^H MO][[oÁ soqo' "uD 9 -8'0 dienusne Á[MOIIEN "up 0 -9˙8 28081118 0) sno1qveog o[onod pagum Ap ysis 0) pa*o1ieu ‘ILAQ Jeseq Áqog1e'q 98081118 $1]D3u2p1220 N əjdınd SO "uD € 6I on sdy proiq *310u$ "ub qg9-c6 dINSITY 0] SNOIGRIC o[onad pasuim A]pvoig 03 pamoireu ‘97 BAO ATPROIG [eseq pose] o1nsirq-Surpeaxdg SUIQNAOL]D `H 10100 v0) IPULI As Ad VIIS AVTIAHd WPIM 32uo25s23qngq ↄdrus zuouloguraay ava! JONHOSHHOdJ WALS ‘H ‘suaqnio.jo sni1upi]2H jo sainjeay [e2orSo[ouduojq jo uosuedwop I 218v] 144 Rhodora [VOL. 61 H. longifolius are linear attenuate while the hybrids’ are merely acute. DISK DIAMETER. The differences here are not pronounced but, in general, the disks of H. longifolius are about 1.0 centimeter in diameter whereas those of H. atrorubens range up to 1.5 centi- meters. DISK COLOR. A clear cut difference exists here in that H. longi- folius has disk corollas which are yellow throughout, while H. atrorubens has purple-tipped corollas. In the field the putative hybrids appear to have yellow corollas but an examination under the microscope reveals that the hybrids have purple pigmentation on their tips. The pictorialized scatter diagram in figure | points out the discordant pattern of variation encountered in the population. a _ 4 * DIAMETER OF DISK IN CM, 09? - 4 WIDTH OF LARGEST LEAF IN CM. FIGURE 1. Pictorialized scatter diagram of a population sample including Helianthus longifolius, H. atrorubens, and putative hybrids. The smallest triangles = attenuate phyllaries, largest triangles obtuse phyllaries, intermediate triangles = acute phyllaries. Open triangles = yellow disk corollas, solid triangles = purple-tipped disk corollas. Triangles without a diagonal line on right side = glabrous leaves, with a long line = scabrous or hirsute leaves, with a short line = remotely setaceous leaves. 1959] Smith & Martin—Helianthus longifolius 145 CYTOLOGY The chromosome numbers of the species, with the exception of H. longifolius, which is reported here for the first time as N=17, have been presented by Heiser and Smith (1955), and are diploid (N=17). Meiosis in the parental species has been reported as normal in all except H. atrorubens, in which Jackson and Guard (1957) have reported numerous abnormalities. Plants of H. atro- rubens which have been grown in the experimental garden have occasionally failed to produce pollen or viable seed, which may be an expression of an abnormal chromosomal condition. This is further suggested by the occurrence of low pollen fertilities in plants from the Alabama population which appear to be "good" H. atrorubens. | : | TABLE Il Comparison of Morphological Features of Helianthus atrorubens, H. longifolius, and their Putative Hybrid H. atrorubens Intermediates H. longifolius LEAF Pubescence Scabrous or Remotely Glabrous hirsute setaceous Width 3.5-5.5 cm. 1.6-3.4 cm. 0.9-1.5 cm. PHYLLARY SHAPE Short, broad, Short, broad, Narrowly attenuate tips obtuse tips acute or acuminate DISK Diameter 1.2-1.5 cm. 1.1-1.3 cm. 0.9-1.3 cm. Corolla color Lobes purple Tips of lobes Lobes yellow purple However, in all the plants from this population which have been examined cytologically, parental species and putative hybrids, meiosis was essentially normal. Buds of several putative hybrids were examined, and of 65 microsporocytes examined, 63 showed perfect pairing and only 2 showed abnormal associations in the form of a single chain of 4 chromosomes. DISCUSSION This report of natural hybridization in Helianthus is especially interesting in that four potentially hybridizing species are repre- sented in a mixed population. Even though the evidence suggests that only three of the four species have actually hybridized, the 146 Rhodora VOL. 61] stage is nevertheless set for the formation of bizarre types of indi- viduals incorporating genetic material from all four species. That such a phenomenon is theoretically possible, has been veri- fied by three and four species hybrids which have been produced experimentally (Smith, unpublished). The actual discovery of hybrids of this type in nature would have special significance relative to the problems concerning the origin of polyploid spe- cies of Helianthus. There are several polyploids which show no close resemblance to extant diploid species or to the F, diploid interspecific hybrids which have thus far been produced. Conse- quently, it is suggested here that certain polyploid species may have originated from hybrids involving more than two species. The relatively high fertility and the high degree of chromosome homology of the hybrids of H. atrorubens X H. longifolius serves to point up the lack of great barriers to crossing between the dip- loid perennial species of Helianthus, even in the presence of a high degree of morphological distinction. Furthermore, the very narrow distribution of H. longifolius contrasts sharply with the extensive range of H. atrorubens, and ecological differences prob- ably also exist, so that the two species are not found in mixed populations unless they are brought together by disturbing in- fluences. It would appear that genetic barriers have been of minor importance in speciation in this group; therefore, with a complete breakdown of the external barriers separating these species, it is conceivable that an amalgamation of them could Occur. As yet, there is little evidence that this is happening even in such a disturbed area as that described here. SUMMARY A mixed population of diploid (N=17) perennial sunflowers composed of Helianthus atrorubens, H. occidentalis, H. longi- folius and H. microcephalus was found in September, 1957, near Albertville, Alabama. One putative hybrid between H. occiden- talis and H. longifolius was found, and no hybrids involving H. microcephalus were evident but a large hybrid swarm of H. atrorubens X H. longifolius was present. These plants were quite variable, with leaf and phyllary characteristics showing [VOL. 61 Smith & Martin—Helianthus longifolius 147 outstanding variation. Very little meiotic abnormality was en- countered in the population, although the occurrence of low pollen fertilities in some of the plants suggests that cryptic struc- tural differences in the chromosomes exist. Speciation has appar- ently proceeded without the development of the sterility barrier, but the external barriers which ordinarily separate them have in this instance been partially broken down by the influence of man. However, even under these conditions there is little indication of the amalgamation of the species in this population. — UNIVERSITY OF KENTUCKY, LEXINGTON, AND UNIVERSITY OF NEW MEXICO, ALBUQUERQUE. LITERATURE CITED Heiser, C. B. Jr. 1949. Study in the Evolution of the Sunflower Species Helianthus annuus and H. Bolanderi. Univ. Calif. Pub. Bot. 23: 157—208. Heiser, C. B. JR. AND D. M. Situ. 1955. New Chromosome Num- bers in Helianthus and Related Genera (Compositae). Proc. Indiana Acad. Sci. 64: 250—253. Jackson, R. C. anp A. T. Guard. 1957. Analysis of some Natural and Artificial Hybrids in Helianthus. Proc. Indiana Acad. Sci., 66: 306—317. Smita, D. M. AND A. T. Guard. 1958. Hybridization between Helian- thus divaricatus and H. microcephalus. Brittonia, 10: 137-145. WALLPAPER CLEANER IN THE HERBARIUM. — In a well-kept herbarium, specimen sheets are handled in such a manner that they do not become excessively dirty. However, where collections have been neglected for long periods or are housed in poor cases they may become badly soiled by dust and smoke. ‘This was the case with a rather large number of sheets in the Herbarium of Yale University, most of which had come as gifts and were dirty when received. Specimens of little value may be discarded but valuable collections should, of course, be preserved even though soiled. If the plant is not glued too tightly it may sometimes be removed and mounted on a clean sheet. However, where this is impossible without severe breakage some other method must be sought. 148 Rhodora [VOL. 61 Since most of the soiled areas are usually near the margin of the sheet, it might be feasible to remove it with an art gum eraser. Actually, this tends to smudge and leaves the paper streaked rather than clean. The idea of using wallpaper cleaner of the pliable putty type occurred to me, and this was tried with very good results. This material can be purchased from most paint or hardware stores and comes in vacuum sealed tins. Several different brands have been used and all proved to be equally satisfactory. When the can is opened, the unused cleaner may be stored in screw-top glass jars to prevent it from drying out. For cleaning a specimen, we use a small portion (about the size of a golf ball) and this is also stored in a screw-top jar. In use, the cleaner must be kneaded thoroughly to keep it pliable and to work the dirt in. It gives satisfactory results until the ball is quite black, when it is dis- carded for a fresh piece. If there is dust close to the specimen, the cleaner may be rolled into a small cylinder and used like a pencil. Care must be exercised, however, since hairs, flower parts, etc., will adhere to the cleaner and thus be removed. This method has been used in the Yale Herbarium for the past six months with excellent results. The idea is offered here in the hope that it may be found equally useful by others who have similar problems. — JOHN EBINGER, OSBORN BOTANICAL LABORA- TORY, YALE UNIVERSITY. ^ RHODODENDRON MAXIMUM IN NEW Hampsuire.! — The valu- able paper by the late C. H. Knowlton (5) on Rhododendron maximum in New England unfortunately contains some errors and omissions for New Hampshire. Recently Iltis (4) has quite understandably accepted. Knowlton's data and has included his stations on a map covering the range of the species. In the past few ycars we have tried to visit all recorded New Hampshire colonies of R. maximum and thus have become aware of the faults in recent publications. But before reporting on these, we have wanted to check all possible clues to new stations. ‘Published with the approval of the Director of the New Hampshire Agriculture Experi- ment Station as Scientific Contribution No. 232. 1959j Hodgdon & Pike—Rhododendron maximum 149 Our first comment relates to the colony, chiefly in the town- ship of Pittsfield, which was first reported by R. J. Eaton (2). It was assumed both by Eaton, and until recently, by the senior author of this paper, that a small part of the stand was in Straf- ford County in the township of Strafford, the remainder or larger part being in Pittsfield. Knowlton, loc. cit. counted this colony as two separate stations presumably because it was reported from two townships and Iltis has perpetuated the error by showing two dots on his map in the vicinity of Pittsfield, New Hampshire. The senior author has visited this area at least a half-dozen times in the years since June 29, 1931 when he was first shown the plants by Robert Varney of Barrington, New Hampshire. These visits, including our most recent one on July 7, 1954, have shown no significant expansion or contraction of the colony in 23 years. In 1954 we made a very careful comparison of map-details with the local topography. It is evident that the plants are not in Strafford County at all and that the nearest plant is somewhat more than 100 feet from the line as shown on the Alton Quad- rangle. But it is interesting to note that the colony does extend westward along the hemlock-shaded shore of Adams Pond into the township of Barnstead in Belknap County. Near Adams Pond 3 townships and 3 counties converge and this single small colony of Rhododendron, covering possibly half an acre, almost stretches into all of them. Recent authors have lost sight of the Manchester Rhododen- drons though at certain times during the past 60 years they may have been more numerous there than in any other part of New Hampshire. W. E. Moore (6) reported the presence of a tre- mendous area of Rhododendrons about 2 miles northwest of Amoskeag Falls. Then on December 11, in the same year (1897) F. W. Batchelder wrote an article for the "Manchester Union" entitled "A Day In My Arboretum" in which he commented on the occurrence in Manchester of Rhododendron, White Cedar and other interesting species of woody plants. Two years later Batchelder (1), this time, for a more critical audience, discussed the local Rhododendrons. Thus, there is good evidence, in the literature, that our plant occurred in Manchester in some abun- 150 Rhodora VOL. 61] dance at the turn of the century. More recently the late Reverend Hubert Sheehan, OSB of St. Anselm’s College collected material from Black Brook Cedar Bog in Manchester and before this in 1935, Dr. Maurice Provost of Vero Beach, Florida, then a student at St. Anselm’s College, discovered another stand from which he collected specimens. ‘The authors of this paper singly or together have now visited both of these stations and in addition have found Rhododendron in one other locality within the confines of Manchester. At least 2 of these 3 colonies probably have been separate for a long time. The other 2 occur in distant parts of the same swamp and quite possibly were joined 60 years ago. With one exception (the Albany station on Mt. Chocorua) we have succeeded in relocating all of the New Hampshire stations listed by Knowlton loc. cit. The evidence is reasonably good in this instance that there was a colony on Mt. Chocorua (3) but from the drastic changes that we have noted taking place in other stands as a result of lumbering operations, swamp-flooding, brows- ing by deer, etc., it is quite likely that the Rhododendrons there may have been completely destroyed or reduced to a few incon- spicuous individuals. Specimens from all stations visited by the authors have been collected and are to be found either in the Herbarium of the University of New Hampshire or that of the New England Botan- ical Club. — A. R. HopGpon AND R. PIKE, DEPARTMENT OF BOT- ANY AND DEPARTMENT OF HORTICULTURE, RESPECTIVELY, UNIVERSITY OF NEW HAMPSHIRE, DURHAM. CLAUDE FAVARGER. Flore ct vegetat on des Alps. I. Etage alpin. II. Etage subalpin. 271 & 274 pp., with 32 + 32 planches and 35 + 41 drawings by Paul A. Robert. Delachaux & Niestle S.A., Neuchatel & Paris 1957 & 1958. Price Swiss Fr. 30.00. FLORA AND VEGETATION OF THE Als. — The Alps are among the regions most botanists and all those interested in the vegetation of mountains and northlands want to visit and study, although only a few of those living in distant countries ever get an opportunity to climb the lofty peaks and enjoy the multitude of flowers. The majority has to be content with descriptions by others and they also must study the 1 CLAUDE Favarcer. Flore et vegetation des Alps. I. Etage alpin. II. Etage subalpin. 271 & 274 pp., with 32 + 32 planches and 35 + 41 drawings by Paul A. Robert. Delachaux & Niestlé S. A., Neuchâtel & Paris 1957 & 1958. Price Swiss Fr. 30.00. 1959] Love—Flora and Vegetation of the Alps 151 flora by aid of manuals, pictures, and herbarium material only, adding life just through imagination. This is sometimes a little difficult, espe- cially because of the fact that descriptive books by botanists tend to be as dry as are their herbarium plants. Botanists with the ability to write literary works, leaving their interest in details for a more general and descriptive treatise, are as rare in the Alps as they are elsewhere on the globe. Recently, one of the best botanists of Switzerland, a man with a wide reputation in modern and classical approaches, has given others of his time and knowledge by writing two delightful volumes on the flora and vegetation of the Alps. This reviewer knows few books which unite botanical correctness with literary language to such a high degree. The many nice pictures, in black and white and color, make it easy for the reader to follow the author, Professor Claude Favarger of the University of Neuchátel, into the field and study with him the variable flora of these best known and most attractive mountains of Europe. Although both the volumes are intended for the interested layman rather than for the specialist, there is hardly another treatment available which at the same time gives a more scientific approach to the subject. In the first volume, the alpine milieu is described in detail and the terminology explained, at the same time as the reader is informed about the climate and different kinds of soil to be expected. The typical characteristics and biology of alpine plants in general are masterly reviewed before the Alpine flora and the origin of its elements are discussed in another chapter. Thereafter, the fundamentals of the phytosociological approach to studies on vegetation are given, followed by a description of the particular communities of Alpine plants con- nected with different kinds of high-alpine conditions, from the rocks to the meadows and mires. The descriptions of these associations are based on their quantitative and qualitative floristic composition, and numerous colored pictures and pen-drawings in the text explain the main features of each community. The last part of the first volume reviews the principal families of Alpine plants, with representative examples. In the second volume, which is devoted to the subalpine regions, the characteristics and limits of this less easily defined zone are discussed, and the transition zones are considered in a concise but very clear chapter. The subalpine zone in the strict meaning of the term is then described in the same detail as were the communities of the alpine zone in the first volume, with a strong emphasis on climax associations and on some pioneering and specialized communities typical of the somewhat more favorable regions. A short chapter on the Jura Moun- tains and the western Alps completes the descriptions of the vegetation, while the floristic review covers many of the most interesting subalpine 152 Rhodora VOL. 61] species typical of these parts of the Alps. The volume is completed with a chapter on botanizing in the Alps, in which a plea is made for caution lest rare species be eradicated, and on the protection of the nature of the Alps so that future generations also will be able to enjoy the same beauty observed by present visitors. There is a short bibli- ography and good indices to both volumes at the end of the second volume. Although the descriptions in these volumes are based on the flora and vegetation of the Alps, they are of great interest also to those who are concerned with the plants of American mountains, notably those in New England. Naturally, the species are rarely the same, and the com- munities are also very different and never as colorful here as in the Alps, but the general character of the vegetation is rather similar and its history may also be somewhat comparable in these formerly periglacial mountain complexes. The delightful volumes by Professor Favarger are a great tribute to the many botanists of the region he treats past and present. The lit- erary abilities of the author and his deep knowledge of the flora and vegetation of his Alps greatly enhance the value of the books. The artist, P. X. Robert, is also worthy of praise, and so is the printer and publisher. The books are to be recommended not only to those inter- ested in mountains in general and the Alps in particular, but also to botanists and botanically minded tourists visiting the Alps. careful study of the volumes before such a visit will greatly increase its value and also open one's eyes to the many features in botany which nowhere are more distinct and better studied than in the Alps in Switzerland. — ÁSKELL LOvE, INSTITUT BOTANIQUE DE L'UNIVERSITÉ DE MONTRÉAL. POTENTILLA ANGLICA IN New YORK. — For over eight years I have observed in the Bronx and in Brooklyn, New York City, a creeping Potentilla that answers to the description of P. anglica Laicharding (P. procumbens Sibth.). In November, 1956 I ob- served, but did not collect, the same species in a yard at Great Neck, Nassau County, Long Island. My collections, deposited in the Herbarium of The New York Botanical Garden, are Mona- chino s.n. (6-22-50), formerly the grounds of The New York Botanical Garden, near the Allerton Avenue entrance, June 22, 1950; s.m. (10-28-53), (N. .), south of the Conservatory, on a lawn near Juniper plantings, at least two major patches, Oct. 28, 1959] Monachino—Pokentilla anglica in N. Y. 153 1953; 622, same station, June 30, 1958; also two fragments col- lected as vouchers and placed in packets are from the Brooklyn Botanic Garden, N. Y. C., on a lawn, June 30, 1956, and June 7, 1958. The species has flowers chiefly 4-petaled and stem leaves predominantly 3-foliolate. It forms a conspicuous, low, dense colony; the repent stems branch and they often root at the nodes; the basal leaves are 5-foliolate, the stem leaves most 3-foliolate. The petioles in the plants examined are up to 3.5 cm. long; the corolla is 10-15 mm. across, the petals are a little longer than the sepals, obovate, slightly emarginate at the apex, bright yellow on the upper side, orange-yellow at the base; the carpels are about 20-25 in number. Many achenes in no. 622 have become enlarged and plump, but are not fully matured. Our plant is the small- flowered form named f. parviflora Domin in Wolf's monograph (1908). Clapham et al. (1952) describe the diameter of the flower of P. anglica as (10—) 14-18 mm. and the number of carpels as 20-50; they call the plant trailing tormentil. The 4-merous flower (4 petals, 4 sepals, 4 epicalyx segments) easily identifies the species in the standard floras. Rydberg (1908) has it in the small group Tormentillae, characterized by the flowers solitary, axillary, on long pedicels, leaflets not tomentose beneath. P. cana- densis is a neighboring species and from a distance P. anglica appears similar, until one approaches and notices that the bright yellow corolla is 4-merous. The colony is conspicuous enough, but “merely a yellow flower on the lawn” does not invite investi- gation, and it may well pass as any low cinquefoil or even a lady- sorrel or a buttercup. In Canada the species is known from Labrador, Newfound- land, Cape Breton Island and Nova Scotia, according to Fernald (1950). For the United States it has been reported from only two states. Rydberg (1908) cited it as introduced in California. Dix (1949) was the first to report it from the East. He collected it near Lake Shehawken, Wayne Co., Pa.; he wrote that its iden- tity was confirmed by Fernald, that it was abundant and grew on a somewhat wet rocky slope with the usual pasture grasses and adventives, that it blossomed late in the fall and that the flowers were 10-13 mm. broad. He added that another locality 154 Rhodora VOL. 61] about a quarter of a mile south was discovered in Sept. 24, 1949. The species was again mentioned for Lake Shehawken, as seen on the Torrey Botanical Club trip to the area, by Louis E. Hand in Bull. Torrey Club (1950) 77: 408. Gleason (1952) does not cite the eastern Pennsylvania station. He persists in using the binomial P. procumbens Sibthorp (pub- lished in 1794) instead of the earlier P. anglica Laicharding (pub- lished in 1790); but I have not looked into the matter of nomen- clature. Until recently when Laicharding's earlier, but obscure, binomial was revived, the name P. procumbens was in uniform use everywhere. Incidentally, in his key Gleason distinguishes the species by its 4-merous flowers from P. reptans L. with flowers “all 5-merous" but the illustration of the latter in the same work shows clearly four petals. Also, for one described as having “leaf- lets 5 or 7” the figure of P. reptans too much suggests some tri- foliolate leaves. I could not trace the actual specimens from which the illustrations of P. reptans and P. anglica were drawn for the New Illustrated Flora, but if they were drawn from the proper species, the widely spreading hairs as shown on stems and pedicels of both must have originated in the artist’s unchallenged imagination. As unquestioned as the identity of the species appears from description, I could not check my specimen of P. anglica with authenticated European material, and such a comparison would be desirable. It has already been hinted how the trailing tor- mentil may be overlooked as a common cinquefoil or some other low-growing yellow-flowered plant. I would predict that inter- ested persons thus alerted will discover P. anglica elsewhere, particularly on lawns, and will eventually prove it to be fairly widespread in its eastern range. — JOSEPH MONACHINO, NEW YORK BOTANICAL GARDEN. LITERATURE CITED CLAPHAM, A. R., Turin, T. G., AND WARBURG, E. F. 1952. Flora of the British Isles, 485, 494. Dix, W. L. 1949. Potentilla procumbens in the United States. Rho- dora 51: 390—391. FERNALD, M. L. 1950. Gray's Manual of Botany, Eighth Edition, 806, 812. 1959] Monachino—Pokentilla anglica in N. Y. 155 GLEASON, H. A. 1952. ‘The New Britton and Brown Illustrated Flora 2: 292, 294. RYDBERG, P. A. 1908. Rosaceae. N. Am. Fl. 22: 294, 302. Worr, T. 1908. Monographie der Gattung Potentilla. Bibliotheca Bot. 71: 642, 650—654. HYMENOPAPPUS FILIFOLIUS var. ALPESTRIS (MAGUIRE) SHINNERS, comb. nov.—H. nudipes var. alpestris Maguire, Amer. Midl. Nat. 37: 144. 1947. H. filifolius var. nudipes (Maguire) Turner, RHODORA 58: 219. 1956. (Illegitimate combination, based on Hymenopappus nudipes.) Dr. Maguire published H. nudipes and its var. alpestris simultaneously. If these are considered to be a single taxon and treated as a variety, it must bear the earliest available epithet in that rank, which is alpestris. Dr. Turner insists (personal communications) that since the Code regulation regarding automatic epithets for varieties including the type of the species is retroactive, an automatic var. nudipes was also pub- lished simultaneously with var. alpestris. His new combination is then to be regarded as based on the automatic but hitherto unlisted var. nudipes. This is hardly in keeping with the spirit of the regulation which states that automatic epithets have no author. It is definitely contrary to the intention of Dr. Gleason's original proposal (for different wordings see Brittonia 7: 17, 1949), whose explicit statement that automatic epithets could not be transferred was unaccoutnably omitted from both the Stock- holm and Paris versions of the Code. Even if Dr. Turner's con- tention is accepted, his combination remains illegitimate because he designated Hymenopappus nudipes and not var. nudipes as basinym. — LLoyp H. SHINNERS, SOUTHERN METHODIST UNIVERSITY, DALLAS, TEXAS. Volume 61, No. 724, including pages 87-126 was issued 30 June, 1959 | Qerenemcs LIBRARY Dodota JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor-in-Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR. IVAN MACKENZIE LAMB Vol. 61 June, 1959 No. 726 CONTENTS A Revision of the Houstonia purpurea Group (Rubiaceae) Edward E. Terrell 157 Paucedanum palustre, an Interesting Additicn to the Flora o: Essex County, Massachusetts. Stuart K. Harris 181 | | | | | | | | The Nru England Botanical Club, Inc. Botanical Museum, Oxford St.. Cambridge 38, Mass. RHODORA.—A monthly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid, in funds payable at par in United States currency in Boston; single copies (if available) 60 cents. Back vol- umes 1-58, with a few incomplete, can be supplied at $5.00 per volume. Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes, relating directly or indirectly to the plants of North America, will be considered by the editorial com: mittee for publication. Articles concerned with systematic botanv and cytotaxonomy in their broader implications are equally accept, able. All manuscripts should be double-spaced throughout. Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance. will he furnished at cost. Address manuscripts and proofs to Reed C. Rollins, Gray Herbarium, 22 Divinity Avenue, Cambridge 38. Mass. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill, Botanical Museum, Oxford Street, Cambridge 38, Mass. Second Class Postage Paid at Boston, Mass. EATON PRESS, Inc. Specialists in Learned Publications 17 SPRING STREET WATERTOWN 72, MASS. Rhodora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 61 June, 1959 No. 726 A REVISION OF THE HOUSTONIA PURPUREA GROUP (RUBIACEAE)! EDWARD E. TERRELL The Houstonia purpurea group or species-complex includes seven named species: H. purpurea L., H. canadensis Willd. ex R. & S., H. tenuifolia Nutt., H. longifolia Gaertn., H. lanceolata (Poir.) Britt., H. montana (Chickering) Small, and H. setiscaphia Carr. These form a unit of closely related species distinct from all others in the genus. Fosberg (1954) placed H. floridana Standl. in this group (as a variety of H. purpurea), but its stout, woody tap root and “stalked” anthers indicate a much closer relation- ship to the H. nigricans group. H. tenuifolia is occasionally confused with H. nigricans (Lam.) Fern. Although both have narrow cauline leaves, they may be distinguished easily by other characters. The former species has many small roots from a short, erect or horizontal, rootstock (a rhizome) and sessile anthers due to adnate filaments; the latter has a stout, woody tap root and anthers stalked by the freedom of the filaments from the corolla-tube. Fosberg stated in 1937 that the genera Houstonia L., Hedyotis L., and Oldenlandia L. should be merged under Hedyotis. He reaffirmed this view in later papers (1941a, 1941b, 1943, 1954). Shinners (1949) accepted these conclusions and transferred Texas ‘Parts of this study were aided by a Grant-in-Aid from the National Science Foundation administered through the Highlands Biological Station during the summer of 1957, also by the tenure in 1952-53 of a Mary S. Muellhaupt Postdoctoral Scholarship in Botany at the Ohio State University. Appreciation is expressed to Dr. R. T. Clausen (Cornell University) for guidance in early parts of this study, to the curators of herbaria from which loans were granted, and to my wife, Bessie Z. Terrell. Dr. G. M. Schulze (Botanical Museum, Berlin) supplied photographs of the type of Houstonia canadensis. Dr. H. L. Blomquist and Dr. R. L. Wilbur (Duke University) made certain suggestions regarding the manuscript. 157 158 Rhodora [VOL. 61 species of Houstonia to Hedyotis. In contrast, Fernald (1950) and Gleason (1952) maintained Houstonia as a separate genus. In the present paper it is retained as a separate genus, a course I prefer to follow until further evidence is obtained on the many, diverse species comprising the large genus Hedyotis (sens. lat.). I have not studied these generic relationships in detail, but my concern has been with relationships within the group that forms the subject of the present paper. MORPHOLOGY AND CyTOLOGY Morphological evidence about the Houstonia purpurea group has come from herbarium and field studies and the growing of plants in pots and in gardens. Field work was most intensive in the central and southern Appalachians and in southern Ohio. All taxa have been collected and observed in the field over parts or most of their respective ranges. Certain collections have been transplanted and grown for short periods. One outdoor trans- plant of H. purpurea has been observed at intervals over a period of several years. Although it has not been possible to transplant all taxa to the same uniform garden, enough transplants have been made to indicate that morphological differences are pri- marily genetic rather than primarily environmental. That this is true is borne out by field observations: where two species grew together in nature they maintained their differences. Many morphological characteristics were investigated to de- termine how the species differ. About a dozen characteristics turned out to be diagnostic, and these are more or less important, depending on the taxa involved. The species differ slightly in certain floral characters, but these are often inconstant and dif- fer within a species from one population to another. In cases in which more constant floral differences exist these are no better than or are inferior in taxonomic usefulness to vegetative char- acters. Over 4000 specimens were examined from the following herbaria (abbreviations from Lanjouw and Stafleu, 1956) : University of Cincin- nati (CINC); Wiegand Herbarium of Cornell University (cv); Duke University (DUKE) ; Chicago Natural History Museum (r); Florida Agri- cultural Experimental Station (As); Florida State University (rsu); 1959] Terrell — The Houstonia purpurea Group 159 University of Georgia (cA); Gray Herbarium of Harvard University (GH) ; State University of Iowa (1A); Indiana University (1ND) ; Univer- sity of Michigan (micH); Missouri Botanical Garden (mo); North Carolina State College (Ncsc); University of North Carolina (NcU); Tulane University (No); New York Botanical Garden (Ny); Oberlin College (oc); Ohio State University (os); Philadelphia Academy of Natural Sciences (pH) ; University of Tennessee (TENN) ; U. S. National Herbarium (us); University of Wisconsin (wis) ; West Virginia Uni- versity (wvA). Other herbaria examined include that of the Great Smoky Mountains National Park and those of Dr. E. Lucy Braun, Mr. C. C. Deam, and Mr. E. J. Palmer. Cytological data on this genus are largely lacking. For H caeru- lea (not in the H. purpurea group) Stevens (1912) reported a gametic number of 16. Fagerlind (1937) corrected this number to 18, and believed this species to be tetraploid since the basic number in the genus is 9. Chromosome counts made by the writer in 1948 from immature anthers yielded the following data: COLLECTION SPECIES NUMBER LOCALITY CHROM. No. H. purpurea var. purpurea 1974 Black Mountain, g=ca.9 Harlan Co., Ky. s=ca. 18 S. J. Smith ‘Tuscaloosa Co., 5157 Alabama s—ca. 18 H. purpurea var. calycosa 2057 Clinton Co., Ohio g—ca. 8-12 Var. calycosa (typical) 1926 cedar glades, s—ca. 18 Wilson Co., Tenn. H. canadensis 2058 Chautauqua Co., N. Y. s—ca. 18 Intergradant, H. tenuifolia 1990 Morgan Co., Ky. g=ca.9 — H. longifolia var. compacta These data suggested that at least three of the taxa in this com- plex are diploid. Sizes of guard cells are sometimes correlated with polyploidy. It was thought that determination of these cell sizes in Houstonia might provide an indication of polyploidy. In 1958 measure- ments of stomatal guard cells were made on dried specimens by the method described by Celarier and Mehra (1958). One to four 160 Rhodora [VOL. 61 different collections of the taxa accepted here were tested. Cell sizes were about the same in all taxa except the following: H tenuifolia varied from about the same as other taxa to about 30% larger; H. purpurea var. montana was slightly smaller. The gen- eral constancy of cell size was impressive. These results merely suggest further that polyploidy may not be involved in this complex: obviously, no more definite conclusions can be reached. ‘TAXONOMY OF THE GROUP Lhe most recent revision of this genus was that of Standley (1918), in which the six then known species of the H. purpurea group were considered to be separate species. Robinson and Fer- nald (1908) and Fernald (1950) maintained those in the Gray's Manual range as distinct, although Fernald previously (4940) had questioned whether they should not all be considered as vari- ations of one species, after secing intergradation in the field in southeastern. Virginia. Gleason (1952) upheld their separation, except in the case of H. lanceolata. Asa Gray in his Manuals had merged them as varieties of one species. Fosberg (1941a) advoca- ted three species, but later (1954) combined them as Gray had done. ‘The present writer (unpublished M. S. Thesis, Cornell University, 1949) tentatively recognized three species. I hat two extreme viewpoints — cach a separate species, or all one species — have been held in regard to this complex, is indica- tive of the nature of the problem it presents. If only typical speci- mens are examined the extremes are so different that many tax- onomists would recognize three or four, or possibly even seven, species. On the other hand, if several hundred specimens of all kinds from throughout the respective ranges are examined, boun- daries between taxa become indistinct and their circumscription difficult. In reference to this group in Indiana, C. C. Deam wrote (in correspondence) that generally the collector with a few speci- mens is able to identify them, but with a larger number, “then it is that you learn that the set are a bunch of renegades and do not conform to any rule or law.” Apparently with the tacit assumption that all of nature is basic- ally orderly, the writer several years ago began studying this problem. After initial attempts at statistical and graphical meth- 1959] Terrell — The Houstonia purpurea Group 161 ods to tame the renegades, the procedures employed evolved into what has been termed the morphological-geographical method. When the group is studied state by state, region by region, local area by local area, and the spatial continuity in morphology is followed, it can be seen that some order obtains and part if not all is not chaotic. Field study has been an important adjunct to this method, for differences and resemblances show up more clearly in the field than in the herbarium. 'The present revision is of the basic taxa and does not include details on putative hybrids except where knowledge of these bears directly on taxonomy. A subsequent paper is planned to deal with such plants in more detail and to consider what evi- dence they provide concerning the origin and recent evolution of this group. Any taxonomic treatment of a species complex must consider intergrading specimens and putative hybrids. ‘They are what make this a complex; if all species were distinct and non-intergrading, this study would have been completed earlier, the taxonomy being obvious. The intergradant populations are now well enough understood so that future findings about them are not likely to change the taxonomic treatment as presented here. Present evidence indicates that much of the variation is the result of hybridization. However, two species seem to cross in some localities but not in others, according to observations in the field. Where they do not cross there may be intermediate types of habitats available for intermediate kinds of plants; therefore, there seem to be genetic barriers, or at least non-ecological bar- riers, present which prevent crossing. All members of this group are heterostylous, a condition which in some other plant groups has been said to favor outcrossing. In this treatment four species are recognized. Two of these have three varieties each and the other two are merely extremely variable. All species vary irregularly or clinally over their ranges and must be studied throughout their ranges before valid conclu- sions are forthcoming. Each species intergrades with the others, but not all varieties intergrade. The term, intergrade, is used here as either a noun or a verb; 162 Rhodora [VOL. 61 when used as a noun it is preferred over the noun intermediate, to indicate not intermediacy but complete intergradation (usual- ly in several characters) between two taxa. In order to determine how many intergrading specimens there actually are, a check was made of the writer's records of collec- tions in two large eastern herbaria, the U. S. National Herbar- ium and the Gray Herbarium. From a total of 676 sheets, 78 sheets or 11.5% represent intergrades (to various degrees) be- tween H. purpurea and H. longifolia var. compacta. Such col- lections are far more numerous than all other kinds of inter- grades combined, which are only 3 or 4% of the total; in view of their obvious importance they are discussed in detail below. These percentages are only approximate, as these two herbaria do not contain by themselves a really complete or accurate sta- tistical representation. Comparable figures from herbaria in North Carolina would show more than 11.5% of the kind of in- tergrade mentioned, and herbaria from certain northern states would show a lower figure. Using the available figures, one ar- rives at a total figure of 14-15% for intergrades of all kinds. This, however, is misleading, since some collections assigned to defi- nite taxa are atypical, or have minor signs of presumed intro- gression, or are problematical in some other way. The more significant kinds of intergrades are as follows (cer- tain other kinds not listed here are cited in the systematic treat- ment): l. H. purpurea var. calycosa — H. canadensis. Scattered and local in southern Ohio, central and western Kentucky, and southernmost In- diana. 2. H. purpurea — H. tenuifolia. Putative hybrids are rather frequent in Missouri and Arkansas, but rarer in the Appalachians. Distribution on Map 4. 3. H. canadensis — H. longifolia var. longifolia. Intermediates occur in northern Illinois (see Map 3 and citations under H. canadensis) . 4. H. tenuifolia — H. longifolia var. compacta. Occur in eastern West Virginia, eastern Kentucky, and western Virginia (cited under former species). 5. H. purpurea — H. longifolia var. compacta. These populations, constituting at least 11.597, of the total 1959] Terrell — The Houstonia purpurea Group 163 complex, range from Ohio south into Georgia and Alabama (see Map 4). They are especially numerous in the southeast, where they recur again and again over a continuous land area in South Carolina and Georgia. This regional race may deserve some sort of formal taxonomic recognition, but it seems best to defer such action until more exact data are available on this morphologic- ally heterogeneous group. Plants of these populations vary through the entire gamut of variation from intermediates into both supposed parental taxa. Leaf shapes vary from ovate, like purpurea, to narrow-elliptic, like compacta. Lower internodes of stems may have long, scattered hairs of the purpurea type, or fine, dense, grayish puberulence of the compacta type, or both types of vestiture may be intermixed on one plant. Mass collections from South Carolina and Georgia often include plants with a variety of leaf shapes and stem vestitures. Plants from the sand- hills of the Carolinas and Georgia tend to be more like var. compacta and grade into it imperceptibly, while plants from the piedmont tend to be more like H. purpurea and likewise grade into it. All of these kinds of populations usually occur in dis- turbed or secondary habitats. In populations from locations north of the Carolinas it is sometimes very difficult to judge whether intermediate types of plants may not actually be appar- ent hybrids between H. purpurea and H. tenuifolia, but it is un- likely that these would constitute more than 2097 of all collec- tions placed in the present grouping. (Intergrades in this pres- ent grouping have been annotated generally by the writer in the past as between H. purpurea and H. tenuifolia, because it was earlier believed that H. longifolia var. compacta should be con- sidered a subspecies of H. tenuifolia.). The various types of intergrades listed above and those cited elsewhere are not the only kinds of problematical populations in this complex. To assign a certain collection to one taxon is sometimes arbitrary; in some of these cases the plants appear to be introgressants — more variable due to the influence of genes from another taxon. Certain other populations suggest hybrid derivation among three basic taxa. The solution of all such prob- lems posed by this complex would seem to depend ultimately on 164 Rhodora [VOL. 61 the use of detailed graphical and statistical methods along with transplant experiments and cytological studies. It may be asked whether the H. purpurea group should not, as has been done by Gray and by Fosberg, be treated as one ex- tremely variable species with several subspecies or varieties. To the present writer the following reasons are to be included among those favoring treatment of the extremes as separate spe- cies: l. ‘The four species in their extremes differ from each other by several well-marked morphological characters. 2. Many populations of the basic taxa occupy large ex- tents of land over which they retain their essential distinctness in spite of a certain amount of intergradation. 3. Two species seem to cross in some geographic areas but not in others; therefore, isolating mechanisms exist between them under certain circumstances. 4. Certain taxa occur sympatrically. In this case they usually occupy somewhat different habitats but may occasionally grow together. The situation in the H. purpurea group is comparable to that in the Gilia tenuiflora-latiflora group, as described by Grant (1957). The latter aggregate is a syngameon, a term which ac- cording to Grant (p. 67) "is the sum total of species or semispe- cies linked by frequent or occasional hybridization in nature; a hybridizing group of species; the most inclusive interbreeding population." In the case of the Houstonia purpurea complex we do not know definitely that hybridization is the cause of the var- iation. However, the general similarity of its patterns of variation to those in the Gilia group is striking. Grant stated (p. 63) that "treatment of Gilia tenuiflora, G. latiflora, and G. cana as a single species would obscure rather than accurately portray their rela- tionships”. The present writer feels that the same is true of the Houstonia purpurea group. SYSTEMATIC TREATMENT The following description includes the more fundamental characteristics of the Houstonia purpurea group as well as the less fundamental ones which all four species have in common. 1959] Terrell — The Houstonia purpurea Group 165 'The taxonomic status of this group in the genus as a whole is in doubt until other species have been studied. Flowers heterostylous. Corollas funnelform. Anthers sessile or with filaments free from corolla-tube as much as 1 mm.; anthers linear to oblong, 0.7-1.6 mm. long in thrum? flowers, 0.3-1.2 mm. long in pin flowers. Styles 1.3-2.2 mm. long in thrum flowers, 3-8 mm. long in pin flowers. Stigmas linear to oblong, 0.2-1.0 mm. long in thrum flowers, 0.5-1.30 mm. long in pin flowers. Capsules subglobose, slightly com- pressed, (A-) 1⁄4 () inferior truncate or rounded or emarginate at apex. Seeds few — 35 per capsule, subcrateriform, oblong or roundish or angular, black, reticulate-pitted; hilum centered in concavity and 14—34 as long as entire seed. KEY TO THE HOUSTONIA PURPUREA GROUP3 A. Median cauline leaves ovate or lanceolate, widest toward the base or at the middle: if ovate then 6-34 mm. wide and 1-3.2 times longer than wide; if lanceolate (4-10 mm. wide and L/W ratio 3.3-6) then with calyx-lobes 4-7 mm. long. B. Calyx-lobes 1-3.9 mm. long; median cauline leaves broad-ovate to ovate-lanceolate, 1-3.2 times longer than wide. C. Lowest internodes sparsely to densely pubescent (with hairs usually at least 0.4 mm. long) or hirsutulous or glabrate; median inter- nodes 2-9 cm. long; median cauline leaves 10-60 mm. long, 6-34 mm. wide; corollas light purple or white. la. H. purpurea var. purpurea. CC. Lowest internodes glabrous or stems only slightly pubescent on lower nodes; median internodes 0.5-4 cm. long; median cauline leaves 8-30 mm. long, 6-13 mm. wide; corollas deep purple; plants from summits of Roan Mt., N.C.-Tenn., and Grandfather Mt., N.C. Ib. H. purpurea var. montana. BB. Calyx-lobes 4-7 mm. long; median cauline leaves varying from narrow- lanceolate to broad-ovate, 1-6 times longer than wide. lc. H. purpurea var. calycosa. AA. Median cauline leaves elliptic or obovate or linear, widest at the middle or near the apex or about equally wide for most of their lengths, or if narrow-lanceolate then calyx-lobes less than 4 mm. long; these leaves 1-6 (-8.5) mm. wide. D. Basal leaves definitely ciliolate and pubescent above (glabrate in plants occurring locally in southcentral Ohio), present during flower- ing and fruiting; stems with 3-6 internodes; lower and/or middle internodes usually much longer than upper ones; median cauline leaves oblonceolate, obovate, or elliptic, 2.5-6 times longer than wide; calyx-lobes 1.6-3.5 (-4.5) mm. long . . . . . 2. H. canadensis. ?In heterostylous flowers, thrum refers to exserted anthers and included style and stigma; pin flowers have exserted style and stigma and included anthers. 8Excluding intergrading specimens. 166 Rhodora [VOL. 61 DD. Not as above in all particulars. E. Lowest internodes densely and finely cinereous-puberulent, with all hairs uniformly less than 0.4 mm. long, or densely papillose- puberulent, or if glabrous then median cauline leaves 9 or more times longer than wide. F. Stems with 3-9 (-11) internodes; median internodes 3-8 cm. long; median cauline leaves linear to narrow-elliptic, 7-20 or more times longer than wide; fertile branches usually with only 1-3 nodes and elongate (to 20 or even 28 cm. long); in- florescence very diffuse and open; calyx-lobes usually less than 2 mm. long and not or scarcely exceeding capsules; mature capsules 1.5-2.5 mm. long m 3. H. tenuifolia. FF. Stems with 6-13 internodes; median internodes 1-4.5 cm. long; median cauline leaves narrow-elliptic, narrow-lanceo- late, narrow-oblong, to linear, 4-11 times longer than wide; fertile branches usually with 2-several nodes and to 12 cm. long; inflorescence not diffuse and open; calyx-lobes 0.5-3.1 mm, long, always exceeding mature capsules; mature cap- sules 1.8-3.0 mm. long. 4b. H. longifolia var. compacta. EE. Lowest internodes not as above. G. Stems with 4-7 internodes, glabrous to somewhat pubescent or scabrous; New England, Great Lakes region, and Canada. 4a. H. longifolia var. longifolia. GG. Stems with 6-13 internodes, glabrous or short-pubescent only at nodes; southwestern North Carolina and adjacent Georgia and South Carolina . . . . 4c. H. longifolia var. glabra. NOTE: H. longifolia var. compacta and var. glabra in most of the herbaria examined in the present study have been annotated by the writer as H. tenuifolia (see discussion of rela- tionships under the systematic treatment of the last species). l. Houstonia purpurea L. Sp. Pl. 1:105. 1753. Perennial herbs. Rhizomes simple or branched, horizontal or erect, sometimes subligneous, shortened or longer and more slender, to 5 cm. long, bearing many small roots. Stems erect or ascending, one-many, tetragonal and slightly winged, 0.4-4.5 dm. high, sparsely to densely pubescent to villous-pubescent below, especially at nodes, with multi- cellular hairs to 1.5 mm. long, or stem glabrous. Internodes numbering 4—12; median internodes 0.5-9.5 cm. long. Stipules lanceolate to deltoid, entire, irregularly toothed, or erose, often with dark glands at apices of teeth, obtuse, acute, or acuminate, to 5 mm. long, to 5.5 mm. wide. Basal leaves forming a rosette in winter, usually withered at time of flowering, rarely present, varying from oval to spatulate, tapering into petiole shorter than to rarely longer than blade, to 3.8 cm. long, to 1.3 cm. wide, glabrous below, pubescent or glabrous above, sometimes ciliolate. Cauline leaves 3-7 nerved; lower leaves sessile or short-petio- late, oblanceolate to ovate; median leaves sessile, broad-ovate and sub- cordate to narrow-lanceolate, 0.8—6.3 cm. long, 0.4-3.4 cm. wide, 1-6 times longer than wide, glabrous or pubescent on nerves below, pubes- 1959] Terrell — The Houstonia purpurea Group 167 cent to glabrous above, ciliolate to glabrous on margins; upper leaves similar to median leaves, reduced in size and becoming bractlike in inflorescence. Branches ascending to spreading, to 19 cm. long, usually arising from uppermost 1-4 nodes, forming an open to rather compact, few to many flowered inflorescence. Pedicels slender, less than 8 mm. long. Calyces glabrous to sparsely pubescent or hirtellous; calyx-lobes erect or spreading, linear-lanceolate to ovate, 1.0-7.0 long, 0.3-3.0 mm. wide, 14 as long to slightly exceeding corolla-tubes, equalling to con- siderably exceeding mature capsules. Corollas purplish to white or deep purple, 4-12 mm. long, granular to densely pubescent within; corolla-tubes 2.0-6.8 mm. long, 1.5-4.0 mm. wide distally; corolla-lobes 1.5-5.0 mm. long, 1.0-2.5 mm. wide. Mature capsules longer than wide, wider than long, or equal, 2.0-4.0 mm. long, 2. 0-4. 0 wide. Seeds 0.70- 1.65 mm. long, 0.50-1.20 mm. wide. A discussion of the relationships among the three varieties of this species appears below under the individual varieties. la. Houstonia purpurea L. var. purpurea Hedyotis umbellata Walt. Fl. Car. 85. 1788. Houstonia varians Michx. Fl. Bor. Am. 1:86. 1808. (in part). (Type in Mus. d'Hist. Nat., Paris; photograph of type in Gray Herbarium). Knoxia purpurea (L.) Lam. ex Poiret, Lam. Encyc. Meth., suppl. 3:225. 1813. Houstonia latifolia Willd. ex Roem. & Schult. Syst. Veg. 3:527. 1818. Anotis purpurea (L.) G. Don, Gen. Hist. 3:585. 1834. (in part). Hedyotis purpurea (L.) T. & G. Fl. N. Am. 2:40. 1841. Oldenlandia purpurea (L.) Gray, Man. ed. 2. 173. 1856. Houstonia purpurea L. var. pubescens Britton, Mem. Torr. Bot. Club 4:125. 1894. Chamisme purpurea (L.) Nieuwl. Am. Midl. Nat. 4:92. 1915. Houstonia purpurea L. f. pubescens (Britt. Fern. Rhod. 38:444. 1936. Hedyotis purpurea (L.) T. & G. var. purpurea f. pubescens (Britt.) Fosberg, Castanea 19:33. 1954. Rhizomes to 2.5 cm. long. Stems 1.0-4.5 dm. high, sparsely to densely pubescent (or rarely glabrate), with hairs (0.3-) 0.4-1.0 (-1.5) mm. long. Internodes numbering (4-) 5-8 (-9); median internodes (1.8) 2.0-5.5 (-9.5) cm. long. Median cauline leaves broad-ovate and sub- cordate to lanceolate-ovate, (I. O-) 2.5-5.0 (-6.3) cm. long (0.6-) 1.2-3.0 (-3.4) cm. wide, 1-3.2 times longer than wide, pubescent to rarely gla- brate above, ciliolate to glabrate on margins. Calyx-lobes linear-lanceo- late to ovate-lanceolate, (1.0-) 2.0-3.9 mm. long, 0.3-1.2 mm. wide, usually less than 34 as long as corolla-tubes, equalling or slightly exceed- ing mature capsules. Corollas purplish to white, (4-) 5-8 (-10) mm. 168 Rhodora [VOL. 61 long. Mature capsules 2.0-3.5 mm. long, 2.0-3.5 wide. Seeds 0.70-1.20 mm. long, 0.50-1.00 mm. wide. VAR. PURPUREA t — H g. PURPUREA MAP l. Distribution of two varieties of Houstonia purpurea. Time of flowering: April or May through July or August, sometimes continuing into September. Type locality: "Habitat in Virginia.” Type: Linnaean Herbarium. Sketch of type in Gray Herb- arium; of two plants on the type sheet only the one on the right side is considered to be the type of var. purpurea. Habitats and distribution: Moist and dry woods, openings, and borders of woods; roadsides; banks; cliffs; meadows; other secondary habitats. Ranges from Fayette County, Pennsylvania and Maryland south to southwestern Georgia, west across Ala- bama and Mississippi to Arkansas and easternmost Oklahoma, north into southwestern Missouri, cast to southern Ohio. Within 1959] Terrell — The Houstonia purpurea Group 169 these boundaries it occurs in all states and all physiographic prov- inces. There are very few collections, however, from the Missis- sippi Embayment. It is centered in the Appalachians and Ozarks, where it is present in more abundance than elsewhere. (Map 1). In its typical form var. purpurea is unique in its ovate or broadly ovate leaves and short calyx-lobes. It is the most abun- dant and widely distributed variety in this complex. Other com- ments about it are noted under var. calycosa. REPRESENTATIVE SPECIMENS: Pennsylvania. FAYETTE: Ohiopyle, Ricker 1225 (us). Maryland. TALBOT: 5 mi. NE of Easton, Tatnall 2964 (GH). Virginia. BRUNSWICK: S of Lawrenceville, Fernald 14855 (cH, PH, Us). GRAYSON: near Troutdale, Gleason 8773 (NY). SUSSEX: sw of Burt, Fer- nald & Long 6399 (GH, NY, Us). West Virginia. CLAv: 1 mi, N of Clay, Core 6343 (WVA). ROANE: Spencer, W. Va. Bot. Exped., 21 June 1928 (GA, wvA). North Carolina. ALAMANCE: E of Burlington, Oosting 33544 (DUKE). DURHAM: 5 mi. sw of Durham, Wiegand & Manning 2999 (cu, GH); Eno River, Oosting 3383 (oc). JOHNSTON: 2 mi. N of Clayton, Fox & Godfrey 1668 (GH, IA, NCSC) . PENDER: 10 mi. N of Burgaw, Godfrey 6591 (GH, us). South Carolina. CLARENDON: 14 mi. s of Manning, God- frey & Tryon 984 (F, GH, MO, NY, US). EDGEFIELD: 4 mi. wsw of Owdoms, Radford 22640 (NcU). MARION: s of Britton Neck, Bell 7833 (Ncv). Georgia: CHEROKEE: w of Canton, Duncan 8387 (GA, IA, MO, NCSC). EARLY: W of Hilton, Thorne 3881 (cu, 1A). Alabama. CULLMAN: Biltm. Herb. 519m. (us) Indiana. JACKSON: 14 mi. E of Chestnut Ridge, Deam 33524 (IND). Kentucky. CARTER: near Cascade Caverns, Smith et al 3535 (GH, vs). Tennessee. BLOUNT: Abrams Creek Ranger Station, Sharp 17482 (TENN). COFFEE: Rutledge Falls near Tullahoma, Sharp et al 4713 (TENN). KNOX: Knoxville, Ruth 3598 (Ny). Missouri. BARRY: Eagle Rock, Bush 45 (F, MO, NY, US). CHRISTIAN: 3 mi. sw of Chadwick, Steyer- mark 23155 (F, MO, NY, TENN, Wis). Arkansas. DREW: Monticello, De- maree 19111 (CU, FSU, MO, NY, TENN, WIS). LINCOLN: Star City, Demaree 19164 (GH, NY). Louisiana. TANGIPAHOA PARISH: 10 mi. E of Kentwood, Ewan 18704 (No). Oklahoma. LEFLORE: near Page, Palmer 21593 (Nx, Palmer). Texas. NEWTON: 2414 mi. Nw of Deweyville, Cory 22373 (cH). Ib. Houstonia purpurea L. var. montana (Small), n. comb. Houstonia montana Small, Fl. S.E. U.S. 1325. 1903. Houstonia purpurea var. montana Chickering, ibid. (as synonym). Hedyotis purpurea (L.) T. & G. var. montana (Small) Fosberg, Cas- tanea 19:33. 1954. Rhizomes to 5 cm. long. Stems 0.4-2.1 dm. high, glabrous or short- pubescent at nodes. Internodes numbering (4-) 5-8 (-9); median internodes 0.5-3.0 (4.4) cm. long. Median cauline leaves ovate, 0.8-3.0 170 Rhodora [VOL. 61 cm. long, 0.6-1.3 cm. wide, 1-2.5 times longer than wide, glabrous to slightly pubescent above, glabrate or glabrous on margins. Branches less than 4 cm. long. Calyces glabrous; calyx-lobes ovate-lanceolate, lan- ceolate, or ovate, 1.0-3.1 (-4.5) mm. long, 0.6-1.2 (-3.0) mm. wide, usually less than three-fourths as long as corolla-tube, equalling or slightly exceeding mature capsules. Corollas deep purple, 8-12 mm. long. Mature capsules 2.0-4.0 mm. long, 2.0-4.0 mm. wide. Seeds (from one collection) 1.40-1.65 mm. long, 0.90-1.20 mm. wide. Time of flowering: June through August or September. Type locality: Roan Mountain, North Carolina. Lectotype: J. W. Chickering, Jr., September 12, 1877. (NY). Duplicates: (F!, PH!) Habitats and distribution: Crevices of exposed rocks and in moist loam on rhododendron bald, Roan Mountain, N. C.-Tenn., and among rocks on Grandfather Mountain, N. C. Known only from summits of these two mountains, at altitudes of about 6000 feet (1829 meters). The specimen I have chosen as lectotype appears to be the orig- inal collection by Chickering, which led Small to publish on this taxon. Chickering’s collection label bears the identification, H. purpurea L. var. montana. He “proposed” the epithet but did not ever publish it in any combination. Small considered the taxon a species and published the varietal combination only as a syno- nym, which did not constitute valid publication (Art 37, Paris rules). In regard to the correct citation for the varietal combina- tion Fosberg's comments (1954, p. 33) are pertinent. Var. montana differs from var. purpurea as follows: stems are glabrous or nearly so; internodes are shorter; leaves smaller; cor- ollas larger and colored deep purple; calyx-lobes, capsules, and seeds are larger. It is obvious that var. montana may, in most characteristics, be merely a high altitude, dwarfed form of var. purpurea. However, until this is proven by transplant experi- ments or by other means, it seems preferable to retain it as a variety. The rarity of occurrence also tends to favor the present course; there are a number of other mountain tops with high altitude balds and exposed rocks — why does not var. montana occur on these? Var. purpurea is common at such altitudes on other mountains in the southern Blue Ridge, but has not been 1959] Terrell — The Houstonia purpurea Group 171 scen by the writer in habitats as exposed as those on Roan and Grandfather Mountains. REPRESENTATIVE COLLECTIONS: Roan Mountain. (summit is on border, Mitchell Co., N. C.-Carter Co., Tenn.). Some early collections are: Gray & Carey, July 1841 (cH) (first known collection) ; Vasey, 1878 (NY, US); Gray, Sargent, Redfield, & Canby, 19 June 1879 (GH, Ny, PH); Chickering, 5 July 1880 (F, Mo, Ncu, Us); Porter, 9 July 1880 (en); J. D. Smith, 10-16 July 1880 (r, vs). Some later collections include: Ball, 15-17 Sept. 1884 (1A, oc, PH, wis) ; Lyon's Bluff, elev. 6350 ft., Small & Heller, 16 July 1891 (r, MO, Ny, PH, US); summit, Blomqutst 1961 (DUKE); near Roan High Bluff, Shanks 3008 (FsU, TENN, US). Grandfather Mountain. (summit on border, Avery-Watauga Cos., N. C). Heller, II Aug. 1890 (F, Mo, NY); Small & Heller 72, 6 Aug. 1891 (cu, r) ; Churchill, 18 June 1899 (cH, TENN) ; Stewart, 6 July 1938 (NCU, NY). Ic. Houstonia purpurea L. var. calycosa Gray, Syn. F1. 1 (2) :26. 1884. Hedyotis lanceolata Lam. ex Poiret, Lam. Encyc. suppl. 3:14. 1813. Anotis lanceolata (Poir.) DC. Prodr. 4:433. 1830. Houstonia macrosepala Nutt. ex T. & G. Fl. N. Am. 2:40. 1840. (as synonym of an unnamed variety of Hedyotis purpurea) . ?Diodia Frankii Steud. & Hochst. Abh. Bóhm. Ges. Wiss. V. 3:86. 1843. (as synonym; this publ. not seen). ?Spermococe lanceolata (Poir. Frank, ibid. (as synonym). ?Hedyotis Frankii Presl, Bot. Bemerk. 86. 1844. (as synonym; publ. not seen). Hedyotis calycosa Shuttlew. ex Gray, PI. Wright. 1:81. 1852. (as synonym). Houstonia calycosa (Shuttlew. ex Gray) C. Mohr, Contr. U. S. Nat. Herb..6:739.: 1901. Houstonia lanceolata (Poir.) Britton, Man. 861. 1901. Houstonia lanceolata f. albiflora Standl. Rhod. 34:177. 1932. Hedyotis purpurea (L.) T. & G. var. calycosa (Gray) Fosberg, Cas- tanea 19:38. 1954. Hedyotis purpurea var. calycosa f. albiflora (Standl.) Fosberg, loc. cit. 34. 1954. Rhizomes to 2.5 cm. long. Stems 1.0-4.5 dm. high, pubescent to gla- brate, usually sparsely pubescent. Internodes numbering 5-12: median internodes (1.8-) 2.0-5.5 (-9.5) cm. long. Median cauline leaves 3-5 nerved, lanceolate to narrow-lanceolate, 1.7-3.3 cm. long, 0.4-1.0 cm. wide, 3.3-6 times longer than wide, pubescent to glabrate above, cilio- late to glabrate on margins. Calyces glabrous to pubescent: calyx-lobes linear-lanceolate to lanceolate, 4.0-7.0 mm. long, 0.3-1.2 mm. wide, one- half as long to slightly exceeding corolla-tubes, considerably exceeding 172 Rhodora [VOL. 61 mature capsules. Corollas purplish to white, (4-) 5-10 mm. long, tend- ing to be pubescent within; corolla-tubes tending to be more widely flared than in var. purpurea. Mature capsules 2.0-3.5 mm. long, 2.0-3.5 mm. wide. Seeds 0.70-1.20 mm. long, 0.50-1.00 mm. wide. Time of flowering: April or May through June or July, rarely August. Type and type locality: The lectotype, collected by Rugel, is inscribed, “Hedyotis calycosa Shuttl. n. sp. In montibus prope Huntsville, Alabama, legit Rugel, Oct. 1843.“ (cH!; duplicate in NY!). Habitats and distribution: Cedar glades, barrens, dry or some- what mesic woods, rocky open places, fields. Scattered and local in New England; Wayne County, Michigan; Barbour County, West Virginia; southern Ohio south to northern Georgia, and Alabama, west to Missouri, Oklahoma, and Arkansas, north to central Illinois and central Indiana. Centered in the Interior Low Plateau (Fenneman) and adjacent Provinces. (Map 1) In the treatment of this complex all names relating to white- flowered forms have been reduced to synonymy. It seems desir- able to recognize color forms only when there is discontinuity in color. In all taxa of this group corolla color varies among colo- nies or within one colony or even among different flowers on the same plant. Color varies imperceptibly from purple or pink- purple to white through various intermediate degrees of purplish tinges and lines on a white background color. White-flowered plants are of frequent occurrence. The name, H. purpurea, isa misnomer, since this species often has white flowers. In regard to the nomenclature of the present variety, it seems clear that under Art. 60, Paris rules, the combination adopted here is correct as it is the earliest in the rank of variety. Although the name, Hedyotis calycosa, is invalid (it was published only as a synonym, that does not mean that the combination, Houstonia purpurea var. calycosa, is illegitimate (Arts. 69 and 72 seem to bear on this point). Gray's description of this variety is adequate. In choosing a type specimen it seems desirable to fix the appli- cation of the epithet, calycosa, by going back to the original spec- imens to which this epithet (in an invalid specific combination) 1959] Terrell — The Houstonia purpurea Group 173 was applied. For this reason the lectotype listed above was chosen; it is rather typical of the natural populations included in the present variety. Gray's Manual, 8th edition, states that the capsules of var. purpurea are broader than high, those of var. calycosa as high as broad, but this distinction does not hold when large numbers of specimens are examined. The Manual also states that calyx-lobes in the latter variety range in length from 5 to 9 mm. The longest lobes seen during the present study were not more than 7 mm. long. The main differences between these two entities are in length of calyx-lobes and in shape of median cauline leaves; the former character is the more important. To deal with the former charac- ter it is preferable to use the absolute length of lobes rather than the ratio, calyx-lobe length to corolla-tube length, because corolla length may vary considerably as a result of environmental in- fluence, whereas lobe length is much less subject to such influ- ence. In addition to the two main differences in these varieties, other lesser differences also exist: var calycosa tends to have more internodes, to be sparsely pubescent rather than ranging from densely pubescent to glabrate, and to have wider, more flaring, corollas which may be more pubescent within. These and other more minor differences are most noticeable in plants from the Cedar Glades of the Nashville Basin. Specimens of typical var. purpurea have short calyx-lobes and ovate leaves. They occur in mesic woods throughout much of the eastern United States (see map). Typical specimens of var. caly- cosa have long calyx-lobes and lanceolate leaves. They occur pri- marily in cedar glades, barrens, and dry woods in the Cedar Glades of the Nashville Basin and in other parts of Tennessee, in central and western Kentucky, in northern and west-central Alabama, in at least two counties in central Mississippi, and in southwestern Missouri and adjacent Arkansas. The preceding paragraph refers to typical specimens. When all available specimens referable to these two varieties are ex- amined, it is evident that intergradation in the two primary diag- nostic characters is complete. Plants with long calyx-lobes pre- 174 Rhodora [VOL. 61 dominate from Ohio, Indiana, and Illinois south into Alabama and Mississippi. These plants may have leaves ranging from nar- row-lanceolate to broad-ovate; therefore, lanceolate leaves are not always correlated with long calyx-lobes. If we were to designate var. calycosa as the variety with lanceolate leaves and long lobes, this would leave unclassified many collections with long lobes and ovate leaves (the latter a feature of var. purpurea). In this study it has been determined that length of lobes is the more im- portant of the two primary characteristics; accordingly, this characteristic is used to provide positive separation. This reflects the definite regional distinction in lobe length: plants with long lobes are primarily midwestern; plants with short lobes have a much wider distribution and tend to grow in more mesic habi- tats. It is believed that gene flow between these two varieties has been quite extensive in the Middle West. The populations in southern Ohio, Indiana, and Illinois are thought to be almost entirely "weedy" biotypes, the result of introgression among “purer” populations in Kentucky and Tennessee. Following in- trogression these “impure” biotypes may have spread northward onto the glaciated Central Lowland. In southern Ohio such plants resemble var. purpurea superficially because of their ovate leaves, but their longer calyx-lobs show the influence of genes from var. calycosa. This intervarietal introgression could have been facilitated by “hybridization of the habitat” after destruc- tion of the original forest cover, and is assumed to have been accelerated since white settlement. Obviously, this theory to explain the existence of the “impure” populations in the Middle West relies heavily on aspects of the rather extensive and well known literature on introgression. New England populations resembling these two varieties are a separate problem in themselves from the standpoint of origin. In morphology nearly all of these fall within the present circum- cription of var. calycosa, but are not really typical of this variety. They are completely disjunct from other populations of both varieties. Apparently, they represent remnants of a formerly more widespread extension of var. calycosa. This may not neces- 1959] Terrell — The Houstonia purpurea Group 175 sarily vitiate the above theory on origin of midwestern forms. Further opinions on their origin is pure conjecture. Some New England plants vary in the direction of H. longifolia. One col- lection from Wayne County, Michigan, resembles New England plants. In summary, vars. purpurea and calycosa are distinct in their H. CANADENSIS u. TENUIFOLIA © rag V4 HERE EUR E bes nea Lor N men 110 eee Co n ECT ATE Fa e? | Ede MUS ato. Ne N E MU CAS Sr GLAS & (LAS | ; ) Map 2. Distribution of Houstonia canadensis and H. tenuifolia. Florida collection ot later species is atypical. 176 Rhodora [VOL. 61 extremes — they have in the past been considered separate spe- cies — but intergrade so greatly that they may only be separated statistically on one character difference — length of calyx-lobes. This probably reflects a regional difference in gene frequency. They are sympatric, with different centers of distribution but with overlapping ranges. Where they occur in the same region they usually occupy different habitats. The intergradation in morphology and in habitat may be the result of introgressive hybridization. Representative collections from New England include the following: Maine. CUMBERLAND: Huston's field, Cumberland Center, Chamberlain 1086 (us). york: North Berwick, Parlin 797 (us). Massachusetts. BERKSHIRE: Shefheld Plain, Townsend, 10 Aug. 1896 (cv). Connecticut. OXFORD: Oxford, Harger, 14 June 1896 (cu). New York. Essex: N of Schroon River P. O., House 27763, 17 Sept. 1941 (cv, GH). Plants with broad-ovate leaves and long calyx-lobes are exemplified by the following: Indiana. HARRISON: 2 mi. sk of Corydon, Deam 16331 (IND) (lobes to 6.9 mm. long). Posey: 9 mi. sw of Mt. Vernon, Her- mann 6655 (MICH, US). The following represent typical var. calycosa, with lanceolate or ovate- lanceolate leaves and long calyx-lobes: Tennessee. pAvIDSON: Mt. View, Svenson 10006 (DUKE, GH, TENN). KNOX: l mi. E of Mascot, Cain et al, 2 June 1937, Plantae Exs. Grayanae 870, (CU, DUKE, F, GH, IA, MICH, MO, NCSC, NCU, NY, PH, TENN, US, WIS, WVA). RUTHERFORD: cedar glade near Lavergne, Harper, 17 May 1923 (GH, Ny, US). WILSON: near Leb- anon, Shanks 1540 (TENN). Illinois. MACON: Clokey 2424 (F, GH, MICH, MO, NY, Us). Missouri. BARRY: Eagle Rock, Bush 3148 (E, Ny, vs). Arkansas. CARROLL: near Elk Ranch, Palmer 39458 (cH). 2. Houstonia canadensis Willd. ex R. & S. Syst. 3:527. 1818. Houstonia ciliolata Torrey, Fl. N. & Mid. U.S. 1:173-174. 1894. Hedyotis ciliolata (Vorr.) Spreng. Syst. 4: cur. post. 40. 1827. Anotis ciliolosa G. Don, Gen. Hist. 3:535. 1834. Oldenlandia purpurea (L.) Gray var. ciliolata (Torr.) Gray, Man. ed. 2. 173. 1856. Houstonia longifolia Gaertn. y ciliolata (Torr) Wood, Class-Book, ed. 1861. 403. 1861. Houstonia purpurea L. var. ciliolata (Torr) Gray, Man. ed. 5. 212. 1868. Chamisme ciliolata (Torr. Nieuwl. Am. Midl. Nat. 4:92. 1915. Hedyotis canadensis (Willd. ex R. & S.) Fosberg, Va. Jour. Sci. 2:110. 1941. (in part) 1959] Terrell — The Houstonia purpurea Group IU Houstonia setiscaphia Carr, Rhod. 46:309. 1944. Hedyotis purpurea (L.) T. & G. var. ciliolata (Vorr.) Fosberg, Cas- tanea 19:34. 1954. Hedyotis purpurea (L.) T. & G. var. setiscaphia (Carr) Fosberg, loc. cit. 35. 1954. Perennial herbs. Rhizomes branched or simple, slender, usually more or less horizontal, to 8.3 cm. long, bearing many small roots. Stems erect or ascending, one to several, tetragonal and sometimes. slightly winged, 0.4 2.5 dm. high, glabrous to pubescent below, especially at nodes, with multicellular hairs to 0.7 mm. long. Intérnodes numbering 3-5 ( 6); median and lower ones 1-8 cm. long, longer than upper ones. Stipules deltoid to ovate-lanceolate, rounded to acute to short-acumi- nate or uppermost ones irregularly lobed or erose, often with dark glands on margins, to 4 mm. long, to 5 mm. wide; stipules of median nodes usually more or less ovate, rounded above, and about as wide as long. Basal leaves forming winter rosettes or in northern part of range forming short offseis, persisting through fruiting stage, varying from oval to spatulate, attenuate below into petioles shorter than to slightly longer than blade, to 3.9 cm. long, to 0.8 cm. wide, glabrous below, hir- tellous to pubescent above, ciliolate (sparsely ciliolate to glabrate in cer- tain populations in south central Ohio), Cauline leaves l-nerved, oblan- ceolate, obovate, elliptic, or narrowly elliptic, (0.5—) 0.8-2.3 (-3.1) cm. long, 0.15-0.85 cm. wide, usually 2.5-6 times longer than wide, glabrous or pubescent on midrib below, pubescent to glabrous above, ciliolate to glabrate on margins. Branches ascending or spreading, less than 8 (rarely to 12) cm. long and usually arising from uppermost 1-3 nodes. forming a rather open, few-many flowered, inflorescence. Pedicels slen- der, to 7 mm. long. Calyces glabrous to hirtellous; lobes erect or spread- ing, linear-lanceolate to lanceolate, 1.6-3.5 (74.5) mm. long, 0.5-1.0 mm. wide, one-fourth as long to nearly equalling corolla-tube, equalling or exceeding mature capsules. Corollas white to purplish, (3-) 4-9 (-11) mm. long, granular to pubescent within; corolla-tubes 2-7 mm. long, 1.5-4 mm. wide distally; corolla-lobes 1.5-5 mm. long, 1-2.5 mm. wide. Mature capsules 2-3.5 mm. long, 2-3.2 mm. wide. Seeds 0.70-1.30 mm. long, 0.50-0.90 mm. wide. Time of flowering: April to June, rarely into August. Type locality: E. Canada". [fype: Willdenow no. 2684, Botanical Museum, Berlin. Habitats and distribution: Dry to moist woods; openings; cliff tops; banks and roadsides; sandy shores; often in rocky soil over shale or limestone. Bruce Peninsula, Manitoulin Island, and Northumberland County, Ontario, southward through Mich- igan, Ohio, western New York, southwestern Pennsylvania and 178 Rhodora [VOL. 61 adjacent West Virginia, southernmost Indiana, central and east- ern Kentucky, southwestern Virginia, Tennessee, to Lookout Mountain area of ‘Tennessce-Georgia. Distribution generally rather disjunct. (Map 2) The locating of the type specimen in the Berlin herbarium establishes that H. canadensis, not H. ciliolata, is the correct name. Two photographs of the type specimen were forwarded to the writer and clearly represent the present species. The relation between this species and H. longifolia was dis- cussed by Standley and by Fosberg. Standley (1936) in a note on the apparent breakdown of characters separating these two spe- cies in specimens from Indiana, advocated either the reduction of H. canadensis to synonymy under H. longifolia or else the rec- ognition "as a rather poorly marked variety H. longifolia var. ciliolata (Vorr.) Wood". Fosberg (1941a) combined the two spe- cies, finding it necessary to adopt a new combination under Hedyotis. Fosberg (1954) made all species of this complex varie- ties of H. purpurea. The two species in question are as distinct as any in this com- plex. Comparison of them in Indiana is rather misleading because H. canadensis occurs in only two counties in southernmost Indi- ana and H. longifolia in its typical form does not occur in the state at all. Specimens seen of H. longifolia from Indiana include two groups: collection from two Ohio River counties not quite typical of var. compacta; a larger number of collections from several central and nothern counties regarded here as "atypical" var. longifolia. Houstonia setiscaphia was described by Carr (Rhod. 46:306- 310. 1944) from cedar glades in southwestern Virginia (Type in Gray Herbarium; isotype in herb. Univ. of Pa.; Lloyd G. Carr 1110, dry glades or barrens among the cedars ca. V5 mile west of Jonesville, The Cedars”, Lee Co., Va., July 10, 1942). Located in the Ridge and Valley Province just south of the Cumberland Mountains, these glades are small in area but very similar ecolog- ically and floristically to the well-known glades in the Nashville Basin, Tennessee, where only H. purpurea var. calycosa occurs. In company with Dr. R. W. Barbour the writer made two col- 1959] Terrell — The Houstonia purpurea Group 179 lections of Carr's new species, one collection five and one-half, the other eight miles southwest of Jonesville, on June 22, 1948. Carr’s published comparison of H. setiscaphia with its nearest relative, H. canadensis, stated that the two species differed in the following characters: height of plants; vestiture of stem-angles (H. setiscaphia has a grayish aspect due to greater amount of vestiture); compactness of inflorescence; shape of cauline and basal leaves; vestiture of calyx; size of corollas. In the present writer’s M. S. thesis (Cornell University, 1949) a tabular comparison was made of two of my collections of H. setiscaphia with sixteen collections of H. canadensis from throughout its range. This comparison demonstrated that very little difference exists between the two taxa in regard to the first three characters listed above as well as in the shapes of cauline leaves. Average sizes of corollas differ somewhat, the average length being 4.7 mm. in Carr’s new species, and 6.4 mm. in H. canadensis. This difference, however, is thought to be due to en- vironmental modification. In all taxa of the H. purpurea com- plex later-flowering plants have smaller corollas, caused perhaps by reduced precipitation in summer. Both Carr and the writer collected H. setiscaphia in later-flowering condition. Environ- mental influence in strong, also, on the density of inflorescences. Shapes of basal leaves in H. seliscaphia are usually oblanceo- late but vary to elliptic or oval; in H. canadensis shapes are more commonly oval or elliptic but may vary to spatulate, oblanceo- late, or obovate. ‘This difference overlaps so much that it is judged not to be significant. The most important characteristic of H. setiscaphia is its “hispid” (or hirsutulous) calyces. The type specimen, examined while on loan from the Gray Herbarium, proved to be slightly more hirsutulous than other collections. Calyces of plants of H. setiscaphia are usually hirsutulous but vary to glabrous. Speci- mens of H. canadensis are usually glabrous but four out of sixteen collections vary from glabrous to pubescent, one collection var- ies from glabrous to hirsutulous, and one collection consists en- tirely of hirsutulous plants. Carr's interpretation of H. canadensis was based apparently on 180 Rhodora [VOL. 61 a preponderance of northern collections. Southern collections are more hairy than northern ones, and are, therefore, more like H. setiscaphia. The three available collections of H. setiscaphia demonstrate more variability than Carr's report suggested. This taxon may even have glabrous calyces, and both taxa overlap in other characters as well. An additional collection, (Carr 775, clinging in crevices of limestone, Natural ‘Tunnel, Scott Co., Va.) from a county adja- cent to Lee, was seen among specimens on loan from the Gray Herbarium. It includes plants with glabrate to rather densely short-pubescent calyces; therefore, it falls within the range of variation possessed by both taxa. Besides the Scott County, Virginia, collection, perhaps more nearly referable to H. canadensis than to the other presumed tax- on other collections show that the two are not well-separated geo- graphically. Collections of H. canadensis, which is likely to be disjunct anyway, from Grainger County, Tennessee, and Whitley County, Kentucky, are only forty and fifty miles, respectively, from Lee County, Virginia. (see Map 2). It is concluded that morphological differences between the two taxa are not even sufficient to maintain H. setiscaphia as a vari- ety of H. canadensis. The former may be thought of as a local extreme or local race which is genetically slightly different. Not giving it formal taxonomic recognition is consistent with the treatment here of the remainder of the H. purpurea complex. Dr. E. Lucy Braun has called attention. (in correspondence) to local populations of H. canadensis which have glabrate basal leaves. These occur in unglaciated south central Ohio in Adams, Highland, and Ross Counties. My own sampling of these popu- lations indicated that they represent a local extreme which over- laps rather extensively in the single character with ordinary H. canadensis. Glabrate plants did not seem to be restricted to any one geological formation. Collections consisted of varying proportions of glabrate, ciliolate, and intergradant plants. Con- sidering the great variability in H. canadensis as a whole this local extreme is not worth taxonomic recognition. (to be concluded) 1959] Harris — Paucedanum palustre in Essex County 181 PEUCEDANUM PALUSTRE, AN INTERESTING ADDITION TO THE FLORA or Essex County, MassAcHUsETTs. On July 9, 1958, while lead- ing a flower walk on Pike Bridge Road in the Artichoke Region of West Newbury, Massachusetts, one of the participants asked me to name a fairly large flowering umbellifer growing under apparently natural conditions near the water and I was somewhat embarrassed to confess that I was unable to do so. In September, Mr. Henry Lewis brought me good fruiting material of what was obviously the same taxon which he had found on Rogers Street about half a mile from where the flowering specimen was col- lected. Even with the fruit I still was unable to name the plant and I turned to Mrs. Claude Weber and Dr. Lincoln Constance for help. ‘They were finally able to identfy it as Peucedanum palustre (L.) Moench., of Europe. Apparently this is the first North American record of the species. The long slender involu- cral bracts of the compound umbels are a very striking feature of the inflorescence. Ihe specimen has been deposited in the Gray Herbarium and I hope to obtain additional material next year. Roadside, Pike Bridge Road, West Newbury, Essex County, Massachusetts, Stuart K. Harris 18059 (9 July 1959). — STUART K. HARRIS, BOSTON UNIVERSITY. Correction FOR May issue. — The literature cited portion of the article on Rhododendron maximum by A. R. Hodgdon and R. Pike was in- advertently omitted. This matter is as follows: LITERATURE. CITED l. BATCHELDER, F. W. Preliminary list of plants in the vicinity of Manchester, New Hampshire. Proc. Manchester Inst. of Arts & Sci. E:110-111, 1899. 2. Earon, R. J. Another New Hampshire Station for Rhododendron maximum. Rhodora 49:161. 1938. 3. FARLow, W. G. Rhododendron maximum in New Hampshire. Rhodora 18:25. 1916. 4. luris, HUGH H. Studies in Virginia Plants II. Rhododendron maxi- mum in the Virginia Coastal Plain and its Distribution in North America. Castanea 21:114—124. 1956. 5. KNowLTON, C. H. Rhododendron maximum in New England. Rhodora 52:215-218. 1950. 6. Moonr, WirLiAM. Contributions to the History of Derryfield. 39, F897: Volume 61. No. 725, including pages 127-156, was published July, 1959. y Qeereenc® LIBRARY i Dodota JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Conducted and published for the Club, by REED CLARK ROLLINS, Editor.in.Chief ALBERT FREDERICK HILL STUART KIMBALL HARRIS RALPH CARLETON BEAN Associate Editors CARROLL EMORY WOOD, JR, IVAN MACKENZIE LAMB Vol. 61 July, 1959 No. 727 CONTENTS A New Species of Justicia from Florida. Joseph Monachino and Emery C. Leonard l... 183 Two Crucifers New to Essex County, Massachusetts. Stuart K. Harris mE zu CDS SA 187 A Revision of the Houstonia purpurea Group (Rubiaceae). Edward E. Terrell. (concluded) „ e 188 The New England Botanical Cluh, Inc. Botanical Museum, Oxford St.. Cambridge 38, Mass. RHODORA.—A monthly journal of botany, devoted primarily to the flora of North America and floristically related areas. Price, $6.00 per year, net, postpaid, in funds payabie at par in United States currency in Boston; single copies (if available) 60 cents. Back vol. umes 1-58, with a few incomplete, can be supplied at $5.00 per volume. Volume 59— available at $6.00. Somewhat reduced rates for complete sets can be obtained upon application. Scientific papers and notes, relating directly or indirectly to the plants of North America, will be considered by the editorial com— mittee for publication. Articles concerned with systematic botany and cytotaxonomy in their broader implications are equally accept- able. All manuscripts should be double-spaced throughout. Please conform to the style of recent issues of the journal. Illustrations can be used only if the cost of engraver's blocks is met through the author or his institution. Forms may be closed five weeks in advance of publication. Extracted reprints, if ordered in advance, will be furnished at cost. Address manuscripts and proofs to Reed C. Rollins, Gray Herbarium, 22 Divinity Avenue, Cambridge 38. Mass. Subscriptions and orders for back issues (making all remittances payable to RHODORA) should be sent to Albert F. Hill. Botanical Museum, Oxford Street, Cambridge 38. Mass. Second Class Postage Paid at Boston, Mass. EATON PRESS, Inc. Specialists in Learned Publications 17 SPRING STREET WATERTOWN 72, MASS. Rhodora JOURNAL OF THE NEW ENGLAND BOTANICAL CLUB Vol. 61 July, 1959 No. 727 A NEW SPECIES OF JUSTICIA FROM FLORIDA Josera MONACHINO AND EMERY C. LEONARD A rich collection of plants gathered in the southeastern United States by John K. Small and co-workers has lain in storage at The New York Botanical Garden for about twenty-five to forty or more years. It has suffered moderately from the ravages of time and neglect. The specimens were originally well-prepared and duplications are frequently ample. Unfortunately they are unnamed and they are often without collection-number or label, and the field-data is poor. Through the support of George R. Cooley, and the intermedi- ation of Bassett Maguire, it is now possible to process this ma- terial so that it can be placed in circulation. The first set will be deposited in the herbarium of The New York Botanical Garden. The first author has been assigned the task of preparing this long-neglected “Small Collection” for herbaria. He is not study- ing it taxonomically, but in the way of naming he is merely sug- gesting the genus for each specimen so that it can be filed in the proper approximate position in the herbarium. Naturally, it 1s impossible to shut the eyes of a taxonomist completely. Instances of apparently new state-records were seen: e.g. Oxytropis Lam- bertii Pursh var. abbreviata (Greene) Barneby and Senecio plat- tensis Nutt from Louisiana, determined by Rupert Barneby and Theodore Barkley, respectively; Rhynchosia galactioides (Nutt.) Endl. from Apalachicola, Florida. There was revealed a good ample collection of Clitoria fragrans Small bearing locality and date of collection identical with those given for the type. As hith- erto there was neither type nor any other specimen of this species 183 184 Rhodora [VOL. 61 in the herbarium of The New York Botanical Garden (at least since 1945), the discovery is important. Quite astonishing too was the discovery of a species of Nyssa from the Florida Keys! It is a fruiting specimen with coriaceous narrow leaf-blades suggest- ing an extreme form of N. biflora Walt., or N. ursina Small. The label reads: J. K. Small and N. L. Britton, hammock, Lignum Vitae Key, Monroe County, December 13, 1919. ‘The most interesting discovery was the new species of Justicia published herewith. One of the old collections, Small 12682, al- though clearly a Justicia, did not appear to be known from Flor- ida. Search in the herbarium revcaled its identity with Cooley 2481 from the same area. Comparisons with other material also proved that the species was undoubtedly a novelty for Florida, and subsequently Mr. Leonard ascertained it to be altogether new to science and he supplied the accompanying description. It is named in honor of George R. Cooley, not only for his collec- tion of the type, but mainly in acknowledgment of his continued generous support both in the present and in other work dealing with the flora of the southeastern United States. The illustration was drawn by Dorothy H. Marsh under the supervision of Carroll Wood at Harvard University from preserved material collected at the same time as Cooley 6334. Justicia Cooleyi Monachino & Leonard, sp. nov. Herba erecta vel ascendens, usque ad 40 cm. alta, caulibus ramosis, subquadrangularibus vel basi teretibus, 2 mm. crassis, pilis plus minusve bifariam pilosis, pilis retrorsis vel aliquando patulis, usque ad 0.75 mm. longis, albidis; laminae foliorum oblongo-ovatae, usque ad 7.5 cm. longae et 3 latae, apice breviter acuminatae (apice ipso subobtuso), basi acutae, in petiolum decurrentes, submembranaceae, integrae, parce vel mediocriter pilosae, pilis usque ad 1 mm. longis, albidis, costa et venis (3— vel 4— paribus) et venulis crasse reticulatis aliquando pormi- nentibus, cystolithis obscuris; petioli usque ad 1.5 cm. longi, dense pilosi, pilis rectis et patulis vel curvatis, usque ad 1 mm. longis; spicae erectae vel ascendentes, usque ad 6 cm. longae, laxae, rectae vel leviter curvatae, internodiis usque ad 13 mm. longis, puberulis, pilis patulis vel retrorsis, 0.13 ad 0.32 mm. longis, pilis alteris longioribus glandulosis, paniculam terminalem formantes, pedunculis usque ad 23 mm. longis, teretibus, puberulis, pilis brevioribus eglandulosis et pilis longioribus glandulosis intermixtis; flores sessiles; bracteae lineares, 2.5-3 mm. longae, 0.5-0.75 mm. latae, acutae, costa prominente, intus glabrae, ex- 1959] Monachino & Leonard — A New Justicia 185 terne. pilosae, pilis 0.18-0.48 mm. longis, paucis, aliquando glandulosis; bracteolae bracteis similes; calycis segmenta 4, lanceolata, 5 mm. longa, l mm. lata, apice subacuta, intus glabra, externe pilosa, pilis patulis vel ascendentibus, pilis eglandulosis et glandulosis intermixtis; corolla 7-8 mm. longa, externe parce pilosa, pilis usque ad 0.16 mm. longis, patulis, intus glabra, pallide purpurea, labio inferiore intus fusco-pur- pureo, medio oblique albido vittato excepto, tubo 4.5 mm. long, basi 1.5 mm. lato, apice 2.5 mm. lato, labio superiore erecto, ovato, ca. 0.3 mm. longo, basi 2 mm. lato, apice angustato, labio inferiore trilobato, patulo, 4 mm. longo, lobis rotundatis, 2 mm. longis, 1.5 mm. latis, mar- ginibus plus minusve crispatis; stamina supra orem tubi corollae ca. 3 mm. exserta, lobis antherarum 0.75 longis, superpositis, obliquis, con- nectivo 0.25 mm. longo, lobo inferiore caudato, cauda plana, spathulata, ca. 0.5 mm. longa: capsulae clavatae, 12 mm. longae, 3 mm. latae, ca. | mm. crassae, hirtellae, pilis rigidis, retrorsis, usque ad 0.25 mm. longis, pilis eglandulosis et glandulosis intermixtis; retinacula curvata, 2 mm. longa, 0.25 mm. lata, apice tenuia, truncata, erosa, nitida; semina plana orbiculata, ca. 2 mm. diam., villosa, pilis patulis, usque ad 0.32 mm. longis, apice ancoraeformi. Fic. 1. Justicia Cooleyi Monachino & Leonard. a, flower and bud, lateral view, X 4; b, opened corolla with one stamen removed, X 4; c, stamen (portion of filament omitted, X 10; d, immature fruit and flower from which corolla has fallen, x 2.. Type in the Herbarium of the New York Botanical Garden, collected in a wet woods on Indian Hill, two miles nw. of Chinsegut Hill, Her- nando Co., Florida, November 20, 1957, by George R. Cooley (No. 27181). Paratypes: Low hammock near Mascotte, Lake County, Florida, December 4, 1925, John K. Small 12682 (ny, us); Brooksville, Hernando Co., Florida, September 5, 1934, Small, West & McFarlin s.n. (NY); common in high hammock on Indian Hill, two miles nw. of Chinsegut Hill, Hernando Co., Florida, December 1, 1958, George R. Cooley 6331 (Nv). Carroll Wood informs us that another sheet of the same 186 Rhodora [VOL. 61 plant in the Gray Herbarium is from "roadside on US 98, 2.7 mi. north of Brooksville, Hernando Co., Florida, Ray Garrett, 11 Aug. 1953.” Justicia Cooleyi differs from J. ovata (Walt.) Lindau, which it superficially resembles, and can be easily separated by the thin- ner leaf blades, these regularly oblong-ovate; by the dense, spreading or retrorse pilosity of the entire plant with more or less of the hairs gland-tipped; by the dark velvety color of the in- ner surface of the lower lip of the corolla; and by the densely villous sceds, their hairs anchor-tipped. The true relationship of Justice Cooleyi is, however with Justicia Pringlei Rob., of Mexico, and Justicia galapagana Lin- dau, of the Galapagos Islands, Ecuador. These three species are extremely close, having in common the ovate to oblong leaf blades, the secund flowers, the mixed glandular and eglandular hairs more or less clothing the entire plant, and the anchor-tipped hairs of the seeds. Compared with J. Pringlei the leaf blades of that species are usually smaller, averaging about 3 cm. in length, the spikes are more spreading, shorter, (averaging 3 cm.) with the rachis hairs practically all gland-tipped and the corolla deep lilac instead of purple. The spreading hairs of the stems are usually more numerous and longer (up to 2.5 mm. long). In J. galapagana the corolla is broader, the spikelets fewer, shorter, usually producing only 3 flowers instead of 6, and more heavily and densely glandular puberulous. Lindau, in his description of J. galapagana, states that the seeds are smooth (laevis) and brown- tomentose in contrast with those of J. Pringlei which are foveo- late and pubescent. The seeds of J. Cooley: are pilose with spreading whitish hairs arising from thickened papilliform bases. This series of Justicia deserves further study with more abundant material available. Writing from Chinsegut Hill near Brooksville, the type local- ity of the endemic Campanula Robinsiae Small, and making ref- erence to the precise geographical station of the type of Justicia Cooleyi, George R. Cooley provided the following information: where Sections 15, 16, 21, and 22, T 21S, R 19E, meet, the land rises from an elevation of 100’ toward the east to a height of 240’, dropping to 80^ one-half mile to the west. This rise is locally 1959] Monachino & Leonard — A New Justicia 187 known as Indian Hill. It is part of a great area of broad-leaved trees known as Annutteliga or Annuttalagga Hammock. Lime- rock mining has destroyed almost all of this extensive wood in which John K. Small and the St. John's found many endemic plants, particularly ferns. In the hammock on Indian Hill evi- dences can be found of lumbering operations many years ago. Probably this wood has been untouched for thirty years. Oaks, hickories, magnolias, and sweet gum trees predominate. Lower trees include Ostrya and Carpinus. The undercover is rather dense and in places is the result of earlier clearings. Perhaps the most abundant of the shrubby plants is Ilex vomitoria Ait., the Yaupon. Among the grasses and herbs grows Justicia attracting attention by its small rich purple flowers. Rarely more than three or four blossoms appear on a plant at one time and usually scat- tered. The flowering season is long, at least from November to March. — The New York Botanical Garden, New York and U. S. National Museum, Washington. Two Crucirers NEw TO Essex County, MASSACHUSETTS. Dur- ing September 1958 I found two uncommon introduced crucifers which were new to the county flora. Alliaria officinalis Andrz. has not yet become abundant in New England. It is well estab- lished in the plantings around the mansion of the Crane Estate on Castle Hill in Ipswich with Galium verum L. which is not common in the country. Castle Hill, Ipswich, Francis C. Wade and Stuart K. Harris 18784 (26 September 1958). The extensive dump off the Salem Turnpike in Saugus con- tains a large number of garden escapes and adventive weeds. Here I found a few plants of Rapistrum rugosum (L.) All. which has a curious transversely two-jointed silique, the upper joint being almost spherical and bearing eight longitudinal ribs while the lower joint is much smaller, unribbed and looks like a pe- duncle. The species was previously known from Massachusetts from two specimens collected in the 1880's on the old ballast flats in South Boston. Dump off Salem Turnpike, Saugus, Stuart K. Harris 18716 (16 September 1958). Specimens have been deposited in the herbarium of the New England Botanical Club.—SruART K. Harris, BOSTON UNIVERSITY A REVISION OF THE HOUSTONIA PURPUREA GROUP (RUBIACEAE) Epwarp E. TERRELL (Continued from page 180) One of the more striking examples of differences between geographically distant colonies was observed in a few plants trans- planted from Chautauqua County, New York, to southern Ohio. The plants were grown for several months and overwintered. In winter they produced a number of short, erect, basal offsets. In contrast, plants from southern locations have merely a basal ro- sette in winter. There are, also, a few other differences between northern and southern plants. Either the differences vary gradu- ally and apparently clinally from north to south, or else they vary in an irregular manner; the natural existence of races and eco- types is not apparent. For such reasons, it seems desirable to leave H. canadensis as a heterogeneous species without recogniz- ing within it any infraspecific taxa. This species intergrades with H. purpurea var. calycosa in southern Ohio, in Clark County, Indiana, and in central and western Kentucky. It is believed that these are actual hybrids. That the two species may occur together without producing in- tergrades was suggested by one observation made in Rowan County, Kentucky. Like other members of the H. purpurea group, the two taxa may sometimes cross, sometimes not. A lack of suitable intermediate habitats (“hybridization of the habitat” of Anderson) may not be the factor restricting survival of inter- mediates. H. canadensis in its typical form is distinct from all other taxa and may usually be distinguished by the following combination of characters: stem with 3-6 internodes; lower and/or middle in- ternodes conspicuously longer than upper ones; basal leaves conspicuous, pubescent above and ciliolate, persisting through the fruiting stage; cauline leaves oblanceolate, elliptic, or obo- vate, usually on one plant at least one pair of the middle and upper cauline leaves are widest near the apex, and often most 188 1959] Terrell — The Houstonia purpurea Group 189 of the leaves are widest near the apex; calyx-lobes 1.6-3.5 mm. long. General differences between this species and H. longifolia var. longifolia were illustrated rather well by John Torrey in his Flora of the State of New York, Vol. 1, Plate 44, 1843. The following collections from northern Illinois are intermediate between H. canadensis and H. longifolia var. longifolia: KANE: Aurora, Boyce, June 1885 (GH). LASALLE: Starved Rock, Greenman et al, 1-7 June 1909 (F, GH, 1A, Ny) ; Starved Rock, Umbach, 29 May 1901 (F, wis). WILL: Joliet Mound, Skeels 2/4 (F, GH, Ny, Us). In Ogle County one collection from Oregon, Gates 2723 (MICH), is closer to H. canadensis, but other collections from this county cited under var. longifolia are typical of that taxon. The writer collected (Terrell 2308) at Starved Rock State Park in LaSalle County in 1951, and confirmed to his own satisfaction that this population is intermediate. Two other collections from other states definitely suggest intermediacy between these same two taxa: New York. GENESEE: LeRoy, Hill, 7 July 1893 (r). Michigan. ST. JOSEPH: Three Rivers, Wheeler & Yoohida, 5 June 1890 (us). One Wisconsin specimen of nearly typical H. canadensis is believed to be an error in location; in spite of intensive collecting in this state no later records exist: MONROE: Sparta, Hale, 1861 (r). Another sheet in herb. wis with an identical inscription is var. longifolia. REPRESENTATIVE COLLECTIONS: — Ontario. BRUCE: damp sandy shores of Lake Huron, Red Bay, Stebbins et al 209 (cu, Gu, WVA) ; Frog Island no. 16, Grassl 4954 (mich) ; Stokes Bay, Krotkov 9416 (Ny, us); Sauble Beach, Gleason, 20 June 1934 (DUKE). LINCOLN: Queenstown Hts., Wilkinson, 12 June 1886 (oc). NORTHUMBERLAND: Marie-Victorin et al 46014 (F, GH). New York. CHAUTAUQUA: Chautauqua Gorge, Muens- cher & Brown 21709 (cU). MONROE: gorge of Genesee R. at Rochester, Matthews 2807 (NcU). NIAGARA: Niagara Falls, Lorenz, June 1916 (Ncsc). Pennsylvania, WESTMORELAND: Big Pucketa Cr. near Parnassus, Patterson, 19 May 1915 (wo). West Virginia. TYLER: Middlebourne, Core 4126 (wvA). Michigan. ALPENA: stony flat, Thunder Bay Island, Ehlers 3172 (MICH, WIS). PRESQUE ISLE: Thompson’s Harbor, Hinshaw, 14 June 1931 (MicH). wAsHTENAW: Ypsilanti, Billington, 19 May 1919 (MicH). Ohio. ERIE: Castalia, Moseley, 28 May 1922 (GH, MICH, os). FRANKLIN: Georgesville, Osburn, 4 May 1895 (os). LAKE: Madison, Werner, 25 May 1885 (ny, os). Indiana. CLARK: 3 mi. Nw of Henry- ville, Deam 55770 (p.) Kentucky. CARTER: Tygarts Creek, Braun 1707 (BRAUN); near Carter Caves, Gilbert, 20 May 1939 (GH). JESSAMINE: High Bridge, King 47 (F); Camp Nelson, McFarland 110 (mo, Us). LINCOLN: 3 mi. sE of Crab Orchard, Wharton 2085 (Ncsc) ; near Stan- 190 Rhodora [VOL. 61 ford, Biltm. Herb. 7899b (vs). PuLAskr just N of Burnside Bridge, Terrell 1915 (cu). Tennessee. MARION: below Hales Bar Dam, Fair- child et al 48-93 (TENN). WHITE: w of DeRossett, Weatherby & Weath- erby 6257 (GH, NY, TENN). 3. Houstonia tenuifolia Nutt., Gen. 1:95. 1818. Hedyotis longifolia (Gaertn.) Hook. B tenuifolia (Nutt.) T. & G. Fl. N. Am. 2:40. 1841. Oldenlandia purpurea (L.) Gray var. tenuifolia (Nutt.) Gray ex Chap- man, Fl. S. U. S. 181. 1860. Houstonia longifolia Gaertn. p tenuifolia (Nutt.) Wood, Class-Book, ed. 1861. 403. 1861. Houstonia purpurea L. var. tenuifolia (Nutt.) Gray, Syn. Fl. N. Am. 1(2):26. 1884. Chamisme tenuifolia (Nutt.) Nieuwl. Am. Midl. Nat. 4:92. 1915. Houstonia tenuifolia Nutt. f. leucantha Standley, Rhod. 34:177. 1932. Hedyotis Nuttalliana Fosberg, Va. Jour. Sci. 2:111. 1941. Hedyotis purpurea (L.) T. & G. var. tenuifolia (Nutt.) Fosberg, Cas- tanea 19:35. 1954. Hedyotis purpurea var. tenuifolia f. leucantha (Standl.) Fosberg, 1. c. 36. 1954. Perennial herbs. Rhizomes simple or branched, erect or horizontal, shortened or to 3.5 cm. long, bearing many small roots. Stems erect or ascending, one-several, round-tetragonal, (0.8-) 1-4 (-5.7) dm. high, densely cinereous-puberulent below, especially at nodes, with hairs on internodes usually less than 0.2 mm. long, or varying to glabrate. Inter- nodes numbering (3-) 4-9 (-11); median internodes 3-6 (8) cm. long. Nodes usually swollen. Stipules broad-ovate or broad-deltoid to ovate-lanceolate, entire, erose, or few-toothed at apex, often with dark glands at apices of teeth, obtuse, acute, or acuminate, to 4 mm. long, to 3 mm. wide; upper stipules often irregularly lobed. Basal leaves forming a rosette in winter, sometimes withered at time of flowering or sometimes persisting through flowering and fruiting (especially in plants of Missouri and Arkansas), varying from oval to spatulate, taper- ing into petioles shorter than to rarely longer than blades, to 3 cm. long, to 1 cm. wide, glabrous below, glabrous or sparsely pubescent above, sometimes sparsely ciliolate. Cauline leaves I-nerved; lower leaves sessile or short-petiolate, oblanceolate to elliptic to linear; median leaves sessile, linear to narrowly elliptic or very narrowly oblanceolate, 1.3-4.7 cm. long, 0.5-4.0 (-5.0) mm. wide, usually at least 7 times longer than wide, glabrous below, glabrous or scabrous-pubescent above; upper leaves similar, smaller. Branches ascending, spreading. divaricate, or deflexed, slender and often ultimately filiform, usually with 1-4 remote nodes bearing reduced leaves, to 20 (-28) cm. long, usually arising from uppermost 3-6 nodes and forming a very diffusely open, few-very many 1959] Terrell — The Houstonia purpurea Group 191 flowered inflorescence; often main stem equal in thickness to its branches, each node therefore appearing to give rise to 3 fertile branches. Pedicels filiform or slender, to 14 (-20) mm. long. Calyces glabrous; calyx-lobes erect, lanceolate, linear-lanceolate, or deltoid- lanceolate (0.5-) 0.8-2.4 (-3.0) mm. long, 0.2-0.9 mm. wide, less than one-half as long as corolla-tube, equalling or slightly exceeding mature capsules. Corollas purplish to white or variously purplish-tinged, (5-) 6-9 mm. long, granular to puberulent (rarely villous-pubescent) within; corolla-tubes 3-6 mm. long, 1.5-3.0 mm. wide distally; corolla-lobes 1.5-3.0 mm. long, 1.2-2.0 mm. wide. Mature capsules 1.5-2.5 (-3.0) mm. long, 1.5-2.5 (-3.0) mm. wide. Seeds 0.50-1.20 mm. long, 0.35-0.90 mm. wide. Time of flowering: Late April or May through July. Earliest recorded flowering date is in late April; latest date is in late August. Fruits from the earliest flowers mature in July or August and continue to mature in order of development through sum- mer and autumn. Type locality: “Near the confluence of Pidgeon river, and the French Broad, Tennessee, on dry gravelly hills". Type: Not seen. Habitats and distribution: Dry or somewhat mesic woods of oak, oak-hickory, oak-pine; thin soil over various kinds of rocks; talus slopes; crevices in cliffs; openings and woodland borders. Commonly in open or lightly shaded places in well-drained, xeric-mesic habitats. Northwestern, southeastern, and south- western Virginia; Fayette County, West Virginia; in North Caro- lina locally abundant in the Piedmont (where often around gran- ite outcrops) and in French Broad drainage in Madison and Haywood Counties (Great Smokies); scattered and local in Pied- mont of South Carolina; frequent in Cumberland Mountains of southeastern. Kentucky and in Great Smokies of ‘Tennessee; northwestern Georgia and adjacent Tennessee south to Heard County, Georgia; central Alabama; southern half of Missouri, northern half of Arkansas, and Leflore County, Oklahoma (main- ly in Ozarks and Ouachitas); Ozark Hills of southern Illinois and adjacent westernmost Kentucky. (Map 2) Nuttall's description is unequivocal for this species. His type has not been located, although specific attempts were made to 192 Rhodora [VOL. 61 check the herbaria of the Philadelphia Academy of Natural Sci- ences, the New York Botanical Garden, and the British Museum. This species has two main centers of distribution — the Ap- palachians and the Ozarks-Ouachitas. It appears to be absent from the territory between these centers. In the Appalachians it is locally disjunct and is absent in some areas which seem to have suitable habitats for it. It was reported from Texas and Mexico by Standley (1918), but no specimens have been seen from either. Shinners (1949) stated that he had not seen any specimens from ‘Texas. Within this species typical Appalachian and typical Ozark- Ouachita plants differ as follows: Ozark-Ouachita Internodes 4-7 Basal leaves often conspicuously large and sparsely ciliolate Middle cauline leaves averaging slightly wider-narrow-elliptic Branches usually less than 10 cm. long, erect or ascending Pedicels less than 8 mm. long Calyx-lobes longer, 1-2.4 (-3.0) mm. long Capsules larger, 1.5-3 mm. long Appalachian Internodes 4-9 Basal leaves usually smaller and glabrous Averaging slightly narrower-linear or very narrowly elliptic Usually 9-20 cm. long, more often spreading to divaricate Less than 14 mm. long Shorter, 0.8-1.5 (-2.3) mm. long Smaller, 1.5-2.5 mm. long This comparison is based on typical plants. Plants from Ala- bama and Georgia tend to be somewhat more like the Ozark plants. A number of collections from the two centers are very similar, e.g., collections from Rich Mountain, Arkansas, have filiform branches or pedicels like many Appalachian collections. Some piedmont, North Carolina, collections are very similar to Ozark collections. In general, the degree of overlap is so great that I have preferred not to distinguish plants of the two centers as separate subspecies, but consider them all part of one variable species. In addition, it is possible that some of the differences may be environmental. It appears that populations in these two cen- ters have been isolated from each other for a rather long time, long enough for certain differences to have arisen. In the past collectors have not always been sure whether Ozark- Ouachita plants should be identified as H. tenuifolia or as H. 1959] Terrell — The Houstonia purpurea Group 193 longifolia. That the plants have been more often referred to the latter is due, apparently, to their wider leaves and shorter branch- es as compared with the narrower leaves and "setaceously pedun- culate" flowers (as described by Nuttall) of the Appalachian group. Within each group populations vary considerably from locality to locality; some of this may be environmental, while chance isolation of certain biotypes in certain local areas may explain other variation. A collection from Walton Co., Florida, is quite atypical, being almost intermediate between this species and H. longifolia var. compacta. Since its characteristics are slightly more than those of H. tenuifolia, it has been placed in this species. The relationship between this species and H. longifolia var. compacta, is another source of confusion in the H. purpurea. group. It is sometimes very difficult to separate them, especially in herbarium collections which are not adequate samples of the population. At other times they appear quite distinct. It was formerly believed by the writer that they were two subspecies of the same species, H. tenuifolia, and on several occasions they were so annotated. Both taxa have densely cinereous-puberulent stems, and the cauline leaves seem basically similar although dif- ferent in average length / width ratios. The writer was impressed, also, by the way they intergrade as one travels north from the Cumberland Mountains into the transitional counties of the Ap- palachian Plateau in eastern Kentucky. Two collections from Floyd and Morgan Counties seemed to indicate introgression into var compacta, while collections from Breathitt and Knott Coun- ties indicate introgression into the other species. This is now interpreted as actual introgression, as originally supposed. Puta- tive hybrids occur, also, in West Virginia and Virginia (cited below). However, introgression has not broken down the dis- tinctness of these two taxa over large parts of their ranges. After observing how basically different they are in Alexander and Ire- dell Counties, western North Carolina, and in parts of south- western Virginia, it was concluded that they should be placed in separate species. One of the more impressive differences is in time of initiation of flowering — this may differ by one to four wecks, 194 Rhodora [VOL. 61 with two to three weeks being an average. In Alexander-Iredell counties they occur within 214 miles of each other in rather simi- lar habitats, but were not found growing together. In this arca they appeared as different species, being different in as many as 6-10 characteristics. Intergrades between H. longifolia var. compacta and H. tenuifolia include the following collections: Virginia. PAGE: roadside, Skyline Drive, 0.7 mi. s of Lewis Mt., Walker 2923 (us). ROCKINGHAM: Swift Run, Boelicher 501. (us). West Virginia. rAvErTE: Kimberly near mouth of Armstrong Cr., Smithson, 14 July 1936 (wvA). GREENBRIER: open upland woods, White Sulphur, L. F. & F. R. Randolph 1291 (cu. n). MCDOWELL: open woods between Barrenshe Cr. and Dry Fork, alt. 1350- 1800 ft., Morris 1162 (us) REPRESENTATIVE SPECIMENS: Virginia. CARROLL: along New R. at mouth of Brush Creek, Small, 12 July 1892 (r, GH, MO, NY, PH, US). MECKLENBURG: 6 mi. N of Clarksville, Fosberg 15441 (GH. TENN, WVA). o MAP 3 9 H. LONGIFOLIA o VAR. LONGIFOL IA VAR. COMPACTA VAR. GLABRA B ATYP. LONGIFOLIA LONG. -COMP. - CAN. -LONG. 0 o b L] OO p » c Jh o | O nre MAP 3. Distribution of the three varieties of Houstonia longifolia; atypical collections of var. longifolia; intergrades between var. longifolia and var. compacta; intermediates between var. longifolia and H. canadensis. Canadian collections of typical var. Jongifolia from vicinities of McKague, Saskatchewan, and Edmonton, Alberta, are not shown on map 1959] Terrell — The Houstonia purpurea Group 195 SOUTHAMPTON: sw of Applewhite Church, Fernald & Long 10431 (F, GH). North Carolina. ALEXANDER: 214 mi. w of Vashti, Radford 13868 (NCU). FRANKLIN: Cedar Rock, near Castalia, Blomquist 13245 (DUKE) . GRAN- VILLE: Oxford, Godfrey 5473 (DUKE, NCU, US); 1 mi. s of Bullock, Fox 4894 (Ncsc). suRRY: Pilot Mt., D. S. & H. B. Correll 14502 (DUKE). South Carolina. LANCASTER: Forty Acre Rock, Ahles & Haesloop 27461 (NcU). Georgia. HEARD: 4 mi. sw of Franklin, Pyron & McVaugh 1766 (GA). Alabama. CHAMBERS: Earle, 22 June 1897 (GH, NY). TUSCALOOSA: 15 mi. above Tuscaloosa, Harper 3498 (GH, NY, US). Kentucky. HARLAN: near Harlan Court House, Kearney, Aug. 1893 (GH, MO, NY, US). HOPKINS: near Dawson Springs, Palmer 17692 (MO, PALMER) . Tennessee. BLOUNT: Cades Cove, Anderson 1297 (GH, 1A, MO). COCKE: Wolf Creek, Kearney 736 (cu, MO, NY, US): vicinity Cosby, Raper & Jennison 3296 (TENN). Illinois. jackson: 114 mi. sE of Gorham, Bauer 2698 (F). Jonsson: 1.5 mi. s of Goreville, Winterringer 984 (ncu). Missouri. BARRY: Eagle Rock, Bush 100 (us, wvA) ; same loc., Bush 15411 (wis). HOWELL: Willow Springs, Palmer 6228 (cu, r, MO). scorr: between Chaffee & Rockview, Steyermark 5011 (F, MO). STODDARD: 214 mi. w of Puxico, Steyermark 66133 (F). Arkansas. CARROLL: Eureka Springs, Palmer 5553 (cu, F, MO). LAWRENCE: Imboden, Demaree 30444 (us). LOGAN: Magazine Mt., Clausen 7661 (cv). Oklahoma. LEFLORE: Rich Mt., Stevens 2667 (GH, MO, NY, US). 4. Houstonia longifolia Gaertner, Fruc. 1:226. 1788. Houstonia longifolia Willd. Sp. Pl. 1 (2) :583-584. 1798. Houstonia longifolia Michx.; Torr. Fl. N. & M. U.S. 173. 1824. Hedyotis longifolia (Gaertn.) Hook. Fl. Bor. Am. 1:286. 1834. Anotis longifolia (Gaertn.) G. Don, Gen. Hist. 3.535. 1834. (in part) Oldenlandia purpurea (L). Gray var. longifolia (Willd.) Gray, Man. d. 2. 178. 1856. (in part) Houstonia purpurea L. var. longifolia (Willd.) Gray, Man. ed. 5. 212. 1868. (in part) Chamisme longifolia (Gaertn.) Nieuwl. Am. Midl. Nat. 4:92. 1915. Hedyotis canadensis (Willd. ex R. & S.) Fosberg, Va. Jour. Sci. 2:110. 1941. (in part) Hedyotis purpurea (L.) T. & G. var. longifolia (Gaertn.) Fosberg, Castanea 19:34. 1954. (in part) Perennial herbs. Rhizomes branched or simple, horizontal or erect, sometimes subligneous, to 7 cm. long, bearing many small roots. Stems erect or ascending, one-many, tetragonal or roundish-tetragonal, some- times slightly winged, 0.5-3.5 dm. high, densely cinereous-puberulent or pubescent or scabrous-pubescent below, especially at nodes, or gla- brous. Internodes numbering (3-) 4-11 (-13); median internodes 1-5 (-6) cm. long. Nodes sometimes rather swollen. Stipules lanceolate to 196 Rhodora [VOL. 61 ovate, entire, erose, toothed (teeth often tipped with dark glands), or irregularly lobed, rounded above or varying from obtuse to acuminate, to 3.3 mm. long, to 3.5 mm. wide. Basal leaves forming a rosette in winter, sometimes persisting through flowering and fruiting or some- times withered, oval, elliptic, or oblanceolate, tapering into petioles shorter or longer than blades, to 3.9 cm. long, 0.2-1.0 cm. wide, gla- brous below, glabrous to sparsely pubescent above, sometimes sparsely ciliolate. Cauline leaves I-nerved, lower ones sessile or short-petiolate; median leaves sessile, varying from narrowly elliptic to oblong to linear or very narrowly lanceolate or oblanceolate, 0.6-3.4 cm. long, 0.15- 0.60 cm. wide, about 4-11] (-14) times longer than wide, glabrous below, glabrous to somewhat pubescent above, margins glabrous to sparsely ciliolate; upper leaves similar, smaller. Branches ascending or spreading, slender, less than 12 cm. long, forming a rather open to rather compact, few-many flowered inflorescence. Pedicels slender, less than 8 mm. long. Calyces glabrous; calyx-lobes erect or spreading, linear-lanceolate to ovate-lanceolate, (0.5-) 1.0-2.5 (-3.1) mm. long, 0.3-0.9 mm. wide, less than two-thirds as long as corolla-tube, equalling to somewhat exceeding mature capsules. Corollas white to purplish or variously purplish-tinged, (4-) 5-8 (-9) mm. long, granular to puberulent within; corolla-tubes 3-6 mm. long, 1.5-4.5 mm. wide dis- tally; corolla-lobes 1.5-3.5 mm. long, 1.2-2.5 mm. wide. Mature cap- sules 1.7-3.8 mm. long, 1.7-3.3 mm. wide. Seeds 0.50-1.40 mm. long, 0.35-1.00 mm. wide. Gaertner described only fruits and seeds of H. longifolia. None of the included drawings or descriptions can be applied with cer- tainty to any particular member of the present group. One draw- ing was of calyces with rather short lobes and could not have depicted H. purpurea var. calycosa. The phrase, "Ex herbario Banksiano", was the only hint as to the source of Gaertner's plants. This suggested to the present writer that Joseph Banks was the collector, and that the type might be located either in the herbaria of Gaertner or Banks. The location of Gaertner’s herbarium apparently is unknown even to taxonomists in Europe. Correspondence with the British Museum, which has the herbarium of Banks, elicited the reply that no specimen resembling the type was to be found there. ‘The remaining possibility is that the type may yet be found in other collections by Banks, which are known to exist in two other herb- aria in Europe. Although the original description is ambiguous and cannot 1959] Terrell — The Houstonia purpurea Group 197 be applied with certainty, I prefer to retain Gaertner's name until all possibilities of finding the type material are exhausted. To replace this name with a new one would seem now slightly premature and would upset established nomenclature perhaps unnecessarily. i> a MAP 4 PE OSH AS PETET i : 1 Poa JU ANS TT O PURP.-TEN. (CH HIE Ac PAIRE A FVV e DLH CE NANON TR Hi C Pa OTH FX MAP 4. Distribution of intergrades between Houstonia purpurea and H. tenuifolia and between H. purpurea and H. longifolia var. compacta. Willdenow’s description in 1798 cited Gaertner’s publication and followed the citation with a question mark, from which it appears that Willdenow was uncertain whether his plants were taxonomically equivalent to those of Gaertner. This implies, also, that Willdenow never saw the specimen described by Gaert- ner. Although included above as a synonym, Willdenow’s des- cripuon is hardly adequate. I consider it probable that he re- ferred to the plants included by modern authors in H. longifolia. The second synonym listed is believed to have resulted from Torrey’s having erroneously ascribed the name to Michaux. Tor- rey cited Willdenow’s publication first; perhaps he intended to put his name in place of “Mich.” In his Flora Boreali-Ameri- cana Michaux did not mention H. longifolia, and no other men- tion of this name by him has been found. Torrey cited as another 198 Rhodora [VOL. 61 synonym, H. angustifolia Michx., but we now know that this name is a synonym of H. nigricans (Lam.) Fern., a species out- side the present complex. In Hooker's description the mention of certain northern lo- cations (Lake Huron, Lake Winnipeg, and Saskatchewan) leaves little doubt that he was describing H. longifolia in its strictest modern sense (group I, as listed below). The first part of Don's description was the same as Hooker's. An added series of phrases applied to H. nigricans, except that no member of the genus has "flowers scarlet." In Gray's Manual, ed. 2, the concept is much like that of mod- ern authors, except that Gray added, “A narrow-leaved slender form is H. tenuifolia Nutt.” In edition 5 the same description was repeated under a new combination. Fosberg (1941) used the name, Hedyotis canadensis, as includ- ing both that species and H. longifolia. Fosberg's (1954) com- bination was based on the judgment that the species of the H. purpurea complex were all varieties of one species. H. longifolia Gaertner has in recent usage been so circum- scribed as to include a variety of elements. These have seemed to intergrade in such a closely interwoven way that there is little wonder that they were never separated. This motley assortment has included what are believed to be hybrids and H. longifolia has served as a receptacle for many specimens that could not be referred to another species. After examination of herbarium specimens and after field observations and study of the popula- tions region by region, it is possible to break down the conglom- eration into the following morphological-geographical groups: Group l. A variable and heterogeneous series of populations ranging from New England across the Great Lakes states west to North Dakota and southern Canada to Alberta. These are restricted. entirely to glaciated territory. Certain populations scem to intergrade with other taxa. This group along with groups 2 and 3 really constitute the main body of traditional H. longi- folia. Group 2. Alleghenian and central Appalachian populations 1959] Terrell — The Houstonia purpurea Group 199 ranging south to South Carolina and Georgia. These are super- ficially somewhat like group | although less variable. Occasional specimens have been identified by collectors as H. tenuifolia, with which it intergrades. It intergrades also with group 5. Group 3. Essentially very similar to group 2. Restricted to southern Blue Ridge. Group 4. Ozark and Ouachita populations. Usually referred to H. longifolia but sometimes to H. tenuifolia. Group 5. Variable populations in southeastern "Virginia, North and South Carolina, Georgia, and Tennessee. Mainly in coastal plain and piedmont. Group 6. Scattered miscellaneous populations elsewhere which resemble one of the other groups closely but have no defi- nite continuous ranges. TABLE I — Comparison of three varieties of Houstonia longifolia. var. glabra var. compacta var. longifolia Stems glabrous or Densely puberulent glabrous to pubescent glabrate No. of internodes: 7-10 6-13 4-7 Stipules often lanceolate, same as glabra ovate, rounded or obtuse, acuminate, smaller larger These two differ by other “tendencies” In aspect, much like compacta These two allopatric, but only 50-75 miles apart; both southern and Appalachian; non-intergrading These two with slightly overlapping ranges and a few intergrading populations; otherwise occu- pying large territories and well-separated geo- graphically; one southern, one northern in glaciated territory Higher altitudes; mesic Lower altitudes; Northern low alt.: habitats in s. Blue Ridge xeric-mesic; xeric-mesic centered in Appal. Plateau and Ridge and Valley Least variable Rather variable; Quite variable and Intergrades with H. pur- heterogeneous, due purea and locally to suspected former with H. tenuifolia hybridization in the present classification these groups are dealt with as fol- lows: Group | — H. longifolia var. longifolia; Group 2 — H 200 Rhodora [VOL. 61 longifolia var. compacta; Group 3 — H. longifolia var. glabra; Group 4 — part of H. tenuifolia; Groups 5 and 6 designated as putative hybrids or hybrid races, in most of which the <